{"meta":{"query_hash":"aac46070376e","filters":{"venue":"Database"},"cohort_total":105,"direct_labels_cover":0,"predictions_cover":105,"exported":105,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/aac46070376e","api":"https://metacan.xera.ac/api/v1/cohort?venue=Database"},"results":[{"id":"W1829370794","doi":"10.1093/database/bav059","title":"G4RNA: an RNA G-quadruplex database","year":2015,"lang":"it","type":"article","venue":"Database","topic":"DNA and Nucleic Acid Chemistry","field":"Biochemistry, Genetics and Molecular Biology","cited_by":52,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"Computer science; G-quadruplex; Information retrieval; Sequence (biology); Computational biology; Theoretical computer science; Data mining; Biology; DNA; Genetics","score_opus":0.03413014463313407,"score_gpt":0.2888167434363996,"score_spread":0.25468659880326555,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1829370794","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.020301664,0.0038558738,0.06356202,0.0005318934,0.00018180035,0.0008217364,0.8332697,0.06359156,0.013883715],"genre_scores_gemma":[0.023578897,0.001574647,0.07059161,0.000384664,0.000045796638,0.00062426174,0.89795554,0.001987455,0.0032571452],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99927276,0.00010227823,0.00019775618,0.0001856782,0.00017357527,0.000067929956],"domain_scores_gemma":[0.9985979,0.00039326964,0.00024685802,0.00024718003,0.00025344364,0.00026133127],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011885324,0.0012166898,0.0014547325,0.006107795,0.00077385636,0.0022534311,0.0031191364,0.0021273477,0.015112947],"category_scores_gemma":[0.004750975,0.0006189358,0.00069381035,0.0058132512,0.00037934133,0.002159591,0.0023176023,0.00097040914,0.020994494],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0033160162,0.0006136356,0.010031043,0.01426906,0.00041125316,0.0020982844,0.00070165016,0.007431303,0.10042159,0.01855902,0.51925147,0.32289574],"study_design_scores_gemma":[0.00050734106,0.00045547943,0.009991849,0.0005894491,0.00026050123,0.0014882146,0.0003901378,0.015439502,0.044425033,0.014120663,0.91209304,0.00023882651],"about_ca_topic_score_codex":0.0014547845,"about_ca_topic_score_gemma":0.0018673039,"teacher_disagreement_score":0.015112947,"about_ca_system_score_codex":0.0006221626,"about_ca_system_score_gemma":0.0019699354,"threshold_uncertainty_score":0.050557792},"labels":[],"label_agreement":null},{"id":"W1967204795","doi":"10.1093/database/bar044","title":"Tripal: a construction toolkit for online genome databases","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":75,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"Agriculture and Agri-Food Canada; Saskatchewan Pulse Growers; Agricultural Research Service; Clemson University; U.S. Department of Agriculture; National Science Foundation","keywords":"Computer science; World Wide Web; Python (programming language); Database; Biological database; User interface; Application programming interface; Bioinformatics; Biology","score_opus":0.0671172462474224,"score_gpt":0.2781780268173692,"score_spread":0.21106078056994682,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1967204795","genre_codex":"methods","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0011700027,0.0005102918,0.5633071,0.00031349045,0.00026271353,0.0006196271,0.06933032,0.35480547,0.009680996],"genre_scores_gemma":[0.012014855,0.0011765704,0.62317824,0.0005084653,0.00011388148,0.0026919702,0.24890853,0.101911314,0.009496172],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9975164,0.0004336353,0.00039340867,0.00051318115,0.00088299596,0.00026042858],"domain_scores_gemma":[0.99531823,0.001787878,0.00040140323,0.0012812653,0.0007336199,0.0004776324],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0034767187,0.002638746,0.0019882596,0.005583419,0.0018577335,0.0064176405,0.008222889,0.0015246753,0.06655944],"category_scores_gemma":[0.011432773,0.0037701211,0.0039886753,0.006774338,0.0010098707,0.0061277277,0.009567615,0.0056531117,0.06454714],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010252752,0.00014138687,0.0011299106,0.0051072864,0.0004113049,0.0014728868,0.0013639962,0.0048152427,0.014177229,0.029546186,0.769829,0.17098036],"study_design_scores_gemma":[0.00026463758,0.00008020945,0.0011007864,0.00043732734,0.00015022859,0.0015921758,0.0003824811,0.017678427,0.013113567,0.03215244,0.93279254,0.0002552508],"about_ca_topic_score_codex":0.0026063963,"about_ca_topic_score_gemma":0.0031502664,"teacher_disagreement_score":0.06655944,"about_ca_system_score_codex":0.0011721356,"about_ca_system_score_gemma":0.00341975,"threshold_uncertainty_score":0.22266346},"labels":[],"label_agreement":null},{"id":"W1974274010","doi":"10.1093/database/bas025","title":"HAltORF: a database of predicted out-of-frame alternative open reading frames in human","year":2012,"lang":"en","type":"article","venue":"Database","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":59,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Université de Sherbrooke","funders":"Canadian Institutes of Health Research","keywords":"Open reading frame; Human proteome project; Human genome; Ribosome profiling; Biology; Frame (networking); Computational biology; Transcriptome; Computer science; Genome; Translation (biology); Genetics; Gene; Gene expression; Messenger RNA; Proteomics; Peptide sequence","score_opus":0.04269381016478672,"score_gpt":0.32927023066012606,"score_spread":0.2865764204953393,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1974274010","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.023239143,0.01253066,0.011264787,0.00034401854,0.00017268094,0.00022074983,0.9304945,0.017024383,0.004709047],"genre_scores_gemma":[0.016977366,0.0027545698,0.020075263,0.00024236344,0.000052452568,0.00029506415,0.9578255,0.0006614748,0.0011159196],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.999493,0.000069361704,0.00011052989,0.00014828285,0.0001041053,0.00007457965],"domain_scores_gemma":[0.99888104,0.00041263748,0.00028567002,0.00012120371,0.00012088614,0.00017861776],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00094548165,0.0021425753,0.0022372887,0.0057717715,0.00083207747,0.001705945,0.0019932608,0.0026004813,0.017949896],"category_scores_gemma":[0.0034066832,0.00070142903,0.0011777355,0.00405228,0.00035590553,0.0015390847,0.0014120593,0.0010337397,0.018696306],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.005594535,0.00030561438,0.015011633,0.017071774,0.00071390095,0.0054536914,0.00070814206,0.0043523973,0.06634769,0.0057654753,0.66281545,0.21585968],"study_design_scores_gemma":[0.0012400246,0.00048673703,0.037637655,0.0016708585,0.00057812885,0.0069712456,0.00039240022,0.0071620704,0.029659934,0.0058814613,0.90795165,0.00036781258],"about_ca_topic_score_codex":0.0025890674,"about_ca_topic_score_gemma":0.0034618324,"teacher_disagreement_score":0.017949896,"about_ca_system_score_codex":0.000981905,"about_ca_system_score_gemma":0.0019994131,"threshold_uncertainty_score":0.0600484},"labels":[],"label_agreement":null},{"id":"W1993679825","doi":"10.1093/database/bau061","title":"Finding needles in haystacks: linking scientific names, reference specimens and molecular data for Fungi","year":2014,"lang":"en","type":"article","venue":"Database","topic":"Plant Pathogens and Fungal Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":481,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Alberta Biodiversity Monitoring Institute","funders":"National Human Genome Research Institute; Agricultural Research Service; College of Pharmacy, University of Michigan; National Institutes of Health; National Brain Research Centre; University of Michigan; Ministry of Agriculture, Forestry and Fisheries; U.S. National Library of Medicine; University of Alberta; Beef Cattle Research Council; Université Catholique de Louvain","keywords":"RefSeq; Biology; Phylogenetic tree; Identification (biology); DNA sequencing; Cistron; Computational biology; Ribosomal DNA; Internal transcribed spacer; Information retrieval; Genome; Computer science; Genetics; DNA; Gene; Ecology","score_opus":0.042030844701357574,"score_gpt":0.282761934017969,"score_spread":0.24073108931661144,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1993679825","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.043648478,0.026489815,0.7472098,0.018939476,0.0091041885,0.0017377445,0.08167106,0.045655716,0.025543818],"genre_scores_gemma":[0.045607965,0.010140573,0.8190668,0.0037774872,0.00096486683,0.00067637576,0.10340765,0.008005426,0.008352903],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98365676,0.006111185,0.004071822,0.0023405342,0.0033408524,0.0004788711],"domain_scores_gemma":[0.9075709,0.03733968,0.013114515,0.027336365,0.011514565,0.0031240694],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0354096,0.0020293728,0.0025148548,0.029701779,0.004344898,0.01446067,0.0038540727,0.003963175,0.016795518],"category_scores_gemma":[0.13728061,0.0023920075,0.001689758,0.025477191,0.0023782344,0.020845363,0.013684553,0.003936099,0.016687613],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011594292,0.00040977873,0.03471677,0.009298661,0.00065739,0.00337705,0.008696296,0.001382406,0.021323677,0.031801175,0.20288654,0.68429077],"study_design_scores_gemma":[0.00007837881,0.00016556185,0.019070689,0.0061611775,0.00046404375,0.0028354404,0.0047222087,0.005141701,0.012052776,0.06583986,0.8829799,0.00048808023],"about_ca_topic_score_codex":0.0048382897,"about_ca_topic_score_gemma":0.007699289,"teacher_disagreement_score":0.0354096,"about_ca_system_score_codex":0.0024799611,"about_ca_system_score_gemma":0.006900495,"threshold_uncertainty_score":0.18726611},"labels":[],"label_agreement":null},{"id":"W2005813445","doi":"10.1093/database/baq031","title":"LPS-annotate: complete annotation of compositionally biased regions in the protein knowledgebase","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":24,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University","funders":"McGill University","keywords":"UniProt; Annotation; Protein sequencing; Computational biology; Genome; Biology; Database; Computer science; Gene; Genetics; Peptide sequence","score_opus":0.062374608092546455,"score_gpt":0.260019375891962,"score_spread":0.19764476779941553,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2005813445","genre_codex":"dataset","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.027950523,0.0027206093,0.33716998,0.00054250733,0.00024215941,0.0005542537,0.40291032,0.21881609,0.009093562],"genre_scores_gemma":[0.05376603,0.0016885346,0.357296,0.00040396897,0.00007429885,0.0010379339,0.56540203,0.0175098,0.002821363],"study_design_codex":"not_applicable","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99832684,0.00020731488,0.00026883913,0.00050617615,0.0005441493,0.00014665442],"domain_scores_gemma":[0.9964574,0.001140635,0.00059364864,0.0008840511,0.0006827396,0.00024140002],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00253512,0.0026151438,0.0017593907,0.0059493766,0.0014238023,0.003615297,0.002664575,0.0022594074,0.012645543],"category_scores_gemma":[0.009431352,0.0013167014,0.0014819616,0.0053883125,0.0006366582,0.0038905202,0.0039463476,0.0019814342,0.016824836],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0032341017,0.00064251083,0.01909129,0.010509484,0.0006751618,0.0018176414,0.0025970498,0.012608311,0.12634477,0.013591774,0.44888112,0.36000678],"study_design_scores_gemma":[0.0004619686,0.00034704673,0.0225528,0.0012071991,0.00040992213,0.0020935258,0.0010048285,0.058344405,0.08701586,0.02229316,0.80378515,0.00048410817],"about_ca_topic_score_codex":0.0028814166,"about_ca_topic_score_gemma":0.0039199777,"teacher_disagreement_score":0.012645543,"about_ca_system_score_codex":0.0012053789,"about_ca_system_score_gemma":0.0030624121,"threshold_uncertainty_score":0.042303562},"labels":[],"label_agreement":null},{"id":"W2047689005","doi":"10.1093/database/baq018","title":"The systematic annotation of the three main GPCR families in Reactome","year":2010,"lang":"en","type":"article","venue":"Database","topic":"Receptor Mechanisms and Signaling","field":"Biochemistry, Genetics and Molecular Biology","cited_by":25,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"National Human Genome Research Institute; Sixth Framework Programme; National Institutes of Health; European Commission","keywords":"G protein-coupled receptor; UniProt; Computational biology; Computer science; Annotation; Database; Biology; Bioinformatics; Receptor; Gene; Genetics","score_opus":0.00734967211467777,"score_gpt":0.22838758613593957,"score_spread":0.2210379140212618,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2047689005","genre_codex":"dataset","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.059687037,0.0036252318,0.34924933,0.0008658515,0.00059212325,0.0013516267,0.49033585,0.07957712,0.014715807],"genre_scores_gemma":[0.06573213,0.0025358342,0.25745165,0.00038567543,0.00008439159,0.0018258485,0.66029584,0.006770581,0.004917995],"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","domain_scores_codex":[0.9974661,0.00052133645,0.0005122841,0.0007718356,0.00050960225,0.00021888706],"domain_scores_gemma":[0.99650264,0.0009531777,0.00052636984,0.0010068667,0.00078647025,0.00022441435],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0033511766,0.0018800613,0.0013598171,0.008866987,0.001302669,0.0015477572,0.0014001724,0.00085366936,0.009304873],"category_scores_gemma":[0.006440854,0.00079222576,0.002722848,0.006620904,0.00056102645,0.0015338368,0.0024323575,0.0016193613,0.007609966],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00456431,0.00033301767,0.0246964,0.018464336,0.001127521,0.0043453565,0.0027780232,0.010306561,0.47601703,0.03503347,0.1639377,0.2583962],"study_design_scores_gemma":[0.00029484215,0.00028051354,0.041758172,0.0013357919,0.00093791157,0.003605118,0.000573908,0.022264812,0.13527797,0.01463866,0.7786707,0.00036159524],"about_ca_topic_score_codex":0.0021203493,"about_ca_topic_score_gemma":0.0029323618,"teacher_disagreement_score":0.009304873,"about_ca_system_score_codex":0.00092356507,"about_ca_system_score_gemma":0.00268595,"threshold_uncertainty_score":0.03112793},"labels":[],"label_agreement":null},{"id":"W2051020903","doi":"10.1093/database/bat087","title":"DBM-DB: the diamondback moth genome database","year":2014,"lang":"en","type":"article","venue":"Database","topic":"Insect Resistance and Genetics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":69,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ministry of Agriculture; University of British Columbia","funders":"National Natural Science Foundation of China; Yuhan","keywords":"Diamondback moth; dBm; Genome; Annotation; Gene Annotation; Plutella; Genomics; Biology; Computer science; Database; Computational biology; Gene; Bioinformatics; Genetics; Larva; Ecology","score_opus":0.007791138652176461,"score_gpt":0.229161108559944,"score_spread":0.22136996990776756,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2051020903","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0027456288,0.0014475584,0.005492151,0.0002621233,0.000096431344,0.00018264566,0.975116,0.009399641,0.005257893],"genre_scores_gemma":[0.0017728646,0.00047695465,0.005903274,0.00009831338,0.000010848586,0.00010848094,0.9904546,0.00045335823,0.0007212602],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9994918,0.00005598999,0.00009325815,0.00014013439,0.0001437216,0.00007501082],"domain_scores_gemma":[0.9993604,0.0000763923,0.0001029354,0.00010348582,0.00016643482,0.00019034464],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00097201875,0.0016435143,0.0020042134,0.004112936,0.00095578836,0.0021170478,0.0033247499,0.0015292984,0.022870801],"category_scores_gemma":[0.0019965046,0.00081886083,0.00088579056,0.0072821216,0.00024403683,0.0022244302,0.0021236683,0.0018301125,0.024368625],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012000768,0.00020371238,0.0029546376,0.005529315,0.00019992089,0.00041292317,0.00030671875,0.0009047548,0.028003957,0.0042983596,0.9030757,0.05290996],"study_design_scores_gemma":[0.00035381416,0.00007844954,0.008912103,0.00036050775,0.00013744918,0.00042563796,0.00015706682,0.0019508594,0.0065977564,0.0034414623,0.97747576,0.00010920294],"about_ca_topic_score_codex":0.007492922,"about_ca_topic_score_gemma":0.0068587125,"teacher_disagreement_score":0.022870801,"about_ca_system_score_codex":0.00077898963,"about_ca_system_score_gemma":0.0023798335,"threshold_uncertainty_score":0.07651049},"labels":[],"label_agreement":null},{"id":"W2099131261","doi":"10.1093/database/baq037","title":"The Biomolecular Interaction Network Database in PSI-MI 2.5","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":97,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"National Human Genome Research Institute; National Institutes of Health; Ontario Genomics; Ontario Genomics Institute; Genome Canada","keywords":"Identifier; Computer science; Database; Download; World Wide Web","score_opus":0.019740439397880078,"score_gpt":0.24994526918287382,"score_spread":0.23020482978499374,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2099131261","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009976448,0.00096817076,0.11125075,0.0009930414,0.00029249038,0.00040837366,0.66520506,0.19446103,0.016444594],"genre_scores_gemma":[0.019724771,0.0009553418,0.07794048,0.00029969693,0.000059200655,0.0007058885,0.88519025,0.011593491,0.0035308448],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99873537,0.00023083112,0.0002111892,0.00026973823,0.00044336045,0.00010944978],"domain_scores_gemma":[0.9980818,0.0005130983,0.00026902804,0.00058218173,0.00033291444,0.0002210212],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025396673,0.0017585099,0.0018581658,0.004220401,0.0010914599,0.0030247506,0.0031968271,0.00094365695,0.03522337],"category_scores_gemma":[0.0062006135,0.0013126482,0.0012917393,0.005935161,0.0004715237,0.004099923,0.0031156184,0.0034621402,0.03055013],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0020069259,0.00024810614,0.0052202055,0.0043487945,0.0005199931,0.0005184501,0.0005690087,0.0054127784,0.02267069,0.030280884,0.86212546,0.066078834],"study_design_scores_gemma":[0.0005253207,0.00014495615,0.012003554,0.00038820313,0.00028699622,0.00076453836,0.00023626407,0.034194157,0.022107169,0.034970827,0.89416385,0.00021424438],"about_ca_topic_score_codex":0.0035829842,"about_ca_topic_score_gemma":0.0035533921,"teacher_disagreement_score":0.03522337,"about_ca_system_score_codex":0.0011558441,"about_ca_system_score_gemma":0.0028588956,"threshold_uncertainty_score":0.11783391},"labels":[],"label_agreement":null},{"id":"W2100190412","doi":"10.1093/database/bat070","title":"Phenol-Explorer 3.0: a major update of the Phenol-Explorer database to incorporate data on the effects of food processing on polyphenol content","year":2013,"lang":"en","type":"article","venue":"Database","topic":"Phytochemicals and Antioxidant Activities","field":"Medicine","cited_by":868,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Institut National Du Cancer; Danone","keywords":"Polyphenol; Database; Phenol; Food science; Computer science; Chemistry; Biochemistry; Organic chemistry","score_opus":0.06410583990095763,"score_gpt":0.2677329195989637,"score_spread":0.20362707969800609,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2100190412","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0064910487,0.0036386384,0.008471671,0.0007582478,0.0001669307,0.00026615223,0.9618809,0.01318966,0.0051367623],"genre_scores_gemma":[0.009309833,0.003090989,0.015395067,0.00047942682,0.00011292143,0.00040889496,0.9661835,0.0024281268,0.0025912442],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99745613,0.00040634323,0.0008746807,0.00036264412,0.000704793,0.00019547368],"domain_scores_gemma":[0.9861458,0.005313838,0.002991164,0.0018621555,0.0024070968,0.0012799774],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0037291648,0.0017435364,0.002403139,0.00938686,0.00042186637,0.0037135456,0.002678786,0.0011324665,0.027734336],"category_scores_gemma":[0.01610573,0.00085715624,0.001456314,0.011361265,0.0004075902,0.0026115556,0.002729984,0.0016496982,0.01864094],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.005221948,0.0003239669,0.023852788,0.026913414,0.0013260993,0.001030786,0.00081152935,0.0025217652,0.014970353,0.009021143,0.714444,0.19956233],"study_design_scores_gemma":[0.0005951766,0.00017259056,0.019536274,0.0011966217,0.000520452,0.00059766846,0.00016580058,0.0018016123,0.0048370445,0.003288564,0.96710825,0.00017996505],"about_ca_topic_score_codex":0.0042239213,"about_ca_topic_score_gemma":0.00455958,"teacher_disagreement_score":0.027734336,"about_ca_system_score_codex":0.00091003184,"about_ca_system_score_gemma":0.0038081482,"threshold_uncertainty_score":0.09278059},"labels":[],"label_agreement":null},{"id":"W2105772474","doi":"10.1093/database/bas031","title":"Phenol-Explorer 2.0: a major update of the Phenol-Explorer database integrating data on polyphenol metabolism and pharmacokinetics in humans and experimental animals","year":2012,"lang":"en","type":"article","venue":"Database","topic":"Phytochemicals and Antioxidant Activities","field":"Medicine","cited_by":188,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"European Social Fund; Instituto de Salud Carlos III; Institut National Du Cancer; Ministerio de Ciencia e Innovación; Centro de Investigación Biomédica en Red-Fisiopatología de la Obesidad y Nutrición; University of Alberta; Danone","keywords":"Polyphenol; Database; Phenol; Chemistry; Computer science; Biochemistry; Organic chemistry","score_opus":0.051556922381446754,"score_gpt":0.32397729840762185,"score_spread":0.2724203760261751,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2105772474","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.027273038,0.022622965,0.023435757,0.0021097392,0.00068572874,0.00075451104,0.8356103,0.07182632,0.015681705],"genre_scores_gemma":[0.01563507,0.009875134,0.027762309,0.001098342,0.00037029755,0.0007555438,0.93044525,0.00867659,0.005381407],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99761814,0.0003880951,0.00057197677,0.0003903942,0.00080741773,0.00022392097],"domain_scores_gemma":[0.9923436,0.0025755512,0.0018397412,0.0010340229,0.0011193053,0.0010877291],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0050058053,0.0022463654,0.003163333,0.0069799554,0.0003709145,0.0038081482,0.0030342743,0.0010055054,0.020916525],"category_scores_gemma":[0.0082492735,0.0009737632,0.0016360593,0.0059161205,0.0004826574,0.003089661,0.0033074438,0.0016813483,0.018875876],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.010761046,0.00046104463,0.016182885,0.021885939,0.0015648371,0.0012479277,0.0007236474,0.0013584121,0.039810322,0.005257327,0.63774854,0.26299813],"study_design_scores_gemma":[0.0008288768,0.00040406748,0.020572267,0.00096666097,0.00069109176,0.0010309574,0.000117795294,0.0017815334,0.009130898,0.0028982877,0.9613092,0.00026834576],"about_ca_topic_score_codex":0.0017239625,"about_ca_topic_score_gemma":0.001685221,"teacher_disagreement_score":0.020916525,"about_ca_system_score_codex":0.00059823814,"about_ca_system_score_gemma":0.0030462327,"threshold_uncertainty_score":0.06997275},"labels":[],"label_agreement":null},{"id":"W2105939477","doi":"10.1093/database/bar053","title":"The Comprehensive Phytopathogen Genomics Resource: a web-based resource for data-mining plant pathogen genomes","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Plant Pathogens and Fungal Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":29,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University; Agriculture and Agri-Food Canada","funders":"","keywords":"Oomycete; GenBank; Genome; Biology; Genomics; UniProt; Genome browser; Computational biology; DNA sequencing; Reference genome; Genome project; Annotation; Genetics; Gene","score_opus":0.05445920231675954,"score_gpt":0.25408823258207636,"score_spread":0.1996290302653168,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2105939477","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009015672,0.0025193733,0.043488726,0.00027985344,0.000067398134,0.00061567116,0.88309,0.054998122,0.0059252325],"genre_scores_gemma":[0.0065058507,0.0010718905,0.074042425,0.00012244252,0.000024787707,0.00050500367,0.9131844,0.0030785482,0.0014645922],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9992494,0.000119770244,0.00010457201,0.00021238624,0.00023008487,0.0000836616],"domain_scores_gemma":[0.99852955,0.00041758432,0.00023285464,0.00028355504,0.00024096076,0.0002955251],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017686203,0.0023407314,0.0023842913,0.008512977,0.0012238707,0.0020855714,0.0026914696,0.0011828352,0.022712452],"category_scores_gemma":[0.003021557,0.001491381,0.0012676374,0.009804459,0.00038817193,0.0025308705,0.0029772667,0.0017465949,0.024632577],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0020163555,0.00068711006,0.010581706,0.009243318,0.0007484902,0.0013231492,0.0007116844,0.004803208,0.07677502,0.004417209,0.70268214,0.18601061],"study_design_scores_gemma":[0.0012979014,0.0003866901,0.037336066,0.0013451515,0.00052420906,0.0017891892,0.00047787305,0.019745128,0.0435831,0.013013204,0.87997353,0.0005280024],"about_ca_topic_score_codex":0.0040643355,"about_ca_topic_score_gemma":0.0060036466,"teacher_disagreement_score":0.022712452,"about_ca_system_score_codex":0.0007438542,"about_ca_system_score_gemma":0.0022983688,"threshold_uncertainty_score":0.07598066},"labels":[],"label_agreement":null},{"id":"W2106831214","doi":"10.1093/database/bar023","title":"The modENCODE Data Coordination Center: lessons in harvesting comprehensive experimental details","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Evolution and Genetic Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":34,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"Basic Energy Sciences; Office of Science; National Institutes of Health; Wellcome Trust; National Human Genome Research Institute; U.S. Department of Energy","keywords":"Computer science; Metadata; Protocol (science); Research center; Encyclopedia; Caenorhabditis elegans; Set (abstract data type); World Wide Web; Information retrieval; Library science; Biology; Programming language","score_opus":0.10123420813286024,"score_gpt":0.33206898989045897,"score_spread":0.23083478175759872,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2106831214","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0034934923,0.0029314298,0.71396196,0.039603505,0.0059833475,0.0028353748,0.091648884,0.1106607,0.02888133],"genre_scores_gemma":[0.00904778,0.0014832431,0.86596596,0.0045784237,0.00087172486,0.0025967457,0.08348162,0.026660427,0.005314066],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9402893,0.020414084,0.010232131,0.008863299,0.018317945,0.0018832004],"domain_scores_gemma":[0.5356182,0.10874885,0.011610346,0.26247048,0.066557206,0.014994894],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.15563415,0.003284753,0.0054937536,0.012780576,0.00567466,0.023108423,0.024687262,0.0052347104,0.030604335],"category_scores_gemma":[0.3158459,0.0054885666,0.0027265665,0.023978902,0.0066683246,0.03902732,0.018486433,0.0117718,0.0438577],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005598241,0.00019252085,0.0037251643,0.0014824987,0.00015059889,0.00020905337,0.001699797,0.0019568244,0.0035581684,0.07123757,0.6955961,0.21963193],"study_design_scores_gemma":[0.00023925275,0.000060396014,0.0017815678,0.0011283627,0.00006891024,0.0002891231,0.00067002064,0.0049873525,0.0050021927,0.063091576,0.92240787,0.00027331858],"about_ca_topic_score_codex":0.025436662,"about_ca_topic_score_gemma":0.01848761,"teacher_disagreement_score":0.84436584,"about_ca_system_score_codex":0.0090310145,"about_ca_system_score_gemma":0.029191868,"threshold_uncertainty_score":0.8230819},"labels":[],"label_agreement":null},{"id":"W2109144185","doi":"10.1093/database/bat068","title":"yApoptosis: yeast apoptosis database","year":2013,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Fondation Chalmers","keywords":"Programmed cell death; Yeast; Computational biology; Apoptosis; Biology; Database; Resource (disambiguation); Computer science; World Wide Web; Genetics","score_opus":0.009843973817082024,"score_gpt":0.22529482374096857,"score_spread":0.21545084992388655,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2109144185","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007322495,0.001882527,0.01895277,0.00026357675,0.00009327365,0.00027463952,0.9222129,0.04317636,0.0058214534],"genre_scores_gemma":[0.0087664435,0.0010700236,0.014329147,0.00008441564,0.000013729787,0.000383114,0.97286993,0.0013728453,0.0011102667],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9996226,0.000049227143,0.0000884692,0.00008988904,0.00011339397,0.00003639498],"domain_scores_gemma":[0.99939716,0.00012996387,0.00009004275,0.000100655736,0.0001810375,0.000101059915],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005528732,0.0014813,0.0012739792,0.0038378886,0.0007039346,0.0018338137,0.0024569726,0.00097274495,0.018232346],"category_scores_gemma":[0.0018869998,0.00061544636,0.00076680706,0.0044175982,0.00018631866,0.0011335636,0.0015921238,0.0010988002,0.019132487],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0025512078,0.0001950992,0.0062429034,0.009028039,0.0003461314,0.0016307329,0.00031778982,0.005374969,0.0262193,0.008009642,0.85653496,0.08354928],"study_design_scores_gemma":[0.000810153,0.0001311696,0.0098777395,0.0005768444,0.00022602556,0.0015650042,0.00025687637,0.016722983,0.018582031,0.010492743,0.9405989,0.00015964091],"about_ca_topic_score_codex":0.0026581618,"about_ca_topic_score_gemma":0.0029847939,"teacher_disagreement_score":0.018232346,"about_ca_system_score_codex":0.00064206624,"about_ca_system_score_gemma":0.0016307406,"threshold_uncertainty_score":0.060993254},"labels":[],"label_agreement":null},{"id":"W2111476723","doi":"10.1093/database/bar032","title":"NetSlim: high-confidence curated signaling maps","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":41,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Scheme for Promotion of Academic and Research Collaboration","keywords":"Computer science; Visualization; Set (abstract data type); Resource (disambiguation); Computational biology; Database; Information retrieval; Data science; World Wide Web; Data mining; Biology; Programming language","score_opus":0.022612611316079403,"score_gpt":0.2250225638603589,"score_spread":0.20240995254427951,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2111476723","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.025031915,0.001595021,0.31073183,0.00060240715,0.00058253936,0.00062857365,0.45124784,0.19180782,0.017772067],"genre_scores_gemma":[0.06650997,0.0010597807,0.26297066,0.0002618671,0.00009204896,0.0013709443,0.64546394,0.0156796,0.0065911976],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9986766,0.0002527795,0.000110285204,0.00034880074,0.0005080299,0.0001035361],"domain_scores_gemma":[0.99777704,0.0008832835,0.00026696545,0.00034863618,0.0005619048,0.00016218392],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019254137,0.002456767,0.0012632308,0.006182654,0.0012689318,0.0026751107,0.0031821143,0.0011969277,0.040016565],"category_scores_gemma":[0.0062507126,0.0010005163,0.0015968995,0.0031384379,0.00040804798,0.0021136005,0.0020108367,0.0019371427,0.0182388],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0025394682,0.00028899967,0.007279297,0.0096538635,0.00064270856,0.0016456343,0.0005695874,0.026254589,0.11663813,0.027146155,0.66425824,0.14308327],"study_design_scores_gemma":[0.0005202089,0.00017861457,0.013573724,0.00078582735,0.00037139474,0.0014540587,0.00027469898,0.12379132,0.12193448,0.034408767,0.7023639,0.00034305445],"about_ca_topic_score_codex":0.0027580895,"about_ca_topic_score_gemma":0.0046784882,"teacher_disagreement_score":0.040016565,"about_ca_system_score_codex":0.0011036011,"about_ca_system_score_gemma":0.0021183374,"threshold_uncertainty_score":0.13386875},"labels":[],"label_agreement":null},{"id":"W2113769137","doi":"10.1093/database/bav008","title":"mycoCLAP, the database for characterized lignocellulose-active proteins of fungal origin: resource and text mining curation support","year":2015,"lang":"en","type":"article","venue":"Database","topic":"Biofuel production and bioconversion","field":"Engineering","cited_by":32,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Resource (disambiguation); Database; World Wide Web; Data curation; Information retrieval","score_opus":0.04390842278676899,"score_gpt":0.25269453764247773,"score_spread":0.20878611485570875,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2113769137","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0068592275,0.008883377,0.028902907,0.00049052015,0.00022485275,0.00075892295,0.91702497,0.028470784,0.008384365],"genre_scores_gemma":[0.0048283287,0.0026793845,0.049638417,0.00024314925,0.000054754437,0.00069338555,0.93796694,0.002175171,0.0017204751],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9980709,0.00025021145,0.00037642717,0.0005278009,0.0006201776,0.00015457271],"domain_scores_gemma":[0.99683595,0.0008536977,0.0008102382,0.00041991632,0.0006902358,0.00039002768],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024116703,0.0041492945,0.004093758,0.013689073,0.0025119486,0.0037704578,0.0032754585,0.002075091,0.016278835],"category_scores_gemma":[0.0086640185,0.001256503,0.001954796,0.014856843,0.0006748195,0.0033827994,0.0030627367,0.0026847608,0.025790507],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0017459284,0.0003240467,0.0045923186,0.033095337,0.0007961937,0.002691192,0.00095594244,0.0015541037,0.11749224,0.0052327896,0.7142575,0.11726235],"study_design_scores_gemma":[0.00041705047,0.00014520381,0.011562252,0.002077484,0.0005446813,0.002115797,0.00024949812,0.0042626336,0.02735299,0.004230516,0.94686186,0.0001801122],"about_ca_topic_score_codex":0.004561638,"about_ca_topic_score_gemma":0.008955815,"teacher_disagreement_score":0.016278835,"about_ca_system_score_codex":0.001036813,"about_ca_system_score_gemma":0.0073812273,"threshold_uncertainty_score":0.05445814},"labels":[],"label_agreement":null},{"id":"W2118284237","doi":"10.1093/database/bas036","title":"Recent advances in biocuration: Meeting Report from the fifth International Biocuration Conference","year":2012,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":11,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Human Genome Research Institute; George Washington University; Wellcome Trust; Georgetown University; Institute of Genetics; University of Oxford; Ontario Institute for Cancer Research","keywords":"Library science; World Wide Web; Promotion (chess); Political science; Computer science","score_opus":0.03594583616335801,"score_gpt":0.3235142753631416,"score_spread":0.2875684391997836,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2118284237","genre_codex":"review","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.010091448,0.6386249,0.006999975,0.23352784,0.047366727,0.00019078929,0.0043178704,0.0008196793,0.058060717],"genre_scores_gemma":[0.087724276,0.67366594,0.03171231,0.053782646,0.031095084,0.0005180248,0.02551178,0.001403174,0.09458679],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9939355,0.0013078229,0.0005917962,0.00072471227,0.0024537705,0.0009863769],"domain_scores_gemma":[0.98013353,0.0031167723,0.0014112083,0.00086629,0.008069858,0.0064023724],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.02507118,0.0012143962,0.0008778839,0.004743088,0.001611072,0.009187875,0.0021180161,0.0039744214,0.017397212],"category_scores_gemma":[0.014121438,0.00046540002,0.0014143209,0.008806737,0.0012780653,0.006810134,0.0059801457,0.0032529968,0.008849174],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00019743554,0.000115142284,0.0031529283,0.0018752407,0.00008266776,0.00018593029,0.0007155735,0.00031476215,0.0012170626,0.004606392,0.755484,0.23205289],"study_design_scores_gemma":[0.000007956158,0.000040964183,0.003126605,0.00061674987,0.000042625354,0.00011211697,0.00056391925,0.00008205102,0.00074920245,0.0006378872,0.99399614,0.000023706152],"about_ca_topic_score_codex":0.006504679,"about_ca_topic_score_gemma":0.0088802185,"teacher_disagreement_score":0.9908121,"about_ca_system_score_codex":0.003259841,"about_ca_system_score_gemma":0.007861492,"threshold_uncertainty_score":0.13259065},"labels":[],"label_agreement":null},{"id":"W2118527670","doi":"10.1093/database/bar026","title":"International Cancer Genome Consortium Data Portal--a one-stop shop for cancer genomics data","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Cancer Genomics and Diagnostics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":672,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"","keywords":"Computer science; Genome; Genomics; Interface (matter); Genome browser; Software; Computational biology; World Wide Web; Biology; Gene; Genetics; Operating system","score_opus":0.12202126051096739,"score_gpt":0.3372931504936901,"score_spread":0.21527188998272268,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2118527670","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0025471265,0.002355534,0.0697664,0.009777087,0.001044517,0.0015750806,0.72745615,0.1294801,0.05599803],"genre_scores_gemma":[0.011575856,0.001792161,0.08738171,0.0021535095,0.00038872135,0.0012993263,0.87056017,0.013441861,0.011406752],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9953271,0.0012974091,0.0007132005,0.0006288419,0.0017203374,0.0003131214],"domain_scores_gemma":[0.9686621,0.008133035,0.0015771629,0.009335243,0.004512797,0.007779618],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011315188,0.0011964565,0.002586964,0.010796427,0.0011667208,0.00745831,0.0048724944,0.0021070724,0.097012386],"category_scores_gemma":[0.028560016,0.001542238,0.0009938002,0.024187898,0.0005985969,0.0073056957,0.008665226,0.0040521743,0.062455587],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00029293212,0.00008682019,0.0022971213,0.00057994085,0.00007635456,0.0001850543,0.00015725079,0.00039820606,0.00071297935,0.006451477,0.90791166,0.080850214],"study_design_scores_gemma":[0.00028608518,0.000047536207,0.0026447896,0.00023636752,0.00003140945,0.00028065775,0.00016142323,0.0013884911,0.0010832391,0.007075644,0.98668915,0.00007531064],"about_ca_topic_score_codex":0.006380338,"about_ca_topic_score_gemma":0.0066801673,"teacher_disagreement_score":0.097012386,"about_ca_system_score_codex":0.0018688177,"about_ca_system_score_gemma":0.006417294,"threshold_uncertainty_score":0.32453877},"labels":[],"label_agreement":null},{"id":"W2125951746","doi":"10.1093/database/bau108","title":"The Virtual Xenbase: transitioning an online bioinformatics resource to a private cloud","year":2014,"lang":"en","type":"article","venue":"Database","topic":"Genetics, Bioinformatics, and Biomedical Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"Eunice Kennedy Shriver National Institute of Child Health and Human Development; National Institute of Child Health and Human Development; University of Calgary","keywords":"Cloud computing; Resource (disambiguation); Computer science; World Wide Web; Operating system; Computer network","score_opus":0.021926973331574562,"score_gpt":0.2858661606597033,"score_spread":0.2639391873281287,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2125951746","genre_codex":"software","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.046406135,0.00302117,0.26570323,0.0128422985,0.004827153,0.0015667796,0.20848814,0.40547273,0.051672377],"genre_scores_gemma":[0.07113745,0.002619128,0.36503854,0.00535239,0.0010146827,0.0009965023,0.49119762,0.040540826,0.022102872],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99743915,0.00036414896,0.0002485064,0.0006032121,0.0010328712,0.00031199513],"domain_scores_gemma":[0.9911205,0.00095347245,0.00034201134,0.0028858022,0.0012372775,0.0034609616],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004909185,0.00095745304,0.0013642288,0.0025672123,0.0014046766,0.0060160346,0.0057035084,0.001324425,0.030423088],"category_scores_gemma":[0.009962313,0.0011462293,0.00083237764,0.004003977,0.00075277366,0.009378922,0.008776582,0.003255787,0.027878273],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0019962548,0.0003418487,0.003994357,0.00041708656,0.00008870217,0.00050697575,0.00029220092,0.001955733,0.010856172,0.012412637,0.8077497,0.1593883],"study_design_scores_gemma":[0.0006264533,0.00022749692,0.006735798,0.00023802099,0.00009294487,0.00053638354,0.00032532576,0.01756733,0.011435343,0.017105168,0.9449179,0.0001917566],"about_ca_topic_score_codex":0.0048018415,"about_ca_topic_score_gemma":0.003639561,"teacher_disagreement_score":0.030423088,"about_ca_system_score_codex":0.0014273104,"about_ca_system_score_gemma":0.00298743,"threshold_uncertainty_score":0.10177535},"labels":[],"label_agreement":null},{"id":"W2131767696","doi":"10.1093/database/bar056","title":"GrameneMart: the BioMart data portal for the Gramene project","year":2012,"lang":"en","type":"article","venue":"Database","topic":"Genetic Mapping and Diversity in Plants and Animals","field":"Biochemistry, Genetics and Molecular Biology","cited_by":23,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Ontario Institute for Cancer Research; European Bioinformatics Institute; Wellcome Trust; National Science Foundation","keywords":"Computer science; Software; Genome; Component (thermodynamics); Resource (disambiguation); World Wide Web; Information retrieval; Database; Biology; Gene; Genetics","score_opus":0.09068718542282403,"score_gpt":0.32453579311027186,"score_spread":0.23384860768744783,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2131767696","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0021216744,0.0011911137,0.020323623,0.0005125885,0.00021956197,0.00014032908,0.8538643,0.09922444,0.022402305],"genre_scores_gemma":[0.0053342683,0.00087419606,0.022790713,0.00024931153,0.000054041102,0.0003284176,0.9428057,0.02327153,0.004291813],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9985417,0.0003273186,0.0002044272,0.00029155833,0.00046074795,0.00017425073],"domain_scores_gemma":[0.9979011,0.0005508321,0.00025116382,0.0006311693,0.0003725796,0.00029320907],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0020649796,0.0025747593,0.0025813647,0.006214486,0.0012669662,0.0049974853,0.0042945407,0.0013890305,0.12580977],"category_scores_gemma":[0.006752735,0.0012597734,0.0012617866,0.010943262,0.0004648587,0.0039954954,0.004712271,0.002901728,0.15286371],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0006538555,0.00008044919,0.0009284656,0.0013952821,0.00013074453,0.0001963073,0.00017640054,0.00045431103,0.003249615,0.0056410045,0.95466846,0.03242517],"study_design_scores_gemma":[0.00015177882,0.000029374321,0.002759186,0.00035972687,0.00008462665,0.00025643554,0.00008560068,0.0021043175,0.004756422,0.0086794365,0.98062354,0.00010957425],"about_ca_topic_score_codex":0.004278901,"about_ca_topic_score_gemma":0.004293368,"teacher_disagreement_score":0.12580977,"about_ca_system_score_codex":0.00079717,"about_ca_system_score_gemma":0.0018307964,"threshold_uncertainty_score":0.4208756},"labels":[],"label_agreement":null},{"id":"W2135504399","doi":"10.1093/database/bas006","title":"Aptamer base: a collaborative knowledge base to describe aptamers and SELEX experiments","year":2012,"lang":"en","type":"article","venue":"Database","topic":"Advanced biosensing and bioanalysis techniques","field":"Biochemistry, Genetics and Molecular Biology","cited_by":93,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Aptamer; Systematic evolution of ligands by exponential enrichment; Computer science; Knowledge base; Molecular recognition; SELEX Aptamer Technique; Computational biology; Nucleic acid; Resource (disambiguation); Chemistry; Biology; World Wide Web; RNA; Molecular biology; Biochemistry; Gene; Molecule","score_opus":0.023836741096483142,"score_gpt":0.32388411801932165,"score_spread":0.3000473769228385,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2135504399","genre_codex":"methods","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.004497092,0.0064985,0.55930305,0.0015848139,0.0005264638,0.0014962599,0.30780998,0.09913089,0.019152861],"genre_scores_gemma":[0.0141374525,0.008219222,0.50304854,0.0008157849,0.00012207028,0.002183927,0.4585862,0.0068915864,0.005995148],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99776244,0.00033047254,0.00067558506,0.00050958426,0.000603233,0.0001186343],"domain_scores_gemma":[0.9936133,0.0029417262,0.000768038,0.0014719963,0.0008705567,0.00033438732],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036511384,0.004323056,0.0040492755,0.0124497,0.0015256577,0.007342673,0.0076821274,0.003908442,0.031353272],"category_scores_gemma":[0.015181533,0.0021376396,0.0022564898,0.012754197,0.00078640744,0.0055972347,0.004765983,0.0032177488,0.030058663],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0013445119,0.0008039154,0.0033626857,0.016805526,0.0012475016,0.0026828044,0.00088220276,0.048211586,0.021347504,0.0416196,0.36132506,0.50036716],"study_design_scores_gemma":[0.0003269978,0.00013934048,0.0011594831,0.0016468808,0.0005579607,0.0019194833,0.00020619364,0.039956786,0.014729996,0.044543926,0.89451337,0.0002995457],"about_ca_topic_score_codex":0.0038420002,"about_ca_topic_score_gemma":0.004684883,"teacher_disagreement_score":0.031353272,"about_ca_system_score_codex":0.0015828242,"about_ca_system_score_gemma":0.0045276755,"threshold_uncertainty_score":0.10488713},"labels":[],"label_agreement":null},{"id":"W2135548107","doi":"10.1093/database/baq026","title":"Literature curation of protein interactions: measuring agreement across major public databases","year":2010,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":65,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Hospital for Sick Children","funders":"Canadian Institutes of Health Research; Hospital for Sick Children","keywords":"Data curation; Database; Focus (optics); Organism; Computer science; Data science; Information retrieval; Biology; Genetics","score_opus":0.02698590957156888,"score_gpt":0.2828197224518299,"score_spread":0.255833812880261,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2135548107","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8328838,0.014547006,0.07482458,0.0017424115,0.00025515398,0.0014046308,0.058297664,0.0032508248,0.012793916],"genre_scores_gemma":[0.8122129,0.0027928709,0.09919943,0.0004353461,0.00018932876,0.0014498375,0.08203059,0.0007003057,0.0009894413],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.8493609,0.04504808,0.043846186,0.016423909,0.043046877,0.0022742082],"domain_scores_gemma":[0.39361504,0.41658768,0.063422695,0.043757476,0.07927584,0.0033413682],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.115507446,0.00089519203,0.0024743772,0.056147404,0.0025028929,0.0066634063,0.003085418,0.0015689961,0.0018614482],"category_scores_gemma":[0.34362793,0.00066859554,0.0018249516,0.044960238,0.0020316946,0.0058972663,0.009493068,0.0011683996,0.00107633],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0029161382,0.0005433309,0.71495324,0.0132495165,0.008126334,0.00067668874,0.010987216,0.009452118,0.017602507,0.0059486376,0.016927434,0.19861686],"study_design_scores_gemma":[0.00033228102,0.00062508555,0.82398117,0.0024281412,0.0037205387,0.0022918514,0.010048938,0.051696368,0.034224812,0.019625533,0.050438005,0.0005873076],"about_ca_topic_score_codex":0.0029174502,"about_ca_topic_score_gemma":0.0037545878,"teacher_disagreement_score":0.8844926,"about_ca_system_score_codex":0.0019267603,"about_ca_system_score_gemma":0.0040673725,"threshold_uncertainty_score":0.61086905},"labels":[],"label_agreement":null},{"id":"W2137417655","doi":"10.1093/database/bap024","title":"Phenol-Explorer: an online comprehensive database on polyphenol contents in foods","year":2010,"lang":"en","type":"article","venue":"Database","topic":"Phytochemicals and Antioxidant Activities","field":"Medicine","cited_by":1414,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Institut National de la Recherche Agronomique; Danone","keywords":"Metabolome; Polyphenol; Food science; Database; Computer science; Biotechnology; Biology; Bioinformatics; Metabolomics; Biochemistry","score_opus":0.09133406312008532,"score_gpt":0.33085742175887245,"score_spread":0.23952335863878713,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2137417655","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0050828387,0.0034894848,0.002022238,0.00018064879,0.000055060944,0.000104125014,0.9779012,0.00788694,0.0032774503],"genre_scores_gemma":[0.010020607,0.0036365439,0.01356264,0.00035592332,0.00007504372,0.00038772047,0.96836823,0.0017791364,0.001814193],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9991786,0.00010701335,0.0001979634,0.00017772711,0.00025012702,0.000088505585],"domain_scores_gemma":[0.9973514,0.001022015,0.00061211607,0.00022174903,0.00045010017,0.00034261466],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00096766383,0.002591402,0.00231313,0.010284316,0.00049191946,0.0025059518,0.0019129261,0.0015797038,0.031226154],"category_scores_gemma":[0.0041823196,0.000704064,0.001278145,0.01195003,0.00027872738,0.0030367898,0.0028469136,0.0013518229,0.021040743],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0036646447,0.0003289322,0.016818993,0.04028424,0.0012894652,0.0015460647,0.0008267675,0.0018117455,0.02492752,0.004254278,0.70795524,0.19629216],"study_design_scores_gemma":[0.00047171352,0.0001642652,0.030514155,0.00152032,0.00040853157,0.0006434107,0.0002574346,0.001381947,0.005576839,0.004184125,0.95466274,0.00021463649],"about_ca_topic_score_codex":0.0029193843,"about_ca_topic_score_gemma":0.0046825157,"teacher_disagreement_score":0.031226154,"about_ca_system_score_codex":0.00062842545,"about_ca_system_score_gemma":0.0015319239,"threshold_uncertainty_score":0.10446191},"labels":[],"label_agreement":null},{"id":"W2138088864","doi":"10.1093/database/bau060","title":"A controlled vocabulary for pathway entities and events","year":2014,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"National Human Genome Research Institute; National Institutes of Health; European Commission; European Bioinformatics Institute","keywords":"Computer science; Readability; Consistency (knowledge bases); Vocabulary; Information retrieval; Controlled vocabulary; World Wide Web; Protocol (science); Database; State (computer science); Programming language; Artificial intelligence; Linguistics","score_opus":0.01190032490260914,"score_gpt":0.25635823974965316,"score_spread":0.24445791484704402,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2138088864","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0035642518,0.00089261605,0.9192972,0.001213357,0.00093051675,0.0022479093,0.044759348,0.010687072,0.016407719],"genre_scores_gemma":[0.04835242,0.0018428749,0.8267744,0.0015212414,0.0005966555,0.005417611,0.10342747,0.0035066402,0.008560718],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9905716,0.001867594,0.0036992014,0.0019086319,0.0016020016,0.00035102968],"domain_scores_gemma":[0.9857248,0.0052800393,0.001574148,0.0036926498,0.0030351295,0.00069329125],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009460259,0.0017432766,0.0018478868,0.008702577,0.0030456348,0.008516918,0.0036277613,0.0027424595,0.011223582],"category_scores_gemma":[0.016498912,0.0014234667,0.0032362232,0.0071583237,0.0033567187,0.012284934,0.004796132,0.00509222,0.007988214],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002546641,0.00011551418,0.001006044,0.0016207618,0.00007884914,0.0006099498,0.0020467394,0.0035563994,0.010328185,0.8717358,0.05236252,0.056284618],"study_design_scores_gemma":[0.0000942851,0.000068568166,0.0004078529,0.00045155192,0.00008098776,0.0006512568,0.0004011137,0.0070856097,0.005009214,0.14131704,0.844343,0.00008955265],"about_ca_topic_score_codex":0.008645179,"about_ca_topic_score_gemma":0.007129062,"teacher_disagreement_score":0.011223582,"about_ca_system_score_codex":0.0034834608,"about_ca_system_score_gemma":0.0099015925,"threshold_uncertainty_score":0.050031185},"labels":[],"label_agreement":null},{"id":"W2138782665","doi":"10.1093/database/bap026","title":"PhosphoGRID: a database of experimentally verified in vivo protein phosphorylation sites from the budding yeast Saccharomyces cerevisiae","year":2010,"lang":"en","type":"article","venue":"Database","topic":"Advanced Proteomics Techniques and Applications","field":"Chemistry","cited_by":101,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Lunenfeld-Tanenbaum Research Institute; University of British Columbia","funders":"Canadian Cancer Society Research Institute; Biotechnology and Biological Sciences Research Council; National Center for Research Resources; Directorate for Biological Sciences; National Institutes of Health; Scottish Funding Council","keywords":"Phosphorylation; Saccharomyces cerevisiae; Protein phosphorylation; Proteome; Biology; Proteomics; Phosphatase; Computational biology; Function (biology); Kinase; Biochemistry; Signal transduction; Yeast; Cell biology; Gene; Protein kinase A","score_opus":0.015524813481317587,"score_gpt":0.2758113290470898,"score_spread":0.26028651556577226,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2138782665","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.018594135,0.0045847744,0.0053253416,0.00012901216,0.00008886897,0.00010814368,0.95543027,0.012145008,0.0035945044],"genre_scores_gemma":[0.011785126,0.0019688548,0.00614327,0.00004716773,0.000011658677,0.00009380388,0.9788728,0.00055532454,0.0005220367],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9996432,0.000038339298,0.000093744646,0.00008733155,0.0000977789,0.000039544393],"domain_scores_gemma":[0.99919814,0.00012905204,0.00020695294,0.00014921451,0.00014534364,0.00017137786],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00047795224,0.0032973676,0.0021637557,0.0068590892,0.000840223,0.0016984936,0.0019517044,0.0013528197,0.01286978],"category_scores_gemma":[0.0017112145,0.00095368206,0.0011739419,0.008421761,0.00029758882,0.0015684576,0.0014498547,0.0014275175,0.012392785],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.008051269,0.00054845033,0.018784571,0.03229386,0.0015422844,0.00511084,0.0010257293,0.011946157,0.17383605,0.009119785,0.61350507,0.12423592],"study_design_scores_gemma":[0.0012265524,0.00035289995,0.042425103,0.00094561325,0.00100701,0.002647295,0.00033106888,0.0052274517,0.04125017,0.007714147,0.8965555,0.00031708018],"about_ca_topic_score_codex":0.006210735,"about_ca_topic_score_gemma":0.009779631,"teacher_disagreement_score":0.01286978,"about_ca_system_score_codex":0.0008325799,"about_ca_system_score_gemma":0.0024758053,"threshold_uncertainty_score":0.043053687},"labels":[],"label_agreement":null},{"id":"W2141158585","doi":"10.1093/database/bar020","title":"Curation of characterized glycoside hydrolases of Fungal origin","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Biofuel production and bioconversion","field":"Engineering","cited_by":105,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Glycoside hydrolase; Enzyme; Ascomycota; Hemicellulose; Biology; Biochemistry; Cell wall; Glycoside; Cellulase; Gene; Database; Cellulose; Botany","score_opus":0.03900584823951406,"score_gpt":0.22278377362014143,"score_spread":0.18377792538062737,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2141158585","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.3371575,0.02645509,0.07964819,0.00049141096,0.00044767678,0.0011519043,0.5240073,0.0075495047,0.023091456],"genre_scores_gemma":[0.14297684,0.009561948,0.14174804,0.00039548118,0.00014901055,0.00096574915,0.69644576,0.0015530409,0.0062040826],"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","domain_scores_codex":[0.9983171,0.00014604419,0.00024915865,0.00045346245,0.0006362319,0.0001979182],"domain_scores_gemma":[0.9968453,0.00078744517,0.00059800723,0.0005702455,0.0009261199,0.00027288028],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013118921,0.0019622273,0.002522608,0.010071586,0.0016572075,0.0019233641,0.0014270928,0.0009751181,0.005435392],"category_scores_gemma":[0.0042435634,0.0004712011,0.001499555,0.009553983,0.00052411377,0.0010772583,0.0015773651,0.001229041,0.0059838374],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010927526,0.0003359019,0.018054897,0.010289294,0.000446831,0.0045327907,0.0008717017,0.0015906903,0.80918103,0.002184602,0.032154094,0.119265504],"study_design_scores_gemma":[0.0001835527,0.00041392725,0.11250914,0.0016397829,0.0015758243,0.0077687902,0.0005858117,0.0073366025,0.2928841,0.0020020092,0.572932,0.00016845332],"about_ca_topic_score_codex":0.0033322321,"about_ca_topic_score_gemma":0.0060206954,"teacher_disagreement_score":0.010071586,"about_ca_system_score_codex":0.0007720492,"about_ca_system_score_gemma":0.003830576,"threshold_uncertainty_score":0.018183172},"labels":[],"label_agreement":null},{"id":"W2146436858","doi":"10.1093/database/bau097","title":"MetaProx: the database of metagenomic proximons","year":2014,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"University of Waterloo; Natural Sciences and Engineering Research Council of Canada; Wilfrid Laurier University","keywords":"Metagenomics; Computer science; Database; Inference; Operon; Information retrieval; Data mining; Computational biology; Biology; Gene; Genetics; Artificial intelligence","score_opus":0.0111480008181741,"score_gpt":0.23363944980633577,"score_spread":0.22249144898816167,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2146436858","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012101694,0.003603148,0.09559548,0.0007611087,0.00022172592,0.00037705756,0.7596766,0.10821281,0.019450368],"genre_scores_gemma":[0.022448575,0.0023492908,0.06471636,0.0002962448,0.00007203811,0.00041073526,0.8991362,0.0067055877,0.0038648797],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9989857,0.00014222,0.00014972028,0.00030246054,0.0003083528,0.0001115576],"domain_scores_gemma":[0.9980375,0.0003381052,0.00034477518,0.00073688873,0.00023986326,0.00030284308],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015851936,0.0017809286,0.0022686557,0.005852313,0.00092015625,0.0050660083,0.0041204565,0.0016965072,0.028142955],"category_scores_gemma":[0.005530332,0.0016098522,0.0011399682,0.008920093,0.00062679633,0.0055054715,0.0034895227,0.0018392868,0.021185737],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0037020585,0.00029416117,0.01106084,0.007524273,0.0005282748,0.0014802543,0.0012641897,0.008311266,0.053437363,0.076004714,0.6234851,0.21290751],"study_design_scores_gemma":[0.0004876656,0.00015769295,0.0070359693,0.00056213455,0.00018203165,0.00078245497,0.0004022689,0.011291797,0.020992454,0.03838907,0.91952384,0.00019258984],"about_ca_topic_score_codex":0.0034637034,"about_ca_topic_score_gemma":0.0028610837,"teacher_disagreement_score":0.028142955,"about_ca_system_score_codex":0.0012479216,"about_ca_system_score_gemma":0.0027193553,"threshold_uncertainty_score":0.09414756},"labels":[],"label_agreement":null},{"id":"W2147119862","doi":"10.1093/database/bar031","title":"The Reactome BioMart","year":2011,"lang":"pt","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":38,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"National Human Genome Research Institute; National Institutes of Health; European Commission","keywords":"Computer science; Suite; Interface (matter); Information retrieval; User interface; World Wide Web; Simple (philosophy); Database; Programming language","score_opus":0.026211106589249227,"score_gpt":0.24010133970795178,"score_spread":0.21389023311870256,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2147119862","genre_codex":"software","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":"software","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008177813,0.013385786,0.25024238,0.0016843441,0.0012622129,0.0012047347,0.33465895,0.34755728,0.04182643],"genre_scores_gemma":[0.033468325,0.009059247,0.23672187,0.002052923,0.00033338837,0.0057950933,0.647151,0.039157197,0.026260953],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99665487,0.000970149,0.00027130457,0.0008387413,0.0010562405,0.00020868286],"domain_scores_gemma":[0.99825305,0.0005906896,0.000243207,0.0004437547,0.00030420918,0.00016512664],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004256066,0.0041919285,0.0040787947,0.0052693333,0.00119749,0.003904779,0.005017523,0.0015646026,0.042017788],"category_scores_gemma":[0.0056336247,0.0024339154,0.003395502,0.0034489792,0.0006459518,0.0021089872,0.004470385,0.003379388,0.05427497],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0031632702,0.00025529682,0.0046234634,0.012517976,0.0025225047,0.0009567584,0.00042863502,0.0102200275,0.059176955,0.034028735,0.7665,0.105606504],"study_design_scores_gemma":[0.0004708806,0.0001668172,0.002627254,0.00064086035,0.00077763683,0.00096212234,0.00007070723,0.017071554,0.05164491,0.018258436,0.9070751,0.00023366934],"about_ca_topic_score_codex":0.0019863292,"about_ca_topic_score_gemma":0.0019947202,"teacher_disagreement_score":0.042017788,"about_ca_system_score_codex":0.00090818247,"about_ca_system_score_gemma":0.002663227,"threshold_uncertainty_score":0.14056349},"labels":[],"label_agreement":null},{"id":"W2148200511","doi":"10.1093/database/bar038","title":"BioMart: a data federation framework for large collaborative projects","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":93,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"","keywords":"Computer science; Server; Annotation; Software; World Wide Web; Selection (genetic algorithm); Database; Operating system","score_opus":0.1032471517953839,"score_gpt":0.34675811195317946,"score_spread":0.24351096015779555,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2148200511","genre_codex":"methods","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0012507862,0.0005506297,0.8493987,0.00088191323,0.0003324737,0.000754635,0.008071653,0.13435237,0.0044068964],"genre_scores_gemma":[0.017756293,0.00091382745,0.9095608,0.00050947507,0.0001957473,0.002489799,0.049480263,0.015371389,0.003722417],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9854634,0.0050416896,0.0028435618,0.0019012328,0.0037028391,0.0010473867],"domain_scores_gemma":[0.97813874,0.005120392,0.0012754564,0.010226476,0.002268934,0.0029700096],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02965756,0.0022298256,0.0030951193,0.006989599,0.0037709668,0.011568882,0.010252211,0.003108311,0.013181375],"category_scores_gemma":[0.034625802,0.0030084262,0.0053033507,0.010938808,0.0019831548,0.012728687,0.018686257,0.0046475246,0.014774028],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0017253333,0.00052819616,0.004633475,0.0023048823,0.001014347,0.0011484076,0.0023454542,0.026064225,0.0069227335,0.2027842,0.49985757,0.25067124],"study_design_scores_gemma":[0.00061051187,0.00013112159,0.0016779449,0.00054692384,0.00019913676,0.0008642483,0.00050493696,0.08185571,0.0060851523,0.19439873,0.71272117,0.00040437278],"about_ca_topic_score_codex":0.011121814,"about_ca_topic_score_gemma":0.008748601,"teacher_disagreement_score":0.02965756,"about_ca_system_score_codex":0.002470118,"about_ca_system_score_gemma":0.008004111,"threshold_uncertainty_score":0.15684605},"labels":[],"label_agreement":null},{"id":"W2149332866","doi":"10.1093/database/baq029","title":"Damming the genomic data flood using a comprehensive analysis and storage data structure","year":2010,"lang":"en","type":"article","venue":"Database","topic":"Algorithms and Data Compression","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Montreal Heart Institute; Université de Montréal","funders":"","keywords":"Computer science; Data mining; Data structure; Search engine indexing; Database; Redundancy (engineering); Data management; Data redundancy; Computer data storage; Data set; Data processing; Software; Normalization (sociology); Big data; Table (database); Information retrieval; Operating system","score_opus":0.05690010147651383,"score_gpt":0.3044548854189972,"score_spread":0.24755478394248334,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2149332866","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.047950674,0.0009407764,0.9186536,0.004775042,0.00047778003,0.0004760521,0.003576275,0.018052157,0.0050977375],"genre_scores_gemma":[0.1056184,0.00092688244,0.8785253,0.001000907,0.00025893244,0.00044591242,0.007648743,0.001049454,0.0045253984],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99789006,0.00038390257,0.0002755346,0.00036359022,0.0009958653,0.00009108717],"domain_scores_gemma":[0.98805064,0.0028842564,0.0006784444,0.006319857,0.0017716553,0.00029522867],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004228822,0.00059487845,0.0010895411,0.003040823,0.0011998696,0.0049608466,0.0021708163,0.0010523775,0.0035057983],"category_scores_gemma":[0.014251824,0.000720195,0.0009560676,0.0051384238,0.0012831753,0.005876242,0.0039697615,0.0022973958,0.0028402319],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00079695735,0.00029019773,0.011564689,0.00046421695,0.00018032143,0.0008751236,0.0011538849,0.01961543,0.04019037,0.077301644,0.0736514,0.7739157],"study_design_scores_gemma":[0.0005685475,0.00071831123,0.013671922,0.0004437986,0.00029417817,0.0029677968,0.0010922813,0.34459254,0.14216755,0.19735491,0.29578623,0.00034181125],"about_ca_topic_score_codex":0.0015464915,"about_ca_topic_score_gemma":0.0016881435,"teacher_disagreement_score":0.0049608466,"about_ca_system_score_codex":0.0011227172,"about_ca_system_score_gemma":0.0032714286,"threshold_uncertainty_score":0.022364378},"labels":[],"label_agreement":null},{"id":"W2149671003","doi":"10.1093/database/bar041","title":"BioMart Central Portal: an open database network for the biological community","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":170,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"Ontario Institute for Cancer Research; Cancer Research UK; Wellcome Trust","keywords":"Computer science; Variety (cybernetics); Interface (matter); Ontology; Database; Resource (disambiguation); World Wide Web; Biological database; Bioinformatics; Artificial intelligence; Computer network","score_opus":0.09830288862401737,"score_gpt":0.3035580759883937,"score_spread":0.2052551873643763,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2149671003","genre_codex":"software","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":"software","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005134425,0.0067657623,0.3548145,0.003661384,0.0013340624,0.0009955061,0.11394837,0.45133704,0.06200897],"genre_scores_gemma":[0.043609064,0.008754971,0.3990171,0.0032539463,0.0009909755,0.0028768026,0.43961662,0.05315271,0.048727803],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9964644,0.0009719475,0.00047283634,0.0005974977,0.0011383469,0.00035494193],"domain_scores_gemma":[0.9894946,0.002237386,0.0009870398,0.0029389304,0.001732246,0.0026097728],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.007771629,0.0016742822,0.002842207,0.0066612377,0.0022964736,0.011870843,0.0055378084,0.0028510643,0.058415104],"category_scores_gemma":[0.018271742,0.0015005671,0.0009823425,0.01186415,0.0012080122,0.01102229,0.008827995,0.0038341752,0.07621755],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011910698,0.00014826779,0.001454628,0.0012808016,0.00014750591,0.00032605618,0.00036359942,0.00074915076,0.006367675,0.027786786,0.8186737,0.14151067],"study_design_scores_gemma":[0.00030924103,0.00009382606,0.0014564523,0.0003143806,0.00009255801,0.00066175323,0.00019199711,0.008532042,0.005061896,0.029812751,0.9532962,0.000176955],"about_ca_topic_score_codex":0.003721316,"about_ca_topic_score_gemma":0.0031321696,"teacher_disagreement_score":0.9944622,"about_ca_system_score_codex":0.0014851437,"about_ca_system_score_gemma":0.005543728,"threshold_uncertainty_score":0.195418},"labels":[],"label_agreement":null},{"id":"W2151991601","doi":"10.1093/database/bau095","title":"WholeCellSimDB: a hybrid relational/HDF database for whole-cell model predictions","year":2014,"lang":"en","type":"article","venue":"Database","topic":"Gene Regulatory Network Analysis","field":"Biochemistry, Genetics and Molecular Biology","cited_by":43,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Prince Edward Island","funders":"National Center for Chronic Disease Prevention and Health Promotion; U.S. National Library of Medicine; National Institute of General Medical Sciences; National Institutes of Health","keywords":"Computer science; Python (programming language); JSON; Metadata; JavaScript; Database; Interface (matter); Relational database; Graphical user interface; Interoperability; SQL; Relational database management system; Information retrieval; World Wide Web; Programming language","score_opus":0.015070836204794637,"score_gpt":0.23656580453771275,"score_spread":0.22149496833291812,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2151991601","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0064573265,0.0012338042,0.32061207,0.0005848356,0.00026568724,0.00028393616,0.4355426,0.2190854,0.015934398],"genre_scores_gemma":[0.0323242,0.0013378477,0.14368542,0.00039513264,0.00006888282,0.0009738806,0.7918278,0.023262203,0.006124611],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9989324,0.00011704948,0.00019844061,0.00024318855,0.00041050318,0.0000984577],"domain_scores_gemma":[0.9975074,0.0006635917,0.00017360489,0.0007773874,0.00066637836,0.00021166625],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0022352652,0.0021079388,0.001837841,0.0037602729,0.0011787079,0.0042524603,0.0070389323,0.0018615243,0.051348403],"category_scores_gemma":[0.006913113,0.0014392924,0.0022718345,0.0039593675,0.00043465203,0.003948003,0.0027013277,0.0017588272,0.031471413],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0006875378,0.0002521817,0.005319727,0.002635956,0.00047468342,0.00075971335,0.00046188867,0.0512798,0.016817942,0.042390767,0.7760544,0.102865316],"study_design_scores_gemma":[0.0004517836,0.000093016584,0.0031775383,0.00032575167,0.00022913667,0.00047625203,0.0002975454,0.19720712,0.030241141,0.039110497,0.7281273,0.0002629341],"about_ca_topic_score_codex":0.014559901,"about_ca_topic_score_gemma":0.011818366,"teacher_disagreement_score":0.051348403,"about_ca_system_score_codex":0.0017073252,"about_ca_system_score_gemma":0.0032071597,"threshold_uncertainty_score":0.17177755},"labels":[],"label_agreement":null},{"id":"W2156564520","doi":"10.1093/database/bar030","title":"Ensembl BioMarts: a hub for data retrieval across taxonomic space","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1542,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Biotechnology and Biological Sciences Research Council; Bill and Melinda Gates Foundation; European Commission; Directorate for Biological Sciences; Ontario Institute for Cancer Research; Wellcome Trust","keywords":"Ensembl; Computer science; Genome; World Wide Web; Annotation; Data science; Biology; Genomics; Artificial intelligence; Gene; Genetics","score_opus":0.10775340630988058,"score_gpt":0.3091795235511951,"score_spread":0.20142611724131454,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2156564520","genre_codex":"software","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":"software","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0052563488,0.005740534,0.19929694,0.0012576281,0.0006652545,0.00047209449,0.2717581,0.49967518,0.01587791],"genre_scores_gemma":[0.020278778,0.0029751232,0.26452625,0.001184547,0.00031244927,0.001079064,0.66124016,0.04035429,0.00804934],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99539363,0.0010354117,0.0008376871,0.0010347972,0.0013573903,0.00034108292],"domain_scores_gemma":[0.99213433,0.002791104,0.000902867,0.0021066498,0.001230752,0.0008343138],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007221583,0.004115785,0.0037790972,0.012958951,0.0024595954,0.0068066902,0.0053895675,0.002393831,0.03808897],"category_scores_gemma":[0.016425366,0.0027321994,0.002942467,0.012932799,0.0010784694,0.00700914,0.0073290756,0.0033038463,0.05655116],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0016693779,0.00018970374,0.0034328788,0.004200155,0.00071300013,0.0008389077,0.0008252222,0.002459156,0.014669704,0.014510548,0.7919351,0.16455626],"study_design_scores_gemma":[0.0004989275,0.00020525759,0.00841735,0.00089300156,0.00033776576,0.0016209927,0.00036709488,0.02060308,0.026477115,0.03303046,0.9070424,0.00050655054],"about_ca_topic_score_codex":0.006717142,"about_ca_topic_score_gemma":0.005610651,"teacher_disagreement_score":0.03808897,"about_ca_system_score_codex":0.0014388415,"about_ca_system_score_gemma":0.0041730967,"threshold_uncertainty_score":0.1274203},"labels":[],"label_agreement":null},{"id":"W2158197251","doi":"10.1093/database/baq023","title":"iRefWeb: interactive analysis of consolidated protein interaction data and their supporting evidence","year":2010,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":223,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada Research Chairs; SickKids Foundation; Hospital for Sick Children; University of Toronto","funders":"Canadian Institutes of Health Research; Hospital for Sick Children","keywords":"Computer science; Interface (matter); Interaction information; World Wide Web; Reliability (semiconductor); Database; Information retrieval","score_opus":0.026028351551974787,"score_gpt":0.3164998662923536,"score_spread":0.2904715147403788,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2158197251","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01749026,0.004095854,0.32319453,0.0016681966,0.0004180349,0.00089900807,0.3613145,0.2808303,0.010089273],"genre_scores_gemma":[0.05571553,0.002489782,0.5955222,0.0004222862,0.00028741185,0.0022820674,0.31762642,0.021758338,0.0038958983],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99653876,0.00077170733,0.00054147595,0.0007560266,0.0012106034,0.00018137977],"domain_scores_gemma":[0.97281265,0.019096153,0.0019531132,0.0027275097,0.0027143045,0.0006961852],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01166786,0.003057122,0.0022786723,0.02897572,0.0011139897,0.006749505,0.0036731584,0.0013167679,0.054533605],"category_scores_gemma":[0.042807166,0.0015600073,0.0027186545,0.01664579,0.00062287576,0.00556478,0.0058605387,0.0021672929,0.013190465],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0018698645,0.00032617484,0.028436404,0.01040083,0.0034920177,0.002175142,0.0014482503,0.0066081127,0.011441689,0.015217459,0.50713444,0.4114496],"study_design_scores_gemma":[0.0009859995,0.00025634494,0.044337902,0.003264862,0.0019039314,0.0036101185,0.0012287343,0.122824565,0.022917125,0.10822224,0.68979454,0.0006536784],"about_ca_topic_score_codex":0.0032044475,"about_ca_topic_score_gemma":0.00638058,"teacher_disagreement_score":0.054533605,"about_ca_system_score_codex":0.0010017129,"about_ca_system_score_gemma":0.0023925025,"threshold_uncertainty_score":0.18243307},"labels":[],"label_agreement":null},{"id":"W2166240318","doi":"10.1093/database/bas056","title":"An overview of the BioCreative 2012 Workshop Track III: interactive text mining task","year":2013,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":75,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute for Research in Immunology and Cancer","funders":"U.S. National Library of Medicine; National Human Genome Research Institute","keywords":"Computer science; Data curation; Annotation; Task (project management); Usability; Information retrieval; Data science; Biomedical text mining; Set (abstract data type); World Wide Web; Text mining; Natural language processing; Artificial intelligence; Human–computer interaction; Engineering","score_opus":0.0459006993030045,"score_gpt":0.3375506222957032,"score_spread":0.2916499229926987,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2166240318","genre_codex":"methods","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.06780125,0.014673149,0.44403687,0.011951915,0.00519114,0.016415225,0.18228161,0.2028161,0.054832704],"genre_scores_gemma":[0.056534875,0.0030029763,0.49981785,0.00364145,0.0006989037,0.016897302,0.34972548,0.014385231,0.055296],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.98964816,0.0026739116,0.0012422174,0.0030248591,0.002667575,0.00074338075],"domain_scores_gemma":[0.9842606,0.0060615833,0.00056260946,0.0026217138,0.005053025,0.0014404637],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014722865,0.0039619636,0.002962233,0.0045310957,0.0047710366,0.005669924,0.0065414086,0.0037722278,0.028855162],"category_scores_gemma":[0.022547869,0.0016435088,0.0032604374,0.0056257234,0.00094138685,0.006200093,0.0062139765,0.0036581913,0.03287193],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012504102,0.0011072248,0.0032956034,0.0041033425,0.00019996594,0.0004895073,0.0029159805,0.0019793215,0.026182737,0.0017668818,0.65146387,0.30524513],"study_design_scores_gemma":[0.0008209833,0.0011639323,0.014740254,0.0008729506,0.00028852595,0.0009937978,0.0019251445,0.031373702,0.055210017,0.0068014152,0.8853098,0.00049954135],"about_ca_topic_score_codex":0.0151302805,"about_ca_topic_score_gemma":0.018797882,"teacher_disagreement_score":0.028855162,"about_ca_system_score_codex":0.0032270534,"about_ca_system_score_gemma":0.0059923353,"threshold_uncertainty_score":0.0965302},"labels":[],"label_agreement":null},{"id":"W2166516661","doi":"10.1093/database/bas017","title":"How to link ontologies and protein-protein interactions to literature: text-mining approaches and the BioCreative experience","year":2012,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":34,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute for Research in Immunology and Cancer","funders":"Biotechnology and Biological Sciences Research Council; National Center for Research Resources; Directorate for Biological Sciences; National Institutes of Health; Office of Research Infrastructure Programs, National Institutes of Health; Canadian Institutes of Health Research; European Commission; Wellcome Trust","keywords":"Computer science; Ontology; Information retrieval; Pipeline (software); Context (archaeology); Data curation; Information extraction; Consistency (knowledge bases); Workflow; Open Biomedical Ontologies; Biomedical text mining; Controlled vocabulary; Process (computing); Annotation; Vocabulary; Natural language processing; Upper ontology; Data science; Suggested Upper Merged Ontology; Artificial intelligence; Text mining; Semantic Web; Database","score_opus":0.045486472278633866,"score_gpt":0.294178056247657,"score_spread":0.24869158396902316,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2166516661","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.026268872,0.06489862,0.78902006,0.06287145,0.0015492897,0.0010708706,0.0070309797,0.0031711725,0.04411873],"genre_scores_gemma":[0.07798959,0.033353418,0.8700514,0.004443796,0.0011903633,0.0007403518,0.004905905,0.00081325544,0.0065120743],"study_design_codex":"design_other","study_design_gemma":"design_other","domain_scores_codex":[0.98950773,0.0049420986,0.0018088757,0.0010784836,0.002459156,0.00020361955],"domain_scores_gemma":[0.94926924,0.037883665,0.0034090162,0.0029713474,0.005687299,0.0007794268],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.017994752,0.0010653986,0.0013331637,0.047663927,0.0023742826,0.012243057,0.0025593732,0.0022079255,0.005392216],"category_scores_gemma":[0.055367894,0.00077948754,0.0014816448,0.027758587,0.004947551,0.022854716,0.004310654,0.0022886838,0.002775411],"study_design_candidate":"design_other","study_design_consensus":"design_other","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00007429823,0.00016516577,0.0044320063,0.006444501,0.00025647428,0.0017093465,0.0153052155,0.0017301318,0.00718023,0.16937564,0.042595156,0.75073177],"study_design_scores_gemma":[0.00003540733,0.000042889515,0.0045387023,0.0064894166,0.00017519502,0.0019417213,0.012165451,0.010038312,0.0045448914,0.39960223,0.5602973,0.00012850098],"about_ca_topic_score_codex":0.0034099533,"about_ca_topic_score_gemma":0.005008338,"teacher_disagreement_score":0.047663927,"about_ca_system_score_codex":0.0023880065,"about_ca_system_score_gemma":0.0030158756,"threshold_uncertainty_score":0.095166445},"labels":[],"label_agreement":null},{"id":"W2168670121","doi":"10.1093/database/bar051","title":"The Chado Natural Diversity module: a new generic database schema for large-scale phenotyping and genotyping data","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":39,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"Division of Emerging Frontiers; National Institute of Allergy and Infectious Diseases; Division of Biological Infrastructure; National Institute of General Medical Sciences; Smithsonian Tropical Research Institute; National Human Genome Research Institute; Agricultural Research Service; National Institutes of Health; National Institute of Food and Agriculture; Smithsonian Institution; U.S. Department of Agriculture; Commonwealth Scientific and Industrial Research Organisation; National Evolutionary Synthesis Center; Agriculture and Agri-Food Canada; National Science Foundation","keywords":"Computer science; Modular design; Genotyping; Biological database; Relational database; Schema (genetic algorithms); Database; Information retrieval; Data mining; Bioinformatics; Biology; Programming language; Genotype","score_opus":0.05710983984248658,"score_gpt":0.25765207301269766,"score_spread":0.20054223317021108,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2168670121","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007985,0.0007591155,0.9266226,0.00095624686,0.00019222323,0.0010885785,0.028844627,0.026551547,0.0070000356],"genre_scores_gemma":[0.05529144,0.0012038187,0.80890864,0.0023891528,0.00013523691,0.0018791218,0.118025854,0.0036407916,0.008525987],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99644744,0.00060490717,0.0009891866,0.0007541241,0.0010106935,0.0001935652],"domain_scores_gemma":[0.9909739,0.0017577008,0.0008260056,0.00365918,0.0022545513,0.00052859157],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011887159,0.00073697616,0.0011287035,0.0032797174,0.00090825645,0.0061959047,0.0048909117,0.0016871674,0.004664177],"category_scores_gemma":[0.010612523,0.0010236182,0.0014889084,0.0032765546,0.0009837567,0.0061689247,0.0040969746,0.0019516151,0.0035708365],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012905143,0.0006293862,0.027145738,0.002662511,0.000777238,0.000958005,0.003058683,0.01488838,0.05144653,0.2531271,0.25036237,0.39365357],"study_design_scores_gemma":[0.0002026807,0.00017962407,0.0061864317,0.00032153862,0.00018118384,0.001518865,0.00045683954,0.027841333,0.01818746,0.028928611,0.9157676,0.00022782745],"about_ca_topic_score_codex":0.0064459736,"about_ca_topic_score_gemma":0.0059083044,"teacher_disagreement_score":0.011887159,"about_ca_system_score_codex":0.0020647105,"about_ca_system_score_gemma":0.003914699,"threshold_uncertainty_score":0.06286609},"labels":[],"label_agreement":null},{"id":"W2169425537","doi":"10.1093/database/bau067","title":"Assisting manual literature curation for protein-protein interactions using BioQRator","year":2014,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":37,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Institute for Research in Immunology and Cancer","funders":"U.S. National Library of Medicine; National Institutes of Health; Ministry of Education, Science and Technology; National Research Foundation of Korea; National Research Foundation","keywords":"Annotation; Computer science; Data curation; Task (project management); Information retrieval; World Wide Web; Artificial intelligence","score_opus":0.030990985643682793,"score_gpt":0.3347915863824545,"score_spread":0.3038006007387717,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2169425537","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.034790702,0.019985508,0.58021075,0.0080769975,0.0023549155,0.009119655,0.12792274,0.19128886,0.026249906],"genre_scores_gemma":[0.025685517,0.004225966,0.88085234,0.0012767988,0.00037944678,0.0037323246,0.070476376,0.005985632,0.0073855687],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.97462374,0.011908915,0.0060562347,0.0029674436,0.0039109187,0.00053278625],"domain_scores_gemma":[0.80969924,0.11087504,0.014512337,0.023002647,0.039082434,0.0028283452],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.046210412,0.0022792318,0.0036799735,0.036880523,0.0027725592,0.006884377,0.0033423572,0.001456599,0.035865944],"category_scores_gemma":[0.122582205,0.0012658915,0.0026639237,0.01974406,0.0010385811,0.0060399566,0.008058565,0.0019429853,0.034864534],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00091344194,0.0003433437,0.007555298,0.032367304,0.0007644474,0.0012741798,0.0051932298,0.00083119335,0.04489646,0.0040295594,0.2701219,0.6317097],"study_design_scores_gemma":[0.00040732318,0.0003256063,0.014505694,0.005728619,0.0010317122,0.0023271898,0.002732445,0.009818451,0.03924037,0.0077654556,0.91564876,0.0004683459],"about_ca_topic_score_codex":0.0036915971,"about_ca_topic_score_gemma":0.009801192,"teacher_disagreement_score":0.046210412,"about_ca_system_score_codex":0.001607876,"about_ca_system_score_gemma":0.0110325655,"threshold_uncertainty_score":0.24438691},"labels":[],"label_agreement":null},{"id":"W2170283197","doi":"10.1093/database/bat026","title":"The PhosphoGRID Saccharomyces cerevisiae protein phosphorylation site database: version 2.0 update","year":2013,"lang":"en","type":"article","venue":"Database","topic":"Fungal and yeast genetics research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":108,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Lunenfeld-Tanenbaum Research Institute; Université de Montréal; Institute for Research in Immunology and Cancer; University of British Columbia","funders":"Biotechnology and Biological Sciences Research Council; Canadian Cancer Society Research Institute; Canadian Institutes of Health Research; National Center for Research Resources; Directorate for Biological Sciences; National Institutes of Health; Wellcome Trust","keywords":"Phosphorylation; Proteome; Saccharomyces cerevisiae; Computational biology; Protein phosphorylation; Protein function; Database; Function (biology); Biology; Yeast; Bioinformatics; Computer science; Biochemistry; Protein kinase A; Cell biology; Gene","score_opus":0.007943784022554383,"score_gpt":0.24274444773718434,"score_spread":0.23480066371462996,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2170283197","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005712513,0.003202908,0.0061249365,0.00026157883,0.00021371298,0.000110853696,0.9566182,0.02333954,0.0044157347],"genre_scores_gemma":[0.0033387013,0.0010956223,0.005721063,0.000079924816,0.000021229978,0.00010398128,0.9874408,0.0014536997,0.00074490235],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99945956,0.000076059085,0.000103712475,0.00011963297,0.00017067169,0.00007043431],"domain_scores_gemma":[0.9987502,0.00016373556,0.00018515135,0.00033301767,0.00033154158,0.00023625819],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013825031,0.003367448,0.0028132454,0.006057416,0.00092045567,0.0031147862,0.003232062,0.0011514012,0.013882363],"category_scores_gemma":[0.0023117936,0.001594007,0.0013323753,0.00928033,0.00030123748,0.0019621544,0.0020454738,0.00177582,0.024347272],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0024717965,0.00015737786,0.0050191255,0.0065362984,0.00052574684,0.0004421557,0.000166391,0.003577404,0.016171386,0.003933766,0.89466643,0.066332184],"study_design_scores_gemma":[0.00047461208,0.0000744069,0.008826509,0.00037065477,0.00035265533,0.0005605058,0.00007374216,0.0033104983,0.007156311,0.004404395,0.9742604,0.00013535128],"about_ca_topic_score_codex":0.009293183,"about_ca_topic_score_gemma":0.009853871,"teacher_disagreement_score":0.013882363,"about_ca_system_score_codex":0.0011520842,"about_ca_system_score_gemma":0.0029056268,"threshold_uncertainty_score":0.046441138},"labels":[],"label_agreement":null},{"id":"W2170802261","doi":"10.1093/database/bar021","title":"A comprehensive manually curated reaction map of RANKL/RANK-signaling pathway","year":2011,"lang":"en","type":"article","venue":"Database","topic":"Bone Metabolism and Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":51,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"The Wellcome Trust DBT India Alliance; Council of Scientific and Industrial Research, India; Department of Biotechnology, Ministry of Science and Technology, India; Wellcome Trust","keywords":"RANKL; RANK Ligand; Osteoclast; NFAT; Activator (genetics); Cell biology; Biology; Signal transduction; Cancer research; PI3K/AKT/mTOR pathway; Transcription factor; Receptor; Biochemistry","score_opus":0.03295461600734513,"score_gpt":0.2515163771703015,"score_spread":0.2185617611629564,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2170802261","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.023223408,0.016818304,0.09014462,0.0006293689,0.0004213739,0.0009886214,0.82869315,0.024259688,0.01482151],"genre_scores_gemma":[0.05014089,0.007062477,0.13954277,0.00017026317,0.00009342695,0.0012172878,0.7942955,0.0014116626,0.006065648],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99827087,0.00018891854,0.0003254546,0.00048015843,0.0006228847,0.00011165024],"domain_scores_gemma":[0.9963535,0.0010435308,0.00045658206,0.0004712411,0.0015520281,0.00012308038],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012083764,0.0029729952,0.0020610392,0.01728913,0.0012426871,0.0021824734,0.0022552987,0.0014383771,0.01652934],"category_scores_gemma":[0.005847316,0.0006612984,0.0022715027,0.013303574,0.00048852654,0.001069597,0.0011576109,0.0008377108,0.009712543],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0020377354,0.00041590197,0.012044724,0.046583623,0.0013921423,0.0065827975,0.0007533109,0.020010917,0.098981984,0.009191753,0.2674911,0.5345141],"study_design_scores_gemma":[0.00028395528,0.00030487034,0.020499332,0.0020287458,0.0021332302,0.0044071493,0.00029768562,0.028979002,0.07172927,0.006919871,0.86210406,0.00031279022],"about_ca_topic_score_codex":0.008134933,"about_ca_topic_score_gemma":0.012319031,"teacher_disagreement_score":0.01728913,"about_ca_system_score_codex":0.001237014,"about_ca_system_score_gemma":0.0056066886,"threshold_uncertainty_score":0.055296183},"labels":[],"label_agreement":null},{"id":"W2336698700","doi":"10.1093/database/baw013","title":"PhyloPro2.0: a database for the dynamic exploration of phylogenetically conserved proteins and their domain architectures across the Eukarya","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":16,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Hospital for Sick Children","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Hospital for Sick Children; University of Toronto","keywords":"Domain (mathematical analysis); Computer science; Visualization; Upload; Phylogenetic tree; Database; Protein domain; Interface (matter); Adjacency list; Data mining; Information retrieval; World Wide Web; Biology; Genetics","score_opus":0.02104871360495447,"score_gpt":0.26985288197856594,"score_spread":0.24880416837361147,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2336698700","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.026354928,0.0051945825,0.11178284,0.00044888255,0.00022764679,0.0002831361,0.6643229,0.17698887,0.014396245],"genre_scores_gemma":[0.034268267,0.0036226455,0.13187338,0.00029692036,0.00007680565,0.0006618236,0.80833817,0.016758293,0.004103578],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.999529,0.000059934304,0.00006828049,0.00012305955,0.00013975988,0.00007992121],"domain_scores_gemma":[0.99870825,0.00041470118,0.00028150345,0.00021744375,0.0001361141,0.00024212834],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013400555,0.0026989062,0.0018586869,0.009369898,0.0012133566,0.0024533055,0.0025811668,0.0016318978,0.029669452],"category_scores_gemma":[0.0029811775,0.0017198772,0.0015449749,0.008729608,0.00054638245,0.0038189688,0.0037022491,0.0018509384,0.021155214],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00405117,0.0003069043,0.015079266,0.012772615,0.0005282913,0.0023509176,0.0014229999,0.004514026,0.10045225,0.014851909,0.598936,0.24473365],"study_design_scores_gemma":[0.00046681243,0.00022583296,0.020218803,0.00089364935,0.00027052467,0.003315646,0.0004039272,0.009843864,0.026669763,0.0122855175,0.9251072,0.00029850632],"about_ca_topic_score_codex":0.0014289069,"about_ca_topic_score_gemma":0.0017626097,"teacher_disagreement_score":0.029669452,"about_ca_system_score_codex":0.0006642371,"about_ca_system_score_gemma":0.0016059176,"threshold_uncertainty_score":0.09925425},"labels":[],"label_agreement":null},{"id":"W2338762894","doi":"10.1093/database/baw007","title":"2P2Idb v2: update of a structural database dedicated to orthosteric modulation of protein–protein interactions","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":80,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canadian Nautical Research Society","funders":"Agence Nationale de la Recherche","keywords":"Protein Data Bank; Protein Data Bank (RCSB PDB); Database; Protein structure database; Protein–protein interaction; Planarity testing; Computer science; Protein structure; Protein ligand; Affinities; Interface (matter); Small molecule; Computational biology; Chemistry; Data mining; Information retrieval; Crystallography; Biology; Molecule; Stereochemistry; Biochemistry","score_opus":0.009674534010869559,"score_gpt":0.2678753414436307,"score_spread":0.25820080743276114,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2338762894","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04968633,0.04034989,0.079851575,0.0038498826,0.0031997731,0.0013038059,0.7000322,0.09652525,0.02520139],"genre_scores_gemma":[0.013929886,0.009058772,0.049228113,0.0008469111,0.0002951769,0.00070866273,0.9139561,0.007091607,0.0048847534],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99782336,0.00032946368,0.0002875311,0.0003249388,0.0009955871,0.00023916294],"domain_scores_gemma":[0.9976676,0.0003508181,0.0002913095,0.00039900493,0.00072727876,0.00056399073],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0042454596,0.0027267768,0.0048512095,0.004258279,0.0010484513,0.0040182834,0.006240549,0.0021358533,0.020095518],"category_scores_gemma":[0.005866092,0.002357207,0.0018110754,0.0050666183,0.0005233378,0.0033085365,0.0038111561,0.0037451377,0.01925966],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0023118213,0.00034041077,0.002796149,0.0053585013,0.0005783666,0.0006613878,0.0003557971,0.0029251317,0.04666774,0.004135788,0.7584915,0.17537737],"study_design_scores_gemma":[0.00091376656,0.00022988765,0.004000787,0.00043611016,0.00038875378,0.0011934903,0.000054672353,0.0059711323,0.012380641,0.0016348176,0.97261524,0.00018067818],"about_ca_topic_score_codex":0.00598897,"about_ca_topic_score_gemma":0.008076391,"teacher_disagreement_score":0.020095518,"about_ca_system_score_codex":0.0015853461,"about_ca_system_score_gemma":0.005329808,"threshold_uncertainty_score":0.06722623},"labels":[],"label_agreement":null},{"id":"W2418383179","doi":"10.1093/database/baw090","title":"MET network in PubMed: a text-mined network visualization and curation system","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"National Human Genome Research Institute; National Institutes of Health; National Institute of General Medical Sciences; Ministry of Science and Technology, Taiwan; European Molecular Biology Laboratory","keywords":"Computer science; Task (project management); Metastasis; Data curation; Visualization; World Wide Web; Process (computing); Cancer; Information retrieval; Data science; Bioinformatics; Medicine; Artificial intelligence; Biology","score_opus":0.014714083094430574,"score_gpt":0.25834995446435843,"score_spread":0.24363587136992787,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2418383179","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01212881,0.005228307,0.16011953,0.0035687604,0.0005148229,0.0020308534,0.60175794,0.20248851,0.012162426],"genre_scores_gemma":[0.04015243,0.0044976287,0.528719,0.0009896293,0.00023527275,0.003404032,0.4102604,0.0068937363,0.0048478474],"study_design_codex":"not_applicable","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99768925,0.00059635076,0.00068750593,0.00050343375,0.00044859308,0.00007481801],"domain_scores_gemma":[0.9889836,0.00557035,0.0017451401,0.0013983999,0.0018086754,0.0004940036],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0051020375,0.002015415,0.0014780894,0.018527497,0.0013503658,0.0032120934,0.0020354697,0.0014642077,0.019403819],"category_scores_gemma":[0.02020289,0.0009985422,0.0018445336,0.011112493,0.000374226,0.00478007,0.004310636,0.0010561303,0.00903887],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0015957769,0.00021306935,0.011715519,0.02640141,0.0012078785,0.0026137752,0.0028903752,0.0049367803,0.02537403,0.020292187,0.6068251,0.295934],"study_design_scores_gemma":[0.0005258167,0.00026682354,0.010474706,0.002571042,0.0007997217,0.0016142441,0.00083631504,0.037264787,0.017190307,0.021327198,0.90686536,0.00026362843],"about_ca_topic_score_codex":0.004194766,"about_ca_topic_score_gemma":0.009580348,"teacher_disagreement_score":0.99489796,"about_ca_system_score_codex":0.0012491036,"about_ca_system_score_gemma":0.0039942428,"threshold_uncertainty_score":0.0649122},"labels":[],"label_agreement":null},{"id":"W2509885322","doi":"10.1093/database/baw121","title":"BioCreative V BioC track overview: collaborative biocurator assistant task for BioGRID","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":32,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Mount Sinai Hospital; Lunenfeld-Tanenbaum Research Institute; Université de Montréal; Institute for Research in Immunology and Cancer","funders":"National Institute of General Medical Sciences; Biotechnology and Biological Sciences Research Council; National Institutes of Health","keywords":"Computer science; Usability; Annotation; Task (project management); Interoperability; Classifier (UML); World Wide Web; Information retrieval; Data curation; Crowdsourcing; Natural language processing; Artificial intelligence; Human–computer interaction","score_opus":0.025597812753064744,"score_gpt":0.3119697022994703,"score_spread":0.2863718895464056,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2509885322","genre_codex":"software","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.035829023,0.0025977823,0.3038677,0.0034863944,0.0014643279,0.006897336,0.24771187,0.35443047,0.043715134],"genre_scores_gemma":[0.039651774,0.00068088266,0.2936366,0.0009860748,0.00029604818,0.0059092394,0.5991034,0.030356215,0.029379824],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9935921,0.0015126264,0.00072201947,0.0015559075,0.0019049147,0.0007124021],"domain_scores_gemma":[0.9805921,0.0039999904,0.0011671623,0.006180745,0.005391428,0.0026685416],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.017059505,0.0027356187,0.0020838818,0.004400817,0.0033523028,0.0053510265,0.005879368,0.0027017584,0.026871176],"category_scores_gemma":[0.019773025,0.0016887544,0.0020269088,0.005271286,0.00057171716,0.005621374,0.0064990637,0.0022481862,0.035236835],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012464983,0.00049029215,0.004354779,0.0010577758,0.00018219509,0.0002166773,0.0010051377,0.0017593723,0.010960668,0.0021165397,0.87807703,0.098533124],"study_design_scores_gemma":[0.000998131,0.00082263036,0.012419098,0.00024849354,0.00017140096,0.0005523171,0.0007160143,0.02359476,0.027296461,0.00377033,0.92911166,0.00029874418],"about_ca_topic_score_codex":0.032511916,"about_ca_topic_score_gemma":0.031033915,"teacher_disagreement_score":0.032511916,"about_ca_system_score_codex":0.002791755,"about_ca_system_score_gemma":0.007993827,"threshold_uncertainty_score":0.09022039},"labels":[],"label_agreement":null},{"id":"W2516150684","doi":"10.1093/database/baw119","title":"Overview of the interactive task in BioCreative V","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Biomedical and Engineering Education","field":"Engineering","cited_by":99,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Institute for Research in Immunology and Cancer; Ontario Institute for Cancer Research","funders":"U.S. National Library of Medicine; National Human Genome Research Institute; National Institute of General Medical Sciences; European Molecular Biology Laboratory; U.S. Department of Energy; Dirección General de Asuntos del Personal Académico, Universidad Nacional Autónoma de México; National Institutes of Health; National Science Foundation","keywords":"Computer science; Task (project management); Natural language processing; Information retrieval; World Wide Web; Artificial intelligence; Engineering","score_opus":0.017758986228998432,"score_gpt":0.252505487475985,"score_spread":0.23474650124698657,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2516150684","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.018023962,0.0049061584,0.7783513,0.0017305859,0.0012100012,0.007830078,0.01343721,0.09907992,0.07543082],"genre_scores_gemma":[0.05265221,0.0022418685,0.8221543,0.001304593,0.00042654824,0.012279758,0.041690633,0.012313239,0.05493683],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98837227,0.004025914,0.0011999306,0.002405139,0.0030138043,0.0009830134],"domain_scores_gemma":[0.982189,0.0077745635,0.00073473604,0.0037783498,0.0037621714,0.0017611439],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009441046,0.0024675345,0.0015249976,0.0048983674,0.002309529,0.0066521475,0.005558063,0.002770407,0.043556638],"category_scores_gemma":[0.022086682,0.0016672782,0.0026034417,0.0031304518,0.0013797014,0.005323962,0.0074713468,0.0022853133,0.03380821],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0020596057,0.0011627388,0.0056970674,0.0054620095,0.00020412977,0.00069158105,0.007188278,0.003379243,0.029806724,0.018982429,0.19337222,0.731994],"study_design_scores_gemma":[0.00025223423,0.0007354466,0.007240812,0.0010194762,0.000101165635,0.0011845262,0.00094054017,0.014097795,0.017370444,0.008262606,0.9485072,0.00028775877],"about_ca_topic_score_codex":0.006556822,"about_ca_topic_score_gemma":0.0075619486,"teacher_disagreement_score":0.043556638,"about_ca_system_score_codex":0.0019412228,"about_ca_system_score_gemma":0.0052974503,"threshold_uncertainty_score":0.14571142},"labels":[],"label_agreement":null},{"id":"W2568859793","doi":"10.1093/database/baw160","title":"FirebrowseR: an R client to the Broad Institute’s Firehose Pipeline","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":224,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Alberta Children's Hospital; University of Calgary","funders":"","keywords":"Computer science; Workflow; Pipeline (software); Application programming interface; Web service; Software; Database; World Wide Web; Interface (matter); Software engineering; Task (project management); Service (business); Operating system","score_opus":0.014081698061545926,"score_gpt":0.2552758913824826,"score_spread":0.24119419332093664,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2568859793","genre_codex":"software","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":"software","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0017425148,0.00026981853,0.15772207,0.00042548703,0.0001236656,0.00024979372,0.010604071,0.8245672,0.004295442],"genre_scores_gemma":[0.061210625,0.00084018323,0.43949345,0.004061606,0.00024646163,0.0026932748,0.080611765,0.38005355,0.030789137],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9977647,0.00044941835,0.00016451704,0.0006039208,0.00075981644,0.00025770444],"domain_scores_gemma":[0.99604577,0.0016855688,0.000299398,0.0010806298,0.00050210505,0.0003864354],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0048516444,0.0023291626,0.0019220628,0.0022366405,0.0009066966,0.0024465874,0.0050365054,0.001988621,0.050531074],"category_scores_gemma":[0.011313851,0.002238985,0.0026735985,0.0012438281,0.00084950915,0.0025101043,0.004089608,0.003309329,0.059082534],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.002095077,0.00021654452,0.00379586,0.0010687902,0.000532473,0.0013057033,0.00054725324,0.0034179103,0.02798356,0.0064579546,0.8266711,0.12590778],"study_design_scores_gemma":[0.0021691034,0.0003044074,0.011761617,0.00047213014,0.00035501205,0.0024995045,0.0002021525,0.11614759,0.12047716,0.031469278,0.71325815,0.0008838694],"about_ca_topic_score_codex":0.0050193635,"about_ca_topic_score_gemma":0.004901885,"teacher_disagreement_score":0.050531074,"about_ca_system_score_codex":0.0011498645,"about_ca_system_score_gemma":0.0020594245,"threshold_uncertainty_score":0.1690433},"labels":[],"label_agreement":null},{"id":"W2573117738","doi":"10.1093/database/baw147","title":"The BioC-BioGRID corpus: full text articles annotated for curation of protein–protein and genetic interactions","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":29,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Mount Sinai Hospital; Lunenfeld-Tanenbaum Research Institute; Université de Montréal; Institute for Research in Immunology and Cancer","funders":"Biotechnology and Biological Sciences Research Council; National Institutes of Health","keywords":"Data curation; Annotation; Crowdsourcing; Computer science; Task (project management); Information retrieval; Gene Annotation; Pipeline (software); World Wide Web; Data science; Natural language processing; Computational biology; Artificial intelligence; Biology; Gene; Genome; Genetics; Engineering","score_opus":0.019403511412564904,"score_gpt":0.2759297630851905,"score_spread":0.2565262516726256,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2573117738","genre_codex":"dataset","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.02363438,0.013649165,0.007220298,0.0024760917,0.00094611605,0.0010179456,0.9286564,0.0030116583,0.01938801],"genre_scores_gemma":[0.02132499,0.004333302,0.015191063,0.00056925457,0.0002523643,0.0017302745,0.95038193,0.0009158172,0.005300917],"study_design_codex":"not_applicable","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9962913,0.00077279407,0.00084486057,0.0007200956,0.001172252,0.00019869814],"domain_scores_gemma":[0.9774449,0.0147281755,0.001787689,0.0014998424,0.003728505,0.0008109412],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0030985258,0.0016414218,0.0013614661,0.021265294,0.0027849877,0.003135238,0.0019892638,0.002031925,0.02024647],"category_scores_gemma":[0.016202081,0.0006109745,0.0007237725,0.026096946,0.0016896743,0.0020637321,0.0036159789,0.001786069,0.014526825],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00048636826,0.00013312469,0.0026490018,0.020831965,0.00015256519,0.002653658,0.0022938934,0.0007540624,0.014003166,0.0039262385,0.8940946,0.058021266],"study_design_scores_gemma":[0.00014066405,0.000033112345,0.012213608,0.0016603625,0.00013472003,0.0009985316,0.0010919318,0.00073396176,0.004950676,0.0012088843,0.9767725,0.000060958933],"about_ca_topic_score_codex":0.010150199,"about_ca_topic_score_gemma":0.028258102,"teacher_disagreement_score":0.021265294,"about_ca_system_score_codex":0.0024817395,"about_ca_system_score_gemma":0.005512308,"threshold_uncertainty_score":0.0677312},"labels":[],"label_agreement":null},{"id":"W2626120835","doi":"10.1093/database/bax048","title":"PigVar: a database of pig variations and positive selection signatures","year":2017,"lang":"en","type":"article","venue":"Database","topic":"Genetic and phenotypic traits in livestock","field":"Biochemistry, Genetics and Molecular Biology","cited_by":44,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"National Key Research and Development Program of China; National Natural Science Foundation of China","keywords":"Selection (genetic algorithm); Database; Computer science; Information retrieval; Artificial intelligence","score_opus":0.011374282493286357,"score_gpt":0.26502520920620565,"score_spread":0.2536509267129193,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2626120835","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.065734334,0.003070363,0.036874782,0.0002613891,0.00027020625,0.0002305176,0.86627495,0.018510586,0.008772944],"genre_scores_gemma":[0.036097378,0.00079410925,0.032868296,0.0002345818,0.0000875018,0.0003843758,0.92557496,0.0021797803,0.0017789883],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9990753,0.00015172137,0.0001538023,0.000324694,0.00020339267,0.0000911231],"domain_scores_gemma":[0.9986563,0.00043499054,0.00036929583,0.0002627029,0.00011220678,0.00016447164],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015110953,0.0011826386,0.0017169514,0.0051700654,0.0005246826,0.0014197208,0.0018081814,0.001354828,0.01788862],"category_scores_gemma":[0.0028477937,0.0006266808,0.0012550773,0.0057444046,0.0004018678,0.00090798776,0.0015396395,0.001009146,0.008701878],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.005580118,0.0005120948,0.10901073,0.007377133,0.0020534804,0.004413277,0.0010903508,0.006104144,0.09159541,0.008711161,0.38648152,0.37707058],"study_design_scores_gemma":[0.00079947687,0.00062411826,0.16196918,0.0006884197,0.00094891945,0.005554378,0.0001790588,0.0060138935,0.011093309,0.008029027,0.8038398,0.00026035067],"about_ca_topic_score_codex":0.001479037,"about_ca_topic_score_gemma":0.002135455,"teacher_disagreement_score":0.01788862,"about_ca_system_score_codex":0.0003682051,"about_ca_system_score_gemma":0.0009452493,"threshold_uncertainty_score":0.05984342},"labels":[],"label_agreement":null},{"id":"W2786427465","doi":"10.1093/database/bax076","title":"Database of resistance related metabolites in Wheat Fusarium head blight Disease (MWFD)","year":2017,"lang":"en","type":"article","venue":"Database","topic":"Mycotoxins in Agriculture and Food","field":"Agricultural and Biological Sciences","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"National Research Council Canada","funders":"National Research Council Canada","keywords":"Fusarium; Head (geology); Blight; Plant disease resistance; Resistance (ecology); Biology; Database; Botany; Computer science; Agronomy; Genetics; Gene","score_opus":0.026633827463476208,"score_gpt":0.25797562512058536,"score_spread":0.23134179765710916,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2786427465","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012192301,0.01180499,0.0016723376,0.00017323243,0.000058275145,0.00016760777,0.9644588,0.0028888788,0.006583463],"genre_scores_gemma":[0.011283458,0.0066094855,0.0062142108,0.00012944064,0.000023011371,0.00014819777,0.97397137,0.0002005418,0.0014201402],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99947935,0.000059746777,0.00015832957,0.000113473136,0.00013343307,0.000055712855],"domain_scores_gemma":[0.9989993,0.00023528267,0.00028990427,0.000101760146,0.0002202388,0.00015360738],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00063524756,0.002033609,0.0023298352,0.009926839,0.00082459685,0.0021011853,0.0011806571,0.0016272903,0.017728135],"category_scores_gemma":[0.0021412596,0.0004870832,0.0011990669,0.015400038,0.0002926016,0.0016529025,0.0016583384,0.00091017963,0.011483284],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0071833446,0.0006509893,0.019411527,0.0923665,0.0017743234,0.005725014,0.0012849761,0.004154569,0.18810108,0.0074269166,0.41377687,0.25814387],"study_design_scores_gemma":[0.00031668763,0.00042685834,0.031100428,0.0020665955,0.00092192437,0.001699611,0.00019372966,0.00092660787,0.014588124,0.0014587275,0.9461841,0.000116647956],"about_ca_topic_score_codex":0.0037929781,"about_ca_topic_score_gemma":0.0061408794,"teacher_disagreement_score":0.017728135,"about_ca_system_score_codex":0.000975818,"about_ca_system_score_gemma":0.0021871668,"threshold_uncertainty_score":0.059306443},"labels":[],"label_agreement":null},{"id":"W2795044975","doi":"10.1093/database/bay032","title":"Biopanning data bank 2018: hugging next generation phage display","year":2018,"lang":"en","type":"article","venue":"Database","topic":"Bacteriophages and microbial interactions","field":"Environmental Science","cited_by":30,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"National Natural Science Foundation of China","keywords":"Biopanning; Computer science; Data bank; Library science; Telecommunications; Biology","score_opus":0.10181567766062573,"score_gpt":0.31198551436850214,"score_spread":0.21016983670787642,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2795044975","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007956052,0.0026841064,0.019581199,0.00045172835,0.00030306837,0.00036352323,0.9179684,0.042116687,0.008575222],"genre_scores_gemma":[0.004118837,0.0007063127,0.010800758,0.00013404094,0.00002211729,0.0002292526,0.9813934,0.0013660874,0.0012293204],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9982987,0.0003065761,0.00033557316,0.00028179455,0.00059720123,0.00018015013],"domain_scores_gemma":[0.9971877,0.00057854905,0.00039438024,0.0007041082,0.0007537722,0.00038146204],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0034578834,0.0022720583,0.003008435,0.0049001994,0.0010692485,0.0040131398,0.003090115,0.0011543059,0.025245592],"category_scores_gemma":[0.0053500775,0.0013413324,0.0012221436,0.0075065694,0.00030828547,0.0019310298,0.002014042,0.0025714722,0.035075027],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0036907312,0.00041020778,0.006024755,0.004541892,0.0003886911,0.00042403533,0.0001864852,0.0027012795,0.040172175,0.0040354338,0.8579666,0.07945773],"study_design_scores_gemma":[0.00072571833,0.0002419782,0.009336488,0.0003605083,0.00023452233,0.00043232125,0.000092896276,0.008133579,0.028954698,0.003149696,0.948122,0.00021545417],"about_ca_topic_score_codex":0.0037951157,"about_ca_topic_score_gemma":0.0026835918,"teacher_disagreement_score":0.025245592,"about_ca_system_score_codex":0.00096149085,"about_ca_system_score_gemma":0.0024903102,"threshold_uncertainty_score":0.084454894},"labels":[],"label_agreement":null},{"id":"W2807602500","doi":"10.1093/database/bay044","title":"CircR2Disease: a manually curated database for experimentally supported circular RNAs associated with various diseases","year":2018,"lang":"en","type":"article","venue":"Database","topic":"Circular RNAs in diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":248,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"National Natural Science Foundation of China","keywords":"Computer science; Database; Circular RNA; Mechanism (biology); Disease; Computational biology; Bioinformatics; Biology; Gene; microRNA; Medicine; Genetics; Pathology","score_opus":0.014955260272099105,"score_gpt":0.28326290417516864,"score_spread":0.26830764390306955,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2807602500","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01907248,0.016991405,0.016869714,0.00044673364,0.00046992453,0.00040979433,0.90883577,0.022504061,0.0144002205],"genre_scores_gemma":[0.022176515,0.0034669742,0.023197701,0.0003041049,0.00015483738,0.00040334987,0.9449851,0.0016470263,0.0036644943],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99919814,0.00008256668,0.00014460394,0.0002770443,0.00019558122,0.00010204184],"domain_scores_gemma":[0.99837744,0.00046505185,0.00044294886,0.000261017,0.00020733113,0.0002461207],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008687141,0.0019757617,0.0028146717,0.008032872,0.0014102081,0.0021942088,0.0019069613,0.0020971359,0.036423437],"category_scores_gemma":[0.0034303651,0.0008330092,0.0012292741,0.0068160323,0.0004135837,0.0010425831,0.0020255665,0.0014189497,0.025776982],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0032198569,0.00022623164,0.01763876,0.023810713,0.0008924484,0.0036744648,0.00050192786,0.001029891,0.13796562,0.0072321044,0.69413614,0.10967185],"study_design_scores_gemma":[0.00046569968,0.00026464878,0.02668333,0.0012138415,0.00077205413,0.00609831,0.0001936947,0.0030860142,0.03167028,0.0034766167,0.925849,0.0002264382],"about_ca_topic_score_codex":0.001994131,"about_ca_topic_score_gemma":0.003126731,"teacher_disagreement_score":0.036423437,"about_ca_system_score_codex":0.0006398744,"about_ca_system_score_gemma":0.0023730644,"threshold_uncertainty_score":0.12184858},"labels":[],"label_agreement":null},{"id":"W2890022719","doi":"10.1093/database/bay086","title":"Spfy: an integrated graph database for real-time prediction of bacterial phenotypes and downstream comparative analyses","year":2018,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Public Health Agency of Canada","funders":"Public Health Agency of Canada","keywords":"Genome; Workflow; Computer science; Genomics; Computational biology; Comparative genomics; Identification (biology); In silico; Population; Biology; Data mining; Database; Genetics; Gene","score_opus":0.050896086236707855,"score_gpt":0.3228503856870912,"score_spread":0.2719542994503833,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2890022719","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009757827,0.0010504498,0.45899057,0.00045330595,0.00029277155,0.00041064713,0.122637436,0.40056744,0.005839493],"genre_scores_gemma":[0.084248155,0.002321869,0.43221375,0.000539529,0.00012433174,0.0014745952,0.44973242,0.024757044,0.0045882477],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9985709,0.00016312844,0.00019631667,0.00045815538,0.00048915454,0.00012230358],"domain_scores_gemma":[0.9973597,0.0008196954,0.00030037304,0.0007617709,0.0004846164,0.00027392033],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0020771043,0.0032695557,0.0019889267,0.00683376,0.0009706513,0.0035185902,0.0058017233,0.0016224734,0.014844398],"category_scores_gemma":[0.006698439,0.0015467574,0.002781823,0.004893663,0.00087454985,0.004420109,0.0036279387,0.0020386113,0.0086780405],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0033115766,0.0005607895,0.012061038,0.004792962,0.0010879486,0.0014381041,0.0008994542,0.08557493,0.039635547,0.059666097,0.4568264,0.33414528],"study_design_scores_gemma":[0.000893365,0.00047178374,0.009162647,0.0004684937,0.00042310794,0.0008723573,0.00042656026,0.41685772,0.03471336,0.13452764,0.40059203,0.00059088634],"about_ca_topic_score_codex":0.012491059,"about_ca_topic_score_gemma":0.011973006,"teacher_disagreement_score":0.014844398,"about_ca_system_score_codex":0.001565742,"about_ca_system_score_gemma":0.0029642277,"threshold_uncertainty_score":0.04965943},"labels":[],"label_agreement":null},{"id":"W2914171828","doi":"10.1093/database/bay147","title":"Overview of the BioCreative VI Precision Medicine Track: mining protein interactions and mutations for precision medicine","year":2018,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":75,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Institute for Research in Immunology and Cancer","funders":"U.S. National Library of Medicine; National Institute of General Medical Sciences; National Cancer Institute; National Institutes of Health","keywords":"Computer science; Task (project management); Precision medicine; Triage; F1 score; Relationship extraction; Precision and recall; Annotation; Information extraction; Named-entity recognition; Information retrieval; Relation (database); Personalized medicine; Natural language processing; Data science; Data mining; Artificial intelligence; Bioinformatics; Medicine; Genetics; Biology","score_opus":0.07326093043933397,"score_gpt":0.3871233921750868,"score_spread":0.31386246173575283,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2914171828","genre_codex":"dataset","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.035572726,0.030569758,0.30553755,0.006009399,0.0013655289,0.0048539913,0.4222121,0.14812417,0.0457548],"genre_scores_gemma":[0.022887422,0.005556324,0.3137175,0.001317606,0.0004376132,0.002119886,0.63377506,0.003149193,0.017039401],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9946636,0.00071374705,0.0006103887,0.0014750483,0.0021694256,0.00036772116],"domain_scores_gemma":[0.9913424,0.0026621327,0.000747712,0.0014191949,0.0029470634,0.0008814612],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0076008355,0.0022829596,0.0016207603,0.011905538,0.0014976878,0.005045407,0.0032464694,0.0021276716,0.015125301],"category_scores_gemma":[0.0103117395,0.0012639721,0.0024261915,0.008776105,0.00040506414,0.0045349346,0.0024544518,0.002001177,0.016549544],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010111572,0.00067164906,0.015609771,0.003946405,0.0006113498,0.00033505083,0.00049329665,0.006309255,0.019873122,0.00404181,0.5350568,0.41204032],"study_design_scores_gemma":[0.00050471874,0.0012565354,0.027476663,0.0009612133,0.0005565912,0.0011741185,0.0002564573,0.061174177,0.025231758,0.0062370882,0.87490696,0.00026369301],"about_ca_topic_score_codex":0.016743204,"about_ca_topic_score_gemma":0.022461848,"teacher_disagreement_score":0.016743204,"about_ca_system_score_codex":0.0019821443,"about_ca_system_score_gemma":0.0047437227,"threshold_uncertainty_score":0.050599158},"labels":[],"label_agreement":null},{"id":"W2920215701","doi":"10.1093/database/baz033","title":"Update on cpnDB: a reference database of chaperonin sequences","year":2019,"lang":"en","type":"article","venue":"Database","topic":"Heat shock proteins research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":39,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Chaperonin; GroEL; Biology; Phylogenetic tree; Computational biology; Identification (biology); Sequence analysis; Sequence (biology); Database; Computer science; Genetics; Ecology; Gene","score_opus":0.03033648656265123,"score_gpt":0.3165121856923562,"score_spread":0.2861756991297049,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2920215701","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.028702421,0.032289486,0.03347457,0.003165719,0.003215843,0.00032516735,0.8514359,0.024201576,0.023189213],"genre_scores_gemma":[0.009995289,0.0067358105,0.0364784,0.00061073294,0.0002912884,0.0001829878,0.9362626,0.0019994418,0.007443412],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9989102,0.00010313956,0.00021265354,0.00019714785,0.0004685176,0.0001083434],"domain_scores_gemma":[0.9957183,0.0004937125,0.00043499787,0.00059394166,0.0022400976,0.0005189653],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001958091,0.0013784386,0.0014280895,0.009665181,0.0008018455,0.00205224,0.0018863021,0.0009143528,0.014381275],"category_scores_gemma":[0.007518647,0.00071360223,0.00073697936,0.010267488,0.00032794397,0.002280723,0.0021213465,0.0019925458,0.01498475],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012302357,0.00016764177,0.004735138,0.0054589068,0.00016471818,0.0004890327,0.00034007747,0.0007138667,0.032295622,0.0030187883,0.63957024,0.31181586],"study_design_scores_gemma":[0.00004147187,0.000041563373,0.006183854,0.00048126365,0.000078238736,0.00047025687,0.00005145531,0.00044099684,0.0040489594,0.0007736717,0.98733133,0.000056952576],"about_ca_topic_score_codex":0.008586314,"about_ca_topic_score_gemma":0.008752844,"teacher_disagreement_score":0.014381275,"about_ca_system_score_codex":0.0010660863,"about_ca_system_score_gemma":0.0032865,"threshold_uncertainty_score":0.048110127},"labels":[],"label_agreement":null},{"id":"W2951817054","doi":"10.1093/database/baz022","title":"AYbRAH: a curated ortholog database for yeasts and fungi spanning 600 million years of evolution","year":2019,"lang":"en","type":"article","venue":"Database","topic":"Fungal and yeast genetics research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Ontario Ministry of Research and Innovation; Natural Sciences and Engineering Research Council of Canada","keywords":"Biology; Phylogenetics; Phylogenetic tree; Genome; KEGG; Phylogenomics; Database; Computational biology; Gene; Genetics; Transcriptome; Clade","score_opus":0.017373495672108906,"score_gpt":0.2865527150731592,"score_spread":0.26917921940105033,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2951817054","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.030114595,0.0108432295,0.018670797,0.00017051588,0.00027812138,0.00023969662,0.8921142,0.0397747,0.007794106],"genre_scores_gemma":[0.018123798,0.0025074654,0.027958624,0.0001228432,0.000039836188,0.00023829723,0.94676447,0.0030426201,0.0012020387],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99900824,0.0001040184,0.00023959519,0.00032469947,0.00021124247,0.00011213548],"domain_scores_gemma":[0.9987336,0.00023818943,0.0002965899,0.0002825003,0.0002384151,0.0002106853],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014606349,0.002537903,0.002415097,0.012315515,0.0020328658,0.0022918251,0.001990112,0.001754856,0.013082329],"category_scores_gemma":[0.0037677526,0.0011583312,0.0015461288,0.010922059,0.0004715383,0.0026582554,0.0038051885,0.0016088568,0.015372218],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0058671767,0.00041634357,0.03515717,0.039544087,0.0015137536,0.0055614905,0.0036390438,0.0027224266,0.17037714,0.011440316,0.525151,0.19861004],"study_design_scores_gemma":[0.00027401582,0.00019288507,0.029518347,0.001308752,0.00057801243,0.0023985335,0.00047440422,0.0017547586,0.011553018,0.0029454844,0.94882494,0.00017685044],"about_ca_topic_score_codex":0.002727997,"about_ca_topic_score_gemma":0.0047714235,"teacher_disagreement_score":0.013082329,"about_ca_system_score_codex":0.00067276886,"about_ca_system_score_gemma":0.0025081865,"threshold_uncertainty_score":0.04376471},"labels":[],"label_agreement":null},{"id":"W2952216057","doi":"10.1093/database/baw015","title":"Wikidata as a semantic framework for the Gene Wiki initiative","year":2016,"lang":"en","type":"article","venue":"Database","topic":"Wikis in Education and Collaboration","field":"Social Sciences","cited_by":61,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"National Institute of General Medical Sciences; National Institute on Drug Abuse; National Institutes of Health","keywords":"SPARQL; Computer science; UniProt; World Wide Web; Dissemination; Metadata; XML; Linked data; JSON; Open data; Semantic Web; Database; RDF; Gene; Biology","score_opus":0.0685358047867713,"score_gpt":0.3994968071533139,"score_spread":0.33096100236654263,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2952216057","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.002515762,0.0012008572,0.7864499,0.0029152585,0.0014118305,0.0009989666,0.039236512,0.13196975,0.033301223],"genre_scores_gemma":[0.017488291,0.0020770486,0.7375495,0.0019409964,0.00056226866,0.0028677436,0.18742444,0.03115321,0.018936632],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99478287,0.0013496762,0.0011832251,0.00094690506,0.0013072861,0.00043003267],"domain_scores_gemma":[0.99214786,0.0023493834,0.0006263845,0.0019019691,0.001513862,0.0014605256],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009885333,0.0017320524,0.0015556626,0.008355692,0.0026952405,0.009285938,0.005663339,0.0024412342,0.011631892],"category_scores_gemma":[0.013517757,0.0018212673,0.002846551,0.0072148615,0.0024799884,0.017013485,0.011753908,0.0059775077,0.021091722],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00041391255,0.00033606176,0.0015212428,0.0024534443,0.00022758581,0.00096819346,0.003846896,0.0048613003,0.00643769,0.4466123,0.41030338,0.122017995],"study_design_scores_gemma":[0.00005666615,0.000029217928,0.00038730341,0.00032300057,0.00003703064,0.00031244653,0.00032933697,0.0066235103,0.0019312195,0.073993884,0.91588306,0.000093390336],"about_ca_topic_score_codex":0.01027841,"about_ca_topic_score_gemma":0.012190842,"teacher_disagreement_score":0.011631892,"about_ca_system_score_codex":0.0021237866,"about_ca_system_score_gemma":0.009033212,"threshold_uncertainty_score":0.052279234},"labels":[],"label_agreement":null},{"id":"W2964307465","doi":"10.1093/database/baz077","title":"Tripal v3: an ontology-based toolkit for construction of FAIR biological community databases","year":2019,"lang":"en","type":"article","venue":"Database","topic":"Research Data Management Practices","field":"Computer Science","cited_by":32,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"National Institute of Food and Agriculture; Agricultural Research Service; Genome Canada; Saskatchewan Pulse Growers; Cotton Incorporated; Washington Tree Fruit Research Commission; U.S. Department of Agriculture; National Science Foundation","keywords":"Computer science; Ontology; World Wide Web; Database; Information retrieval","score_opus":0.2573459647153449,"score_gpt":0.4068510443479609,"score_spread":0.149505079632616,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2964307465","genre_codex":"methods","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0028246942,0.0005712578,0.5703159,0.0007241382,0.000358242,0.0014166704,0.17507666,0.22771297,0.020999469],"genre_scores_gemma":[0.01226416,0.0008215485,0.62956244,0.00051013933,0.00007067284,0.0046365005,0.2945529,0.04946454,0.00811702],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99458873,0.0012005,0.0011276696,0.0009936343,0.0016935059,0.0003960641],"domain_scores_gemma":[0.98819155,0.004698694,0.00093870336,0.0033743198,0.0017652448,0.0010313976],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.011193616,0.0019688443,0.0018185118,0.0093153175,0.002627432,0.00817646,0.0075542624,0.0013670556,0.045057185],"category_scores_gemma":[0.02515407,0.0028281708,0.0035223323,0.0099138105,0.0015258227,0.008002181,0.013571587,0.00430679,0.037413668],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007168205,0.00020122532,0.0040111323,0.0057989387,0.0005099672,0.0010126142,0.0049242373,0.0049415207,0.010092428,0.07688365,0.715728,0.17517944],"study_design_scores_gemma":[0.00009649991,0.000048098642,0.0015789622,0.0006272621,0.00012491996,0.00053392607,0.00096732803,0.007941796,0.004759764,0.045074955,0.9380518,0.00019470777],"about_ca_topic_score_codex":0.009417748,"about_ca_topic_score_gemma":0.013123291,"teacher_disagreement_score":0.99244577,"about_ca_system_score_codex":0.0022404864,"about_ca_system_score_gemma":0.0077454853,"threshold_uncertainty_score":0.15073127},"labels":[],"label_agreement":null},{"id":"W2967909176","doi":"10.1093/database/baz089","title":"PICEAdatabase: a web database for Picea omics and phenotypic information","year":2019,"lang":"en","type":"article","venue":"Database","topic":"Molecular Biology Techniques and Applications","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria","funders":"","keywords":"Pinaceae; Picea abies; Biology; Transcriptome; Omics; Ecology; Bioinformatics; Botany; Pinus <genus>; Genetics; Gene","score_opus":0.007577377018596199,"score_gpt":0.2544863433214788,"score_spread":0.2469089663028826,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2967909176","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009948499,0.0044276207,0.022525199,0.00016894778,0.00021991383,0.0003356467,0.91809005,0.03577203,0.008512019],"genre_scores_gemma":[0.008682671,0.0012211589,0.019817255,0.00015002681,0.000034787645,0.00053238956,0.96575534,0.0025618754,0.0012444239],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9994106,0.000059375623,0.00014029736,0.00018155327,0.00014476165,0.000063366526],"domain_scores_gemma":[0.99901915,0.00018554737,0.00015677582,0.00023835889,0.00023088857,0.00016929554],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011297157,0.0029114417,0.0020798708,0.008695183,0.0011652923,0.002820677,0.00265144,0.00092768326,0.015809277],"category_scores_gemma":[0.0028662046,0.0010560709,0.0013445402,0.009380494,0.00032323675,0.0024293547,0.0029403693,0.0014696011,0.014440121],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.003358538,0.00037926296,0.020031568,0.016065184,0.0017749323,0.0022708194,0.0014258854,0.0029958184,0.072612554,0.009568907,0.6654411,0.2040754],"study_design_scores_gemma":[0.00024755017,0.00010742628,0.027292551,0.00077069266,0.0005095352,0.0008971763,0.00023761371,0.003099621,0.011318988,0.0046076262,0.95073104,0.00018009418],"about_ca_topic_score_codex":0.0040365974,"about_ca_topic_score_gemma":0.0044932324,"teacher_disagreement_score":0.015809277,"about_ca_system_score_codex":0.0007389836,"about_ca_system_score_gemma":0.0018436102,"threshold_uncertainty_score":0.05288726},"labels":[],"label_agreement":null},{"id":"W2992521260","doi":"10.1093/database/baz123","title":"Reactome and ORCID—fine-grained credit attribution for community curation","year":2019,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":20,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"National Human Genome Research Institute; National Institutes of Health; Canada First Research Excellence Fund; National Institute of General Medical Sciences; European Molecular Biology Laboratory; Alfred P. Sloan Foundation","keywords":"Computer science; World Wide Web; Annotation; Attribution; Visibility; Interface (matter); Information retrieval; Data curation; Data science; Artificial intelligence","score_opus":0.017588765438800596,"score_gpt":0.25961206249851887,"score_spread":0.24202329705971828,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2992521260","genre_codex":"dataset","genre_gemma":"methods","domain_codex":null,"domain_gemma":"evaluation","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":"evaluation","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0019590145,0.0014292857,0.1493868,0.03125735,0.090821594,0.008705099,0.50819856,0.102373675,0.105868645],"genre_scores_gemma":[0.04321814,0.0021811975,0.2329895,0.014956743,0.014210104,0.027951764,0.3867914,0.13254552,0.14515565],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.939511,0.0130286915,0.010362708,0.012741329,0.018642271,0.005714075],"domain_scores_gemma":[0.6682616,0.054173674,0.021078248,0.10457477,0.12839277,0.023518968],"candidate_categories":["metaresearch","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.05996387,0.003354587,0.0067014447,0.017402487,0.007835189,0.02120647,0.009914666,0.0089541,0.6950373],"category_scores_gemma":[0.30514917,0.0037843913,0.0035913135,0.022020375,0.0065508513,0.01728478,0.020477353,0.008840304,0.5060241],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00024002025,0.000026296724,0.00041150616,0.0014077423,0.000054198063,0.000102341466,0.00019364496,0.00011881366,0.0008665701,0.009533335,0.9708283,0.016217245],"study_design_scores_gemma":[0.00034526986,0.00004190808,0.001127595,0.0014296796,0.00008259414,0.0003695729,0.00029613805,0.0017454983,0.0031364618,0.018299408,0.97294927,0.00017660913],"about_ca_topic_score_codex":0.0026403314,"about_ca_topic_score_gemma":0.0042274166,"teacher_disagreement_score":0.9400361,"about_ca_system_score_codex":0.0069214627,"about_ca_system_score_gemma":0.01922754,"threshold_uncertainty_score":0.43499225},"labels":[],"label_agreement":null},{"id":"W3022201072","doi":"10.1093/databa/baaa022","title":"Why data citation isn't working, and what to do about it","year":2020,"lang":"en","type":"article","venue":"Database","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":18,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Discovery Centre","funders":"Wellcome Trust; Wellcome","keywords":"Computer science; Citation; Citation database; Information retrieval; Granularity; Process (computing); Database; Data science; Value (mathematics); World Wide Web; Scopus; MEDLINE","score_opus":0.3412421372747082,"score_gpt":0.4181307828633919,"score_spread":0.07688864558868369,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3022201072","genre_codex":"commentary","genre_gemma":"methods","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.017423868,0.0762479,0.1973644,0.6596877,0.016508756,0.00031129844,0.0033754169,0.008036172,0.021044519],"genre_scores_gemma":[0.21497948,0.08606043,0.5324068,0.09981653,0.027334156,0.0014046467,0.0070490977,0.008361919,0.022586888],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.8985743,0.03614869,0.01539325,0.009678601,0.038397945,0.0018071922],"domain_scores_gemma":[0.50971884,0.2842106,0.02848784,0.045893855,0.12366795,0.0080208825],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.11872613,0.0009174472,0.002744847,0.0149153,0.0057256734,0.033613373,0.0068988577,0.007588559,0.0049060434],"category_scores_gemma":[0.46989852,0.001817642,0.0014638348,0.036666855,0.01084454,0.066578954,0.007079297,0.00916097,0.008830341],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00020210106,0.000111191366,0.021524828,0.005062525,0.0003095304,0.00025082036,0.004394021,0.0012227069,0.0014558827,0.12102369,0.25312015,0.59132254],"study_design_scores_gemma":[0.000086371314,0.00011646886,0.006873059,0.0048358566,0.0002620097,0.0013687436,0.0070625837,0.0056373053,0.0041933404,0.32762545,0.641394,0.00054478383],"about_ca_topic_score_codex":0.011713891,"about_ca_topic_score_gemma":0.007704909,"teacher_disagreement_score":0.9663866,"about_ca_system_score_codex":0.006153333,"about_ca_system_score_gemma":0.010336389,"threshold_uncertainty_score":0.6278913},"labels":[],"label_agreement":null},{"id":"W3034243427","doi":"10.1093/database/baaa045","title":"Curation of cancer hallmark-based genes and pathways for in silico characterization of chemical carcinogenesis","year":2020,"lang":"en","type":"article","venue":"Database","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"Kaohsiung Medical University Chung-Ho Memorial Hospital; National Health Research Institutes; Taipei Medical University; Kaohsiung Medical University; Ministry of Science and Technology, Taiwan","keywords":"Toxicogenomics; In silico; Carcinogen; Post hoc; Computer science; Data curation; Computational biology; Database; Biology; Data mining; Gene; Genetics; Medicine","score_opus":0.0536775691608744,"score_gpt":0.30342496570728045,"score_spread":0.24974739654640604,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3034243427","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.24275453,0.0054934807,0.42765695,0.0016036965,0.0003509608,0.001575864,0.2841564,0.025306113,0.0111020375],"genre_scores_gemma":[0.3078013,0.0036163111,0.4279023,0.00059923955,0.00005972122,0.0015979531,0.2533207,0.001511646,0.003590791],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995777,0.00011377635,0.000035353893,0.00013178779,0.00010629696,0.000035047604],"domain_scores_gemma":[0.9992949,0.00034737366,0.00006926092,0.000109146386,0.00014263217,0.000036826103],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00076941104,0.0010724441,0.00090575544,0.0028870597,0.0005633555,0.00091753644,0.0006508134,0.0004671009,0.0057847667],"category_scores_gemma":[0.0019318443,0.0003177962,0.001756157,0.0021358817,0.0002167957,0.0004944161,0.00071534526,0.00065874046,0.0013296366],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0025818606,0.0012423448,0.08310789,0.011282578,0.0020081287,0.0020054597,0.0005613567,0.15967198,0.39280638,0.015002383,0.07416999,0.25555968],"study_design_scores_gemma":[0.00060235383,0.0010673928,0.060408417,0.0005420463,0.0021746878,0.001423263,0.0005393511,0.44370872,0.23494583,0.033973694,0.22042888,0.00018542899],"about_ca_topic_score_codex":0.0037402555,"about_ca_topic_score_gemma":0.0082474705,"teacher_disagreement_score":0.0057847667,"about_ca_system_score_codex":0.00085370115,"about_ca_system_score_gemma":0.0025130492,"threshold_uncertainty_score":0.019352019},"labels":[],"label_agreement":null},{"id":"W3036787676","doi":"10.1093/databa/baaa033","title":"FOBI: an ontology to represent food intake data and associate it with metabolomic data","year":2020,"lang":"en","type":"article","venue":"Database","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":45,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"European Institute of Innovation and Technology","keywords":"Ontology; Computer science; Open Biomedical Ontologies; Visualization; Data science; Ontology-based data integration; Information retrieval; Annotation; Workflow; Identification (biology); Data mining; Suggested Upper Merged Ontology; Database; Semantic Web; Artificial intelligence","score_opus":0.1022735384683664,"score_gpt":0.3213407420082993,"score_spread":0.21906720353993292,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3036787676","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005567043,0.0012128929,0.86975026,0.0027391063,0.0007471949,0.0010616268,0.07875679,0.021685747,0.018479431],"genre_scores_gemma":[0.032867026,0.0032105567,0.79244184,0.001976302,0.00022148297,0.002237681,0.15281121,0.0042663654,0.009967577],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99776566,0.00035112636,0.0004919975,0.00041095953,0.00075894594,0.00022135845],"domain_scores_gemma":[0.9968374,0.0010996402,0.00034960674,0.00059559336,0.00084523985,0.00027254727],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0036790504,0.0015221976,0.0011487347,0.006549979,0.0017304264,0.003440042,0.0030360362,0.0023834885,0.0063185846],"category_scores_gemma":[0.0071310946,0.0010087072,0.0030042813,0.0059308554,0.0012429565,0.0076045054,0.0041435356,0.003123901,0.0044436036],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00064796035,0.0006887136,0.009844965,0.006075376,0.00071411306,0.0025315091,0.0051034316,0.0146924835,0.026985561,0.3224946,0.25207677,0.3581445],"study_design_scores_gemma":[0.00004434841,0.000040852978,0.0029855606,0.00080704555,0.00014818272,0.00079102686,0.00051354733,0.016133163,0.004166579,0.06536332,0.9088603,0.00014610981],"about_ca_topic_score_codex":0.027697368,"about_ca_topic_score_gemma":0.021533158,"teacher_disagreement_score":0.027697368,"about_ca_system_score_codex":0.002760962,"about_ca_system_score_gemma":0.005787151,"threshold_uncertainty_score":0.055072248},"labels":[],"label_agreement":null},{"id":"W3041921794","doi":"10.1093/database/baaa052","title":"Autophagy and Tumor Database: ATdb, a novel database connecting autophagy and tumor","year":2020,"lang":"en","type":"article","venue":"Database","topic":"Autophagy in Disease and Therapy","field":"Medicine","cited_by":23,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Fundamental Research Funds for the Central Universities; Zhejiang University; Natural Science Foundation of Zhejiang Province; National Natural Science Foundation of China","keywords":"Autophagy; Carcinogenesis; Biology; Database; Context (archaeology); microRNA; Computational biology; Bioinformatics; Cancer; Computer science; Cancer research; Gene; Genetics; Apoptosis","score_opus":0.035899491770201165,"score_gpt":0.29222322821871366,"score_spread":0.2563237364485125,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3041921794","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0055756075,0.0065620574,0.015097999,0.0007580812,0.0002571988,0.00030993932,0.9422511,0.022337457,0.0068506296],"genre_scores_gemma":[0.0112827085,0.002546689,0.015785541,0.00045105963,0.00008370277,0.00036075033,0.9670934,0.0011313007,0.0012648259],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9986261,0.0001745577,0.00046209188,0.0003380153,0.00026461974,0.00013460143],"domain_scores_gemma":[0.99728966,0.00074408646,0.000516009,0.00050921383,0.000446616,0.00049449207],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015149162,0.0022355076,0.0031628339,0.012849574,0.0009217427,0.0036457202,0.0028826268,0.0020280357,0.018482335],"category_scores_gemma":[0.0056110234,0.0008197625,0.0016994673,0.012288378,0.00047279696,0.003495442,0.003645258,0.0017383925,0.016506635],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0019162777,0.0002854328,0.015232385,0.016430205,0.0009198519,0.001819293,0.00052014523,0.0042964644,0.012352542,0.01336496,0.790967,0.1418955],"study_design_scores_gemma":[0.0004487859,0.00017929357,0.012131685,0.0010934517,0.0006306579,0.001893655,0.00033698176,0.008174742,0.009006085,0.013160051,0.9527425,0.0002021049],"about_ca_topic_score_codex":0.0049656085,"about_ca_topic_score_gemma":0.005077755,"teacher_disagreement_score":0.018482335,"about_ca_system_score_codex":0.0011478022,"about_ca_system_score_gemma":0.0033819922,"threshold_uncertainty_score":0.061829567},"labels":[],"label_agreement":null},{"id":"W3099269015","doi":"10.1093/database/baaa096","title":"The IMEx coronavirus interactome: an evolving map of <i>Coronaviridae</i> –host molecular interactions","year":2020,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":43,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Discovery Centre; University Health Network","funders":"H2020 European Research Council; National Institute of Diabetes and Digestive and Kidney Diseases; National Institute of Allergy and Infectious Diseases; National Institute of General Medical Sciences; National Cancer Institute; Wellcome Trust; National Human Genome Research Institute; National Institute of Mental Health; National Heart, Lung, and Blood Institute; National Eye Institute; National Institute on Aging; Associazione Italiana per la Ricerca sul Cancro; Fondation pour la Recherche Médicale; Canada Foundation for Innovation; European Molecular Biology Laboratory; Natural Sciences and Engineering Research Council of Canada; International Business Machines Corporation","keywords":"Interactome; Coronaviridae; Coronavirus disease 2019 (COVID-19); Coronavirus; Computational biology; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Pandemic; Proteomics; Biology; Computer science; Disease; Medicine; Genetics; Infectious disease (medical specialty); Gene","score_opus":0.020227507019509622,"score_gpt":0.28447117888153034,"score_spread":0.2642436718620207,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3099269015","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.051344518,0.016864011,0.0055181636,0.0015919378,0.00017568063,0.000109930064,0.9137073,0.0027603162,0.007928076],"genre_scores_gemma":[0.022913147,0.004685189,0.008653675,0.0004477145,0.00005471132,0.00016337288,0.9617709,0.00018788473,0.0011234533],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.99856645,0.00022201621,0.0001811957,0.0004889528,0.00038837772,0.00015300293],"domain_scores_gemma":[0.9979365,0.00060882984,0.00046233615,0.00037685488,0.00027627527,0.00033927095],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010548749,0.0011964286,0.0014200927,0.00742335,0.0010140216,0.0025797382,0.0012525527,0.0012542974,0.00514852],"category_scores_gemma":[0.004021565,0.00037359184,0.0010947755,0.010055035,0.00047319755,0.0020276448,0.003222458,0.0013019355,0.0044191857],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0022983109,0.00030974427,0.087940745,0.022637067,0.0015289894,0.0029882418,0.0021704643,0.005658235,0.06797586,0.020568548,0.6592479,0.12667596],"study_design_scores_gemma":[0.00009806584,0.00014431925,0.1180712,0.001698161,0.00048038698,0.0022996461,0.0007522329,0.005292032,0.00894178,0.010428736,0.851673,0.000120468416],"about_ca_topic_score_codex":0.0047465577,"about_ca_topic_score_gemma":0.009451559,"teacher_disagreement_score":0.00742335,"about_ca_system_score_codex":0.0011580989,"about_ca_system_score_gemma":0.0017545201,"threshold_uncertainty_score":0.017223537},"labels":[],"label_agreement":null},{"id":"W3101981777","doi":"10.1093/database/baaa079","title":"Measurement Recorder: developing a useful tool for making species descriptions that produces computable phenotypes","year":2020,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa; Agriculture and Agri-Food Canada; University of Manitoba","funders":"National Science Foundation","keywords":"Computer science; Usability; Software; Set (abstract data type); Ontology; Reuse; Variation (astronomy); Character (mathematics); Information retrieval; Data science; Software engineering; Human–computer interaction; Programming language","score_opus":0.17239052536197122,"score_gpt":0.3014869736680668,"score_spread":0.1290964483060956,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3101981777","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.002117612,0.00008867311,0.8229253,0.00041662832,0.00012596705,0.00047757768,0.005045022,0.16623864,0.0025644852],"genre_scores_gemma":[0.018327292,0.00027269308,0.93472695,0.00032394473,0.00005176407,0.0013364516,0.010204919,0.029230118,0.0055258702],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99465066,0.0014386947,0.0009184904,0.0010107182,0.0018230913,0.00015846611],"domain_scores_gemma":[0.96754074,0.018385846,0.0023158563,0.0074496884,0.00347752,0.0008303328],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01597781,0.0020912865,0.0012905526,0.004860613,0.0012280213,0.0040742396,0.004606051,0.0018098991,0.026852971],"category_scores_gemma":[0.0458105,0.0026993272,0.0026672436,0.0037252763,0.0014680044,0.010542324,0.005288737,0.003289466,0.011710982],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.001017628,0.00047450114,0.007987884,0.0045930594,0.0005447965,0.0014757321,0.008319071,0.00910005,0.04303045,0.06252945,0.22644223,0.6344852],"study_design_scores_gemma":[0.00042430515,0.00046578835,0.0051404396,0.00097265723,0.00020506972,0.0018281466,0.0011478007,0.08461258,0.07358461,0.056796275,0.7741722,0.000650087],"about_ca_topic_score_codex":0.0022901522,"about_ca_topic_score_gemma":0.0024157693,"teacher_disagreement_score":0.026852971,"about_ca_system_score_codex":0.0016500123,"about_ca_system_score_gemma":0.003352416,"threshold_uncertainty_score":0.08983213},"labels":[],"label_agreement":null},{"id":"W3121051245","doi":"10.1093/database/baaa072","title":"People are essential to linking biodiversity data","year":2020,"lang":"en","type":"article","venue":"Database","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":35,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Agriculture and Agri-Food Canada","funders":"Horizon 2020 Framework Programme; European Cooperation in Science and Technology; European Commission","keywords":"Digitization; Metadata; Biodiversity; Computer science; Herbarium; Data curation; Data science; Incentive; Citizen science; World Wide Web; Information retrieval; Ecology; Biology","score_opus":0.07942744792015947,"score_gpt":0.2676297436586908,"score_spread":0.1882022957385313,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3121051245","genre_codex":"methods","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.02545389,0.0062975027,0.5430284,0.045341834,0.0066053746,0.0014892248,0.11303739,0.018449279,0.24029712],"genre_scores_gemma":[0.19995585,0.008291228,0.49743932,0.017303975,0.0025383255,0.0017863532,0.2029852,0.010805805,0.058893863],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.975244,0.0076497886,0.0031859502,0.004582831,0.008271003,0.0010664932],"domain_scores_gemma":[0.9216528,0.024269035,0.004558371,0.034190916,0.011940012,0.0033889445],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01946875,0.0008895065,0.0015452913,0.009315424,0.004013826,0.0118017765,0.0034472982,0.0031801788,0.03521431],"category_scores_gemma":[0.0981567,0.0015305781,0.0013816778,0.016000045,0.0026742634,0.030846277,0.016388115,0.004271352,0.023417879],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00030718977,0.00022101679,0.029327983,0.0019136049,0.0003466813,0.00083124446,0.0069150273,0.0020095895,0.0020881526,0.2623779,0.29579878,0.39786276],"study_design_scores_gemma":[0.000020052139,0.000022027547,0.0036412347,0.00066142733,0.00009088821,0.00039494483,0.0015252748,0.0013628702,0.0013007872,0.157846,0.8330619,0.00007262463],"about_ca_topic_score_codex":0.010702241,"about_ca_topic_score_gemma":0.0078065945,"teacher_disagreement_score":0.03521431,"about_ca_system_score_codex":0.0016768243,"about_ca_system_score_gemma":0.0050462303,"threshold_uncertainty_score":0.11780357},"labels":[],"label_agreement":null},{"id":"W3130317325","doi":"10.1093/database/baab006","title":"Curation of over 10 000 transcriptomic studies to enable data reuse","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":60,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada's Michael Smith Genome Sciences Centre; Genome British Columbia; University of British Columbia","funders":"National Institute of Mental Health; Natural Sciences and Engineering Research Council of Canada","keywords":"Metadata; Computer science; Reuse; Software; Data curation; Gemma; World Wide Web; Database; Information retrieval; Data science; Biology","score_opus":0.09921660307901177,"score_gpt":0.3764284461266624,"score_spread":0.2772118430476506,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3130317325","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04222474,0.015830563,0.17955367,0.0042521367,0.002202909,0.0020216398,0.69257987,0.033482403,0.02785204],"genre_scores_gemma":[0.038044132,0.0043595936,0.20567296,0.0018332121,0.00038230926,0.0027504398,0.7339167,0.0077211666,0.005319431],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9890205,0.0018700408,0.0022484479,0.0035298087,0.0027158067,0.00061528146],"domain_scores_gemma":[0.9659796,0.007825448,0.0025631925,0.013540557,0.008262176,0.0018290825],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.018223032,0.0018311074,0.0027862394,0.016684202,0.0035900392,0.006542317,0.002553085,0.0016687679,0.013553361],"category_scores_gemma":[0.038978256,0.001178475,0.003314947,0.017826324,0.00097275054,0.0034671382,0.009165266,0.003077101,0.010082454],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0014514377,0.00021277007,0.032185405,0.01224538,0.0023204947,0.001772744,0.003295031,0.002845425,0.13060343,0.013361306,0.5089692,0.29073742],"study_design_scores_gemma":[0.0001770744,0.00011519836,0.034095727,0.0016656389,0.0011004603,0.0010827578,0.0005985456,0.002727247,0.022556787,0.014850733,0.92084116,0.00018860548],"about_ca_topic_score_codex":0.00635498,"about_ca_topic_score_gemma":0.019653156,"teacher_disagreement_score":0.9974469,"about_ca_system_score_codex":0.002322788,"about_ca_system_score_gemma":0.0113002835,"threshold_uncertainty_score":0.09637374},"labels":[],"label_agreement":null},{"id":"W3159409148","doi":"10.1093/database/baab021","title":"Increasing metadata coverage of SRA BioSample entries using deep learning–based named entity recognition","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":21,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Canadian Institutes of Health Research; National Institutes of Health; National Institute of General Medical Sciences; Canadian Institute for Advanced Research","keywords":"Metadata; Computer science; Named-entity recognition; Information retrieval; Artificial intelligence; Deep learning; Natural language processing; World Wide Web","score_opus":0.040499717845373796,"score_gpt":0.29040079590886664,"score_spread":0.24990107806349285,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3159409148","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.7660538,0.0017028359,0.14144291,0.0009786191,0.00030667768,0.00017630671,0.05096965,0.031771813,0.00659751],"genre_scores_gemma":[0.6188905,0.00054789276,0.23124553,0.00064094935,0.00009824872,0.00032377042,0.14315642,0.00092333945,0.004173242],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992398,0.00010312361,0.00007126324,0.00035090008,0.00016685668,0.00006802417],"domain_scores_gemma":[0.9977678,0.00079824793,0.00029866575,0.0004402101,0.000597556,0.00009746701],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0016393727,0.0008448887,0.00057585083,0.0020175185,0.0005313392,0.00089185516,0.00094720337,0.0006944997,0.0014222872],"category_scores_gemma":[0.0047716135,0.00026257156,0.00088306353,0.0013985683,0.00033723513,0.0018735273,0.0013267178,0.0010109749,0.0019574342],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012904335,0.00085281907,0.1779504,0.0009631697,0.0004879589,0.00088519795,0.00094143767,0.05530151,0.17998274,0.0024709424,0.053553745,0.52531976],"study_design_scores_gemma":[0.000067215224,0.00036413717,0.058262005,0.00018696126,0.00022129354,0.0003728359,0.0006635878,0.7698882,0.13170463,0.0057821567,0.032358523,0.00012852416],"about_ca_topic_score_codex":0.0065763067,"about_ca_topic_score_gemma":0.014799806,"teacher_disagreement_score":0.99836063,"about_ca_system_score_codex":0.0005611786,"about_ca_system_score_gemma":0.0005910556,"threshold_uncertainty_score":0.013076067},"labels":[],"label_agreement":null},{"id":"W3163423203","doi":"10.1093/database/baab030","title":"Integration of 1:1 orthology maps and updated datasets into Echinobase","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Echinoderm biology and ecology","field":"Agricultural and Biological Sciences","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"Eunice Kennedy Shriver National Institute of Child Health and Human Development; National Science Foundation","keywords":"Echinoderm; Strongylocentrotus purpuratus; Resource (disambiguation); Biology; Genome; Variety (cybernetics); Computer science; Data science; Evolutionary biology; Sea urchin; Computational biology; Ecology; Gene; Artificial intelligence; Genetics","score_opus":0.017294260163247997,"score_gpt":0.2432262885103951,"score_spread":0.2259320283471471,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3163423203","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007622362,0.000996535,0.03245251,0.0004623846,0.000403521,0.00027532206,0.90140915,0.0428254,0.0135529],"genre_scores_gemma":[0.0044680396,0.0004694251,0.030259617,0.00013593977,0.00003112602,0.00030925957,0.9587546,0.003904057,0.0016679763],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9981639,0.00018034395,0.00038001905,0.00053655205,0.00059368927,0.00014556474],"domain_scores_gemma":[0.9963486,0.0006879913,0.00041161408,0.0011139917,0.0009224859,0.00051517016],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0035697438,0.001919368,0.0017603132,0.009894475,0.0018567691,0.0038613891,0.0036406904,0.0016566514,0.037598558],"category_scores_gemma":[0.010589826,0.0014963691,0.0012162217,0.011153834,0.00046331718,0.0044352384,0.006607585,0.0030862687,0.03894026],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0015553064,0.00034269053,0.011779282,0.0065361364,0.00053365756,0.001230696,0.0016588732,0.0034526158,0.034046885,0.014878328,0.8082707,0.115714826],"study_design_scores_gemma":[0.00017020291,0.00007716589,0.011137463,0.0004896438,0.00016271291,0.0005407263,0.0002771773,0.0019991607,0.010173283,0.007740817,0.9670949,0.00013672627],"about_ca_topic_score_codex":0.0047555566,"about_ca_topic_score_gemma":0.0070849103,"teacher_disagreement_score":0.037598558,"about_ca_system_score_codex":0.001291345,"about_ca_system_score_gemma":0.0030701947,"threshold_uncertainty_score":0.12577969},"labels":[],"label_agreement":null},{"id":"W3170771812","doi":"10.1093/database/baab069","title":"OBO Foundry in 2021: operationalizing open data principles to evaluate ontologies","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":206,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"National Human Genome Research Institute; National Institutes of Health","keywords":"Computer science; Operationalization; World Wide Web; Data science","score_opus":0.20312183623828936,"score_gpt":0.4157198848331276,"score_spread":0.21259804859483825,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3170771812","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.066896185,0.00060236367,0.82981217,0.015964601,0.0015451371,0.004429367,0.00895909,0.023754694,0.04803633],"genre_scores_gemma":[0.15025926,0.00027434833,0.81475496,0.0021071844,0.00017445472,0.0021337671,0.018640442,0.0050770133,0.006578508],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.8808183,0.055065706,0.014028128,0.008171415,0.036778785,0.0051376563],"domain_scores_gemma":[0.7500297,0.071702175,0.0142851975,0.076058626,0.080444194,0.0074800663],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.1664337,0.0015754693,0.0021855016,0.009912846,0.005139089,0.023840865,0.005191683,0.003453957,0.0045147403],"category_scores_gemma":[0.23832104,0.0015126322,0.0026370236,0.005486563,0.008647444,0.021849371,0.020314638,0.0064283097,0.002579347],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011592534,0.0010290712,0.033302665,0.0010029402,0.00027595038,0.0006103182,0.00978562,0.023028875,0.008816028,0.58681285,0.07928936,0.2548871],"study_design_scores_gemma":[0.0002992352,0.000578221,0.009434195,0.0020681322,0.00015521262,0.00039152292,0.0053740684,0.118314326,0.028204838,0.4501958,0.38450563,0.00047891852],"about_ca_topic_score_codex":0.024880331,"about_ca_topic_score_gemma":0.018325223,"teacher_disagreement_score":0.9948083,"about_ca_system_score_codex":0.011909521,"about_ca_system_score_gemma":0.023264822,"threshold_uncertainty_score":0.8801961},"labels":[],"label_agreement":null},{"id":"W3176752316","doi":"10.1093/database/baab035","title":"Which methods are the most effective in enabling novice users to participate in ontology creation? A usability study","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa; Government of Canada; Agriculture and Agri-Food Canada; University of Manitoba","funders":"Canadian Institutes of Health Research; National Science Foundation","keywords":"Usability; Wizard; Computer science; Ontology; Interoperability; World Wide Web; Set (abstract data type); Think aloud protocol; Data curation; Data science; Information retrieval; Human–computer interaction","score_opus":0.03992296776808297,"score_gpt":0.40249750443283333,"score_spread":0.36257453666475037,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3176752316","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9641719,0.0019655684,0.025899585,0.000875129,0.000091461356,0.0017752115,0.00023430456,0.0008941712,0.004092736],"genre_scores_gemma":[0.88873184,0.0013783606,0.10452113,0.0005424241,0.000059461436,0.0026806179,0.00035206415,0.00043911612,0.0012949957],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.9609003,0.025766708,0.00517904,0.0028189803,0.004163184,0.0011718156],"domain_scores_gemma":[0.630496,0.3292925,0.009905518,0.009320769,0.018300751,0.0026843941],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.049793772,0.001453377,0.0012832681,0.00274489,0.0014058748,0.004731631,0.0011797897,0.0021468715,0.0015513285],"category_scores_gemma":[0.18332829,0.00096932845,0.001206735,0.0015210327,0.0014216974,0.005027238,0.0020677124,0.0009613841,0.0005276149],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.009032154,0.0054921205,0.14908518,0.014006149,0.0009158808,0.00093727617,0.09004281,0.0017331031,0.035839852,0.0014574755,0.0061892155,0.68526876],"study_design_scores_gemma":[0.0057352125,0.054193586,0.52140737,0.008442292,0.004615962,0.004356417,0.1614803,0.055073056,0.09001995,0.011460832,0.08078992,0.0024251668],"about_ca_topic_score_codex":0.0015334507,"about_ca_topic_score_gemma":0.0031144666,"teacher_disagreement_score":0.049793772,"about_ca_system_score_codex":0.0011560633,"about_ca_system_score_gemma":0.0013086294,"threshold_uncertainty_score":0.26333773},"labels":[],"label_agreement":null},{"id":"W3177341850","doi":"10.1093/database/baab032","title":"MetamORF: a repository of unique short open reading frames identified by both experimental and computational approaches for gene and metagene analyses","year":2021,"lang":"en","type":"article","venue":"Database","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":42,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Agence Nationale de la Recherche; Canadian Institutes of Health Research; Aix-Marseille Université","keywords":"ORFS; Open reading frame; Genome; Biology; Gene; Genome browser; Computational biology; Genetics; Computer science; Genomics","score_opus":0.08106924100586585,"score_gpt":0.34013706402289134,"score_spread":0.25906782301702547,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3177341850","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.006960764,0.004176952,0.041502416,0.0002226442,0.0002650476,0.00029117323,0.83991605,0.102985404,0.0036794886],"genre_scores_gemma":[0.0061046304,0.0015661309,0.043560512,0.00009704687,0.00006826336,0.00054439326,0.93957794,0.0077969166,0.0006841397],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99665254,0.0005463525,0.0005554535,0.0010079904,0.0009164354,0.00032131365],"domain_scores_gemma":[0.9952696,0.0015433047,0.0010391746,0.0012441752,0.00041994135,0.00048378584],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005057709,0.005902174,0.0051205726,0.01807148,0.0015524331,0.004386589,0.0057134926,0.0036486955,0.020284798],"category_scores_gemma":[0.011034656,0.0020493143,0.0048354226,0.01565607,0.0008568917,0.0047804327,0.005689653,0.0032092137,0.031501986],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0051089204,0.00079798076,0.015889721,0.039179787,0.0031156652,0.0039121984,0.0016018115,0.008901876,0.106270574,0.016516209,0.6010596,0.19764566],"study_design_scores_gemma":[0.0011773533,0.00067034207,0.02463995,0.0035980793,0.0017169554,0.0026186323,0.00040047042,0.012543228,0.04345247,0.017435208,0.8910244,0.0007229786],"about_ca_topic_score_codex":0.001742287,"about_ca_topic_score_gemma":0.002219433,"teacher_disagreement_score":0.020284798,"about_ca_system_score_codex":0.0011108192,"about_ca_system_score_gemma":0.00358173,"threshold_uncertainty_score":0.06785941},"labels":[],"label_agreement":null},{"id":"W3177366615","doi":"10.1093/database/baab038","title":"ReMeDy: a platform for integrating and sharing published stem cell research data with a focus on iPSC trials","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Pluripotent Stem Cells Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Center for Advancing Translational Sciences; National Heart, Lung, and Blood Institute; Canadian Institutes of Health Research; National Institutes of Health","keywords":"Focus (optics); Data sharing; Computer science; Data science; Stem cell; World Wide Web; Research data; Internet privacy; Medicine; Biology; Data curation; Alternative medicine","score_opus":0.1962832745826133,"score_gpt":0.39653851683930136,"score_spread":0.20025524225668806,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3177366615","genre_codex":"methods","genre_gemma":"software","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03994757,0.0032472652,0.45987356,0.008614705,0.00096386904,0.0045288024,0.25874153,0.18849736,0.035585273],"genre_scores_gemma":[0.16383532,0.0034196912,0.49579,0.0023308152,0.0005334302,0.0027342339,0.31211373,0.008706299,0.010536536],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9920717,0.0017152914,0.0023490784,0.0013167396,0.0022641674,0.00028297355],"domain_scores_gemma":[0.9695403,0.0086054,0.0042591654,0.0117432345,0.0033899012,0.0024619643],"candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.016541025,0.0010735476,0.0015310111,0.0148258945,0.0012160722,0.006811014,0.0040544304,0.001982401,0.01808311],"category_scores_gemma":[0.035357453,0.0008336814,0.0011897903,0.012056749,0.00093577226,0.008118778,0.011258959,0.0016421699,0.011071586],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0031979766,0.0012716063,0.028969912,0.005574515,0.0009663571,0.0018640296,0.0040440713,0.00787662,0.028842412,0.04625903,0.29605788,0.5750757],"study_design_scores_gemma":[0.0010253624,0.0007537169,0.02708472,0.0013951366,0.0006917286,0.0018131054,0.0022226817,0.025476027,0.025041915,0.04114277,0.8726943,0.00065856386],"about_ca_topic_score_codex":0.0024637452,"about_ca_topic_score_gemma":0.0024142805,"teacher_disagreement_score":0.9959456,"about_ca_system_score_codex":0.0013064762,"about_ca_system_score_gemma":0.004502738,"threshold_uncertainty_score":0.08747834},"labels":[],"label_agreement":null},{"id":"W3181498676","doi":"10.1093/database/baab040","title":"Standardization of assay representation in the Ontology for Biomedical Investigations","year":2021,"lang":"en","type":"review","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"National Institute of Allergy and Infectious Diseases; National Human Genome Research Institute; Bill and Melinda Gates Foundation; National Institutes of Health; National Science Foundation","keywords":"Standardization; Ontology; Computer science; Process (computing); Open Biomedical Ontologies; Software engineering; Representation (politics); Hierarchy; Process ontology; Information retrieval; Ontology Inference Layer; OWL-S; Term (time); World Wide Web; Semantic Web; Suggested Upper Merged Ontology; Programming language; Semantic Web Stack","score_opus":0.11938547991577698,"score_gpt":0.4300302172784989,"score_spread":0.3106447373627219,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3181498676","genre_codex":"review","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0020294618,0.8701327,0.093468636,0.00958527,0.003373534,0.00039609344,0.0011276198,0.00039220433,0.019494465],"genre_scores_gemma":[0.012446586,0.8472579,0.12530643,0.005439796,0.0007192548,0.0005598336,0.0036813535,0.0001550638,0.0044338047],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99654007,0.00095385185,0.0007682963,0.0004586613,0.0011306343,0.00014860554],"domain_scores_gemma":[0.9940519,0.0027977908,0.00052856305,0.00073272386,0.0017429687,0.00014603714],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014602051,0.0009517474,0.0021333606,0.0066903858,0.0006307993,0.003886955,0.0030668438,0.0021519996,0.0016261349],"category_scores_gemma":[0.009668819,0.00057605415,0.0016838834,0.007293423,0.0023020424,0.0057595912,0.0025907836,0.0049639544,0.0018518885],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00006271723,0.00011871926,0.00053469726,0.018993568,0.00019504182,0.0003148557,0.0006291843,0.0011044288,0.0057113348,0.16941312,0.021656342,0.781266],"study_design_scores_gemma":[0.000009901303,0.000031054693,0.0005125029,0.00510426,0.00014933059,0.000609671,0.00018623222,0.0003335836,0.0022560991,0.0145310275,0.97624105,0.000035346107],"about_ca_topic_score_codex":0.0048957113,"about_ca_topic_score_gemma":0.0035263563,"teacher_disagreement_score":0.014602051,"about_ca_system_score_codex":0.0039474647,"about_ca_system_score_gemma":0.0101489145,"threshold_uncertainty_score":0.07722396},"labels":[],"label_agreement":null},{"id":"W3190629400","doi":"10.1093/database/baab051","title":"OUP accepted manuscript","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Banana Cultivation and Research","field":"Agricultural and Biological Sciences","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"","keywords":"Computer science; Interoperability; Metadata; Schema (genetic algorithms); World Wide Web; Open source; Database; Software; Information retrieval","score_opus":0.06820757546597621,"score_gpt":0.27493018157125854,"score_spread":0.2067226061052823,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3190629400","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.003065669,0.005523904,0.0063431333,0.009740767,0.022548052,0.00047327307,0.013447642,0.004233297,0.93462425],"genre_scores_gemma":[0.008691816,0.002563252,0.0026981602,0.0017392645,0.0016421726,0.00014304092,0.009085169,0.0012498555,0.9721872],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99735034,0.0004081306,0.00019068309,0.00074082,0.0009387497,0.0003711922],"domain_scores_gemma":[0.9959881,0.0004289467,0.00018414955,0.0009345886,0.0015845455,0.00087970257],"candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.002635064,0.0012644917,0.0014324608,0.0026893984,0.0024515209,0.010195076,0.002910927,0.0042078705,0.78304636],"category_scores_gemma":[0.007591971,0.00078240514,0.0011757863,0.0022637532,0.001402049,0.0031136214,0.0063332804,0.0030081247,0.7153288],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00046609872,0.00022049557,0.0013798294,0.0010845484,0.000064771004,0.00090570934,0.00032927896,0.00026739037,0.0049873292,0.017476678,0.71147364,0.26134422],"study_design_scores_gemma":[0.000017616516,0.000026186275,0.00042687586,0.0000894504,0.000005305432,0.00014721487,0.00006649158,0.00006399515,0.0005023173,0.0009373834,0.9977107,0.0000065409085],"about_ca_topic_score_codex":0.0026844009,"about_ca_topic_score_gemma":0.00354528,"teacher_disagreement_score":0.21695364,"about_ca_system_score_codex":0.0017766722,"about_ca_system_score_gemma":0.0033863138,"threshold_uncertainty_score":0.30945802},"labels":[],"label_agreement":null},{"id":"W3203751410","doi":"10.1093/database/baab062","title":"Classifying domain-specific text documents containing ambiguous keywords","year":2021,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"Eunice Kennedy Shriver National Institute of Child Health and Human Development","keywords":"Computer science; Set (abstract data type); Domain (mathematical analysis); Information retrieval; Artificial intelligence; Naive Bayes classifier; Overfitting; Identifier; Machine learning; Support vector machine; Ambiguity; Data mining; Artificial neural network","score_opus":0.027044701699820522,"score_gpt":0.290375614861075,"score_spread":0.26333091316125445,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3203751410","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.41285816,0.11671267,0.15473124,0.01113487,0.0046367166,0.0067344606,0.23573327,0.013085415,0.04437329],"genre_scores_gemma":[0.30902308,0.032948438,0.51168305,0.0028687215,0.0014586045,0.002540709,0.12884879,0.0008441078,0.009784491],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99483657,0.0010444738,0.0017965898,0.00079994317,0.0013637651,0.00015870672],"domain_scores_gemma":[0.93145716,0.050163794,0.0066127367,0.0028036013,0.0082093,0.0007533582],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00884917,0.0010974831,0.0017116887,0.035691865,0.0012953893,0.0045972858,0.0011123196,0.0014949875,0.0067792516],"category_scores_gemma":[0.03963737,0.00031444276,0.0016046237,0.02888635,0.00055102084,0.0028796485,0.0011377129,0.0005608382,0.003566945],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0015703838,0.0004481158,0.037727963,0.05686154,0.0012508805,0.00480184,0.0020073126,0.0028222543,0.050806627,0.007982881,0.08699573,0.74672455],"study_design_scores_gemma":[0.00083298434,0.0015520273,0.10284069,0.013410776,0.004321645,0.013380378,0.009624744,0.022519663,0.11479493,0.038255874,0.6779065,0.0005597386],"about_ca_topic_score_codex":0.0017132073,"about_ca_topic_score_gemma":0.0036635222,"teacher_disagreement_score":0.035691865,"about_ca_system_score_codex":0.0011187114,"about_ca_system_score_gemma":0.0036396284,"threshold_uncertainty_score":0.04679948},"labels":[],"label_agreement":null},{"id":"W4210447837","doi":"10.1093/database/baac001","title":"Authors’ attitude toward adopting a new workflow to improve the computability of phenotype publications","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Agriculture and Agri-Food Canada; University of Ottawa; University of Manitoba","funders":"","keywords":"Workflow; Ambiguity; Computer science; World Wide Web; Knowledge management; Database","score_opus":0.04305262788612411,"score_gpt":0.30935208021017413,"score_spread":0.26629945232405,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4210447837","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reporting","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reporting","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.96861035,0.00047479555,0.013925717,0.007869333,0.00014195198,0.0001683161,0.000093541035,0.0002289712,0.008487062],"genre_scores_gemma":[0.97990304,0.00054706057,0.01591503,0.0014026016,0.00010127009,0.00011365597,0.00013703371,0.00007524936,0.0018049706],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.94992065,0.023392592,0.0073907776,0.002812623,0.0148430625,0.0016402426],"domain_scores_gemma":[0.5851863,0.23606034,0.07675066,0.032346208,0.05456159,0.015094836],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.08703584,0.00032245094,0.00034878493,0.0030441063,0.0022882072,0.0081009185,0.0012914626,0.0017985984,0.0020788903],"category_scores_gemma":[0.24119855,0.0005927975,0.0010708985,0.0019332719,0.0032609443,0.005340707,0.0034839758,0.0020158112,0.0008866987],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0005558903,0.0004540347,0.66463864,0.0009989887,0.000290583,0.0012171022,0.14873518,0.002587598,0.020390933,0.007319152,0.004580356,0.14823155],"study_design_scores_gemma":[0.00028595235,0.0023758123,0.6059825,0.0017144007,0.0005038195,0.0049431017,0.19437326,0.011897688,0.016660266,0.01703836,0.14324059,0.0009842027],"about_ca_topic_score_codex":0.0022863334,"about_ca_topic_score_gemma":0.0026768567,"teacher_disagreement_score":0.91296417,"about_ca_system_score_codex":0.0024617913,"about_ca_system_score_gemma":0.0050076037,"threshold_uncertainty_score":0.46029502},"labels":[],"label_agreement":null},{"id":"W4212800016","doi":"10.1093/database/baac009","title":"Evaluating the predictive accuracy of curated biological pathways in a public knowledgebase","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada Research Chairs; University of New Brunswick; University of Toronto; Ontario Institute for Cancer Research","funders":"National Human Genome Research Institute","keywords":"Biological pathway; Downregulation and upregulation; Computer science; Signal transduction; Computational biology; Biology; Genetics; Gene; Gene expression","score_opus":0.09095347415119943,"score_gpt":0.32690795826394903,"score_spread":0.2359544841127496,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4212800016","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8607601,0.00635109,0.07291612,0.0009590505,0.00017388226,0.00046846955,0.037006382,0.014275154,0.0070897085],"genre_scores_gemma":[0.88746136,0.0010980616,0.05879706,0.00016220506,0.000038791906,0.00022587868,0.051366284,0.00033587945,0.00051452446],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99102247,0.0029906183,0.00093467114,0.002327828,0.0024152799,0.00030912898],"domain_scores_gemma":[0.9227055,0.061763577,0.0030287988,0.0071901428,0.0047905357,0.00052147324],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.013107626,0.0017456656,0.00083856256,0.008054216,0.0009110859,0.0031371193,0.0022621255,0.0021896972,0.0019744723],"category_scores_gemma":[0.07743516,0.00055093266,0.0013490944,0.004676393,0.00083682477,0.0024625852,0.002049605,0.0009321189,0.001002394],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0036325858,0.0009381722,0.1958781,0.004794851,0.0028447958,0.0013454591,0.0008710019,0.536251,0.0104975505,0.004129931,0.016156182,0.22266036],"study_design_scores_gemma":[0.00021854808,0.00038467452,0.03341779,0.0005557314,0.0006630202,0.0006043238,0.00035355816,0.9338834,0.016791914,0.004289092,0.008759391,0.00007864805],"about_ca_topic_score_codex":0.010488111,"about_ca_topic_score_gemma":0.010298818,"teacher_disagreement_score":0.013107626,"about_ca_system_score_codex":0.001692102,"about_ca_system_score_gemma":0.002747686,"threshold_uncertainty_score":0.06932056},"labels":[],"label_agreement":null},{"id":"W4224325974","doi":"10.1093/database/baac069","title":"Multi-label classification for biomedical literature: an overview of the BioCreative VII LitCovid Track for COVID-19 literature topic annotations","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Topic Modeling","field":"Computer Science","cited_by":40,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"U.S. National Library of Medicine; Medical Research Council; National Institutes of Health","keywords":"Computer science; Coronavirus disease 2019 (COVID-19); Annotation; Named-entity recognition; Scientific literature; Information retrieval; Data science; Artificial intelligence; Natural language processing; World Wide Web; Task (project management); Medicine; Infectious disease (medical specialty)","score_opus":0.2046024088696558,"score_gpt":0.4089928303778776,"score_spread":0.2043904215082218,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4224325974","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012055004,0.073803544,0.1060613,0.011625832,0.0045686616,0.003988834,0.6112822,0.14208667,0.034527957],"genre_scores_gemma":[0.0055425055,0.0098674055,0.15123805,0.0030053922,0.00079790497,0.0022309653,0.8128781,0.006275997,0.008163751],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9812501,0.003814021,0.0032635592,0.004289219,0.00638824,0.0009948596],"domain_scores_gemma":[0.955544,0.015401774,0.0046804706,0.007638965,0.012559704,0.0041751866],"candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.019823898,0.0028758228,0.0031121548,0.042533647,0.004193842,0.011486439,0.0051621133,0.0036162399,0.022048146],"category_scores_gemma":[0.047683895,0.0016000797,0.004641766,0.03224641,0.0011904661,0.008785834,0.009696248,0.003386465,0.039940547],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00054124574,0.00032313002,0.0058385152,0.010145025,0.00042799243,0.00039076182,0.0007372942,0.000972156,0.010533012,0.0032573398,0.73061574,0.2362178],"study_design_scores_gemma":[0.00014299745,0.00017197097,0.007483691,0.0021608411,0.00016065376,0.00049917685,0.00025668746,0.0045755873,0.004261619,0.0032111774,0.9769165,0.00015915488],"about_ca_topic_score_codex":0.012929982,"about_ca_topic_score_gemma":0.031409204,"teacher_disagreement_score":0.95746636,"about_ca_system_score_codex":0.0047586537,"about_ca_system_score_gemma":0.01145212,"threshold_uncertainty_score":0.10484004},"labels":[],"label_agreement":null},{"id":"W4225573257","doi":"10.1093/database/baac034","title":"GrainGenes: a data-rich repository for small grains genetics and genomics","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Wheat and Barley Genetics and Pathology","field":"Agricultural and Biological Sciences","cited_by":96,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Agriculture and Agri-Food Canada","funders":"Agricultural Research Service","keywords":"Genome browser; Genome; Triticeae; Genomics; Biology; Interoperability; Metadata; World Wide Web; Genetics; Database; Computational biology; Computer science; Gene","score_opus":0.06471702717790244,"score_gpt":0.24476727253286695,"score_spread":0.1800502453549645,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4225573257","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0026491224,0.0013060508,0.024153544,0.00041479454,0.00020281824,0.00017016349,0.89416087,0.06659598,0.010346512],"genre_scores_gemma":[0.0044541177,0.00063391926,0.01898898,0.00020899416,0.000033630375,0.0002552538,0.9684483,0.0050874506,0.0018893946],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99828017,0.0001687998,0.0002335432,0.0005422727,0.0005870715,0.0001882753],"domain_scores_gemma":[0.99682754,0.0006656638,0.00043185643,0.0009135698,0.0006183254,0.0005431197],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.0027310695,0.0019611958,0.0023504035,0.008139328,0.0021931597,0.004235085,0.004507068,0.001915485,0.026701463],"category_scores_gemma":[0.007322497,0.0013271638,0.0012827378,0.015758796,0.0007310605,0.0040993863,0.004899969,0.002814269,0.03733344],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010400782,0.00016004781,0.0051620556,0.0038003507,0.0003121534,0.0006419606,0.00095745386,0.0022996443,0.020524664,0.015824892,0.8982358,0.051040914],"study_design_scores_gemma":[0.00030023966,0.000048017962,0.005958024,0.00038805872,0.00015092119,0.00027209477,0.0002542588,0.002129147,0.007896246,0.009882139,0.97259074,0.00012998846],"about_ca_topic_score_codex":0.011750264,"about_ca_topic_score_gemma":0.018088298,"teacher_disagreement_score":0.99549294,"about_ca_system_score_codex":0.0018298519,"about_ca_system_score_gemma":0.0049115075,"threshold_uncertainty_score":0.08932525},"labels":[],"label_agreement":null},{"id":"W4297260078","doi":"10.1093/database/baac088","title":"ChemBioPort: an online portal to navigate the structure, function and chemical inhibition of the human proteome","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada Research Chairs; University of New Brunswick; Structural Genomics Consortium; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Mitacs; Québec Consortium for Drug Discovery","keywords":"Druggability; Computer science; Function (biology); Domain (mathematical analysis); Chemical database; Human proteome project; Ontology; Architecture; Proteome; Computational biology; Information retrieval; Bioinformatics; Chemistry; Biology; Proteomics; Biochemistry","score_opus":0.029578339846526556,"score_gpt":0.29806326734176203,"score_spread":0.26848492749523545,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4297260078","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0045503727,0.0049532703,0.05812339,0.0019981593,0.00046228504,0.00038530133,0.6943542,0.19389103,0.04128201],"genre_scores_gemma":[0.024889078,0.006549307,0.08905111,0.002108081,0.00019679639,0.0011519896,0.8240487,0.026276344,0.0257285],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9994653,0.00008942536,0.000060746206,0.00010592547,0.00022275139,0.00005583579],"domain_scores_gemma":[0.9983759,0.0007247676,0.00017568271,0.00026922618,0.00018704127,0.000267341],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017347769,0.0019386506,0.0013928422,0.0033643895,0.0006139819,0.003061645,0.0025707728,0.001770753,0.14837132],"category_scores_gemma":[0.003717104,0.0010027321,0.0008614164,0.0041817278,0.0003578105,0.003694692,0.0039116265,0.001616182,0.077359214],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00035570704,0.000089753885,0.0011721329,0.0013695522,0.0001016275,0.00020878995,0.000086550725,0.00047052372,0.0025510122,0.0038857562,0.942211,0.047497462],"study_design_scores_gemma":[0.00029171552,0.000054281747,0.0018720258,0.00026858822,0.000052044918,0.00047113173,0.00005967998,0.0029581964,0.0071799597,0.00705609,0.9796683,0.00006799438],"about_ca_topic_score_codex":0.0026816817,"about_ca_topic_score_gemma":0.0044905753,"teacher_disagreement_score":0.14837132,"about_ca_system_score_codex":0.0009962659,"about_ca_system_score_gemma":0.0019628978,"threshold_uncertainty_score":0.49635154},"labels":[],"label_agreement":null},{"id":"W4302275377","doi":"10.1093/database/baac084","title":"Overview of the COVID-19 text mining tool interactive demonstration track in BioCreative VII","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Institute for Research in Immunology and Cancer","funders":"U.S. National Library of Medicine; National Institute of General Medical Sciences; National Human Genome Research Institute; Canadian Institutes of Health Research; National Institutes of Health","keywords":"Coronavirus disease 2019 (COVID-19); Computer science; Track (disk drive); Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); 2019-20 coronavirus outbreak; Information retrieval; Natural language processing; Artificial intelligence; World Wide Web; Virology; Medicine","score_opus":0.05752538837657176,"score_gpt":0.34534285423608096,"score_spread":0.2878174658595092,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4302275377","genre_codex":"software","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.022978513,0.0014360823,0.35889694,0.0014966483,0.0007053954,0.005642165,0.09800353,0.46152675,0.04931394],"genre_scores_gemma":[0.053225115,0.00105182,0.6056722,0.0015344328,0.00028163023,0.008377775,0.21160801,0.045988392,0.07226072],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99741346,0.00046763447,0.0003375694,0.00058853644,0.0010091335,0.00018369887],"domain_scores_gemma":[0.9917549,0.005209488,0.00028000027,0.00079377944,0.0012777062,0.00068422913],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0048654475,0.0018562152,0.0010910326,0.0042309244,0.0010782768,0.0041075004,0.004466458,0.002000463,0.08958161],"category_scores_gemma":[0.010109628,0.0014260419,0.001485525,0.0022789466,0.00050388306,0.004605745,0.0027993466,0.0019138933,0.04419575],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0020786885,0.0015404074,0.006586143,0.003352842,0.0002397862,0.002019366,0.0027269567,0.00338381,0.032061554,0.004220346,0.5974427,0.3443474],"study_design_scores_gemma":[0.0008245967,0.0010771417,0.009266864,0.0007348295,0.00010968974,0.0016073092,0.0005503328,0.04023557,0.024903271,0.0033575161,0.9170151,0.00031779334],"about_ca_topic_score_codex":0.005144983,"about_ca_topic_score_gemma":0.0063827806,"teacher_disagreement_score":0.08958161,"about_ca_system_score_codex":0.0008259601,"about_ca_system_score_gemma":0.0017038633,"threshold_uncertainty_score":0.29968035},"labels":[],"label_agreement":null},{"id":"W4303685210","doi":"10.1093/database/baac086","title":"HSDatabase—a database of highly similar duplicate genes from plants, animals, and algae","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University; Dalhousie University","funders":"","keywords":"Algae; Gene; Database; Biology; Computational biology; Computer science; Botany; Genetics","score_opus":0.016419549707939383,"score_gpt":0.23257291969946828,"score_spread":0.2161533699915289,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4303685210","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.016022451,0.0060172845,0.011152071,0.00016893966,0.00036961766,0.00024982027,0.94374526,0.014888435,0.0073860846],"genre_scores_gemma":[0.004623722,0.001292197,0.015301679,0.000075087744,0.00003515411,0.00015938246,0.97653025,0.00091998075,0.0010624591],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9989643,0.00007719544,0.000264209,0.00032110608,0.00027736777,0.000095822936],"domain_scores_gemma":[0.9984517,0.00025307445,0.0003828642,0.00024700083,0.0003941668,0.00027114156],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012244671,0.0022973893,0.0027245835,0.009270991,0.0021680756,0.0024828233,0.0028150144,0.001720489,0.018477298],"category_scores_gemma":[0.0035402807,0.0012435741,0.0013598663,0.012348975,0.0005470359,0.002928047,0.0026966326,0.0019285529,0.02604697],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0038644867,0.0004314221,0.016092865,0.02480276,0.0007671251,0.0025954044,0.0021893745,0.0022437032,0.1508184,0.007364376,0.61876214,0.17006794],"study_design_scores_gemma":[0.00026029133,0.00021486278,0.016084425,0.00075499323,0.00032316428,0.0017466423,0.00035181767,0.0014480358,0.01556115,0.0030198377,0.9600403,0.00019456791],"about_ca_topic_score_codex":0.0039852993,"about_ca_topic_score_gemma":0.0042260718,"teacher_disagreement_score":0.018477298,"about_ca_system_score_codex":0.0010791725,"about_ca_system_score_gemma":0.0040707276,"threshold_uncertainty_score":0.0618127},"labels":[],"label_agreement":null},{"id":"W4304757261","doi":"10.1093/database/baac081","title":"KinMod database: a tool for investigating metabolic regulation","year":2022,"lang":"en","type":"article","venue":"Database","topic":"Microbial Metabolic Engineering and Bioproduction","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Ontology; Homogeneous; Data mining; Missing data; Metabolic network; Systems biology; Database; Computational biology; Machine learning; Biology; Mathematics","score_opus":0.015007313350069081,"score_gpt":0.24721977959840397,"score_spread":0.2322124662483349,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4304757261","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.013444031,0.0018440089,0.07565435,0.00027821952,0.00016130775,0.00027738075,0.78250563,0.117414884,0.008420232],"genre_scores_gemma":[0.034734234,0.0016382523,0.09486992,0.000260811,0.00004289203,0.00081465207,0.85988474,0.0057489886,0.0020054749],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9991466,0.000100517515,0.00018137829,0.00025730053,0.0002477058,0.00006638912],"domain_scores_gemma":[0.9981323,0.0006809731,0.0003344298,0.00044046395,0.00027244806,0.00013935725],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013932355,0.0020882823,0.0017013531,0.0057204827,0.0007430642,0.0024785826,0.0030840193,0.0015116457,0.022453008],"category_scores_gemma":[0.0043622516,0.0010275793,0.0016661072,0.0063603725,0.00030669567,0.0027960488,0.002212705,0.0014576981,0.012940003],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0040584863,0.0007037005,0.032836128,0.014043309,0.0009884222,0.0028783015,0.0011304241,0.032827243,0.040082034,0.03175053,0.6475698,0.19113164],"study_design_scores_gemma":[0.00045691335,0.0002180477,0.021715144,0.0008007601,0.00030818646,0.001800101,0.0005078413,0.07636275,0.032847416,0.032032438,0.8326184,0.00033208347],"about_ca_topic_score_codex":0.0035606814,"about_ca_topic_score_gemma":0.0036199943,"teacher_disagreement_score":0.022453008,"about_ca_system_score_codex":0.00091750844,"about_ca_system_score_gemma":0.0017913551,"threshold_uncertainty_score":0.07511282},"labels":[],"label_agreement":null},{"id":"W4313706978","doi":"10.1093/database/baac109","title":"Developing a disease-specific annotation protocol for <i>VHL</i> gene curation using Hypothes.is","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Genomics and Rare Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Sinai Health System; Princess Margaret Cancer Centre; Canada Research Chairs; Hospital for Sick Children; University Health Network; Toronto General Hospital; University of New Brunswick; University of Toronto; Ontario Institute for Cancer Research","funders":"National Human Genome Research Institute; VHL Alliance; Princess Margaret Cancer Foundation","keywords":"Annotation; Data curation; Protocol (science); Computer science; Disease; Computational biology; Biology; World Wide Web; Bioinformatics; Medicine; Pathology","score_opus":0.1263615011844818,"score_gpt":0.3555056672103924,"score_spread":0.2291441660259106,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4313706978","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012766732,0.0012161206,0.60906154,0.0055010314,0.0030381712,0.012477165,0.20255564,0.1119565,0.041427135],"genre_scores_gemma":[0.022425508,0.00070977537,0.5819163,0.0035904464,0.00041153838,0.019934969,0.3316391,0.021121282,0.018251099],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9852645,0.0037212274,0.0033682222,0.0033813342,0.003442065,0.00082263094],"domain_scores_gemma":[0.9597691,0.012679062,0.0021219747,0.009909475,0.013789222,0.0017310185],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.02285855,0.0014804482,0.0011999495,0.008532689,0.003596677,0.003566413,0.0030682234,0.0019077309,0.041407388],"category_scores_gemma":[0.047201946,0.0013709521,0.0018683723,0.0041141724,0.0013274179,0.0029611615,0.009644828,0.0034380269,0.03590464],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0008513096,0.00021219107,0.005237592,0.0036751232,0.00019717608,0.0016356816,0.0050145644,0.0013898276,0.06630998,0.012585586,0.77268606,0.13020493],"study_design_scores_gemma":[0.00019585415,0.0001061021,0.005756054,0.0009314533,0.000100899226,0.00086914067,0.001150041,0.0050228806,0.025271706,0.012697427,0.9476763,0.00022218915],"about_ca_topic_score_codex":0.007525464,"about_ca_topic_score_gemma":0.016378747,"teacher_disagreement_score":0.97714144,"about_ca_system_score_codex":0.00215746,"about_ca_system_score_gemma":0.008833938,"threshold_uncertainty_score":0.13852149},"labels":[],"label_agreement":null},{"id":"W4319062633","doi":"10.1093/database/baac108","title":"Automatic Extraction of Medication Mentions from Tweets—Overview of the BioCreative VII Shared Task 3 Competition","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Topic Modeling","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"U.S. National Library of Medicine","keywords":"Timeline; Computer science; Task (project management); Natural language processing; Artificial intelligence; Class (philosophy); Information retrieval; World Wide Web","score_opus":0.06896966406581621,"score_gpt":0.32966677997371774,"score_spread":0.2606971159079015,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4319062633","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.23910318,0.018506227,0.101561256,0.014266493,0.0072061825,0.009696382,0.49079096,0.08224292,0.036626376],"genre_scores_gemma":[0.04804764,0.0012855891,0.09225586,0.0017127909,0.0006902334,0.005232583,0.83587146,0.0048865033,0.010017293],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98589176,0.0043043764,0.0013714862,0.0035820305,0.003554755,0.0012956532],"domain_scores_gemma":[0.9817442,0.0070872274,0.000913395,0.002997856,0.005191868,0.0020653855],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.017024947,0.005010755,0.0035256306,0.0073593785,0.0054556644,0.005365973,0.0057120137,0.0048898514,0.0084301075],"category_scores_gemma":[0.022108763,0.0015204897,0.0037105952,0.0054175216,0.0017267723,0.004106556,0.00931365,0.0042882487,0.016842458],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0027895952,0.0018156077,0.00889563,0.0049155112,0.00070593756,0.0017740236,0.0021528124,0.0038457206,0.048015453,0.0021121784,0.7245072,0.19847038],"study_design_scores_gemma":[0.002447852,0.0020241085,0.051277254,0.0008238924,0.0007657585,0.0024385334,0.0044012433,0.06941837,0.07171553,0.007688433,0.78636247,0.00063655357],"about_ca_topic_score_codex":0.019928603,"about_ca_topic_score_gemma":0.03728904,"teacher_disagreement_score":0.019928603,"about_ca_system_score_codex":0.003681621,"about_ca_system_score_gemma":0.0080420915,"threshold_uncertainty_score":0.090037584},"labels":[],"label_agreement":null},{"id":"W4319825965","doi":"10.1093/database/baac112","title":"Coronavirus Immunotherapeutic Consortium Database","year":2023,"lang":"en","type":"article","venue":"Database","topic":"SARS-CoV-2 and COVID-19 Research","field":"Medicine","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Fast Grants; GHR Foundation; Mastercard Foundation; Wellcome Trust; National Institute of Allergy and Infectious Diseases; Bill and Melinda Gates Foundation","keywords":"Coronavirus disease 2019 (COVID-19); Coronavirus; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); 2019-20 coronavirus outbreak; Virology; Database; Computer science; World Wide Web; Library science; Medicine; Infectious disease (medical specialty); Outbreak; Pathology","score_opus":0.1494326824191115,"score_gpt":0.4207830660524239,"score_spread":0.2713503836333124,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4319825965","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0004745749,0.00096812577,0.00040576846,0.00019017622,0.00003546441,0.00007523692,0.99468493,0.00061522826,0.002550555],"genre_scores_gemma":[0.0011048447,0.0007486766,0.00096414436,0.00019788947,0.000013987272,0.0001832608,0.9962037,0.00007699217,0.0005065164],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9983742,0.00024241928,0.00042888068,0.0003748488,0.00042368646,0.0001558758],"domain_scores_gemma":[0.99613285,0.0011235344,0.00070376735,0.00063222804,0.0008421386,0.00056536566],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017351998,0.0014827409,0.0024274222,0.0058201854,0.00077917933,0.002891141,0.0028795525,0.0021237067,0.045759592],"category_scores_gemma":[0.007693746,0.00045983904,0.0011377725,0.00977772,0.00032290656,0.0011199676,0.0019300587,0.0016779166,0.034571826],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00057311676,0.00009240765,0.0030585895,0.0065037096,0.00016398834,0.00016832545,0.00006091055,0.0008780497,0.0014229867,0.0024653263,0.96298856,0.021623941],"study_design_scores_gemma":[0.00025879618,0.00005338433,0.0034966655,0.0008756988,0.00010821679,0.00017982193,0.00004542199,0.0006988977,0.00085051835,0.0017792452,0.9916224,0.0000308877],"about_ca_topic_score_codex":0.0067937244,"about_ca_topic_score_gemma":0.008435402,"teacher_disagreement_score":0.045759592,"about_ca_system_score_codex":0.0015508272,"about_ca_system_score_gemma":0.0041638017,"threshold_uncertainty_score":0.15308106},"labels":[],"label_agreement":null},{"id":"W4323519786","doi":"10.1093/database/baad005","title":"Chemical identification and indexing in full-text articles: an overview of the NLM-Chem track at BioCreative VII","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"H2020 Marie Skłodowska-Curie Actions; U.S. National Library of Medicine; Fundação para a Ciência e a Tecnologia; Natural Sciences and Engineering Research Council of Canada; European Commission; University of Oxford; Albaha University; National Institutes of Health; Nvidia","keywords":"Computer science; Search engine indexing; Named-entity recognition; Identification (biology); Information retrieval; Natural language processing; Task (project management); Normalization (sociology); Artificial intelligence","score_opus":0.06945455733307294,"score_gpt":0.3394277437636137,"score_spread":0.26997318643054075,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4323519786","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.039046157,0.042066477,0.2854689,0.01134251,0.009719356,0.009462548,0.26817256,0.27306092,0.0616606],"genre_scores_gemma":[0.025926676,0.008824418,0.341594,0.002278804,0.0019244646,0.003363805,0.5512696,0.017660294,0.04715796],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.97117525,0.005458975,0.004570467,0.0053793103,0.011975932,0.0014400429],"domain_scores_gemma":[0.92494494,0.019710723,0.0053402185,0.012188633,0.030755932,0.0070595494],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.041788716,0.0034959638,0.0034113163,0.029708823,0.0037559094,0.013816978,0.0059196856,0.0038300431,0.03323917],"category_scores_gemma":[0.049935557,0.00226855,0.0042811627,0.01986606,0.0011761334,0.015032706,0.008765157,0.0033885916,0.070258155],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00083313725,0.0008148982,0.006158555,0.0052013854,0.00054337754,0.0004501589,0.0008034177,0.0025312952,0.019172555,0.0019748975,0.5349363,0.42658013],"study_design_scores_gemma":[0.00040050177,0.001247764,0.01471431,0.0012761527,0.00026939483,0.0015167107,0.00055565895,0.026843755,0.038201995,0.0036552139,0.910898,0.00042057593],"about_ca_topic_score_codex":0.0060010552,"about_ca_topic_score_gemma":0.009404777,"teacher_disagreement_score":0.9582113,"about_ca_system_score_codex":0.003936198,"about_ca_system_score_gemma":0.008512912,"threshold_uncertainty_score":0.22100246},"labels":[],"label_agreement":null},{"id":"W4362515160","doi":"10.1093/database/baad016","title":"PurificationDB: database of purification conditions for proteins","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Canadian Institutes of Health Research","keywords":"UniProt; Computer science; Database; Function (biology); Column (typography); Information retrieval; Chemistry; Biology","score_opus":0.0200547654024743,"score_gpt":0.2963510366646304,"score_spread":0.2762962712621561,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4362515160","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0013621034,0.0076503106,0.02528965,0.0005016239,0.00026961227,0.0004922598,0.91813767,0.037245765,0.009050965],"genre_scores_gemma":[0.0011160944,0.0025433425,0.01815498,0.00027724245,0.000025875292,0.0003313947,0.974606,0.0017641151,0.0011809387],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99734855,0.00032782758,0.0007329777,0.0005981794,0.0007802168,0.00021223059],"domain_scores_gemma":[0.99492115,0.0011562839,0.00090796984,0.0008582028,0.001753454,0.0004030264],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029160718,0.005212747,0.0051224898,0.009248908,0.0021330998,0.005518036,0.0072020544,0.0037648755,0.045730785],"category_scores_gemma":[0.010562841,0.0022236311,0.0021316092,0.013316956,0.00075423456,0.005828745,0.0036298663,0.0033181515,0.09954935],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0006365538,0.00018296226,0.0013999116,0.012946596,0.0002824593,0.0003310012,0.0001998224,0.0011282819,0.00902203,0.005346449,0.9244883,0.044035707],"study_design_scores_gemma":[0.00016722355,0.00006107728,0.0016766102,0.0010158144,0.00017492501,0.0004161081,0.0000703093,0.0012386964,0.0054950053,0.0041741594,0.9853977,0.00011226408],"about_ca_topic_score_codex":0.006682705,"about_ca_topic_score_gemma":0.0054171747,"teacher_disagreement_score":0.045730785,"about_ca_system_score_codex":0.0026307907,"about_ca_system_score_gemma":0.009163815,"threshold_uncertainty_score":0.15298474},"labels":[],"label_agreement":null},{"id":"W4365135058","doi":"10.1093/database/baad021","title":"NbThermo: a new thermostability database for nanobodies","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Monoclonal and Polyclonal Antibodies Research","field":"Medicine","cited_by":39,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"","keywords":"Thermostability; Database; Computer science; Melting temperature; Chemistry; Materials science; Biochemistry","score_opus":0.11078268293147928,"score_gpt":0.3938353857990054,"score_spread":0.28305270286752615,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4365135058","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.019465603,0.013499477,0.054128584,0.00036767474,0.00026777416,0.00033664602,0.8032154,0.09859713,0.010121784],"genre_scores_gemma":[0.017423961,0.003624172,0.03939733,0.00031688964,0.00004280646,0.00074338436,0.9318912,0.0045211,0.0020391298],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.999183,0.000074482676,0.0002113034,0.00021275341,0.0002465346,0.00007192164],"domain_scores_gemma":[0.998789,0.00028835252,0.00032006137,0.00027457197,0.00019612037,0.00013195246],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011434653,0.0022127104,0.0025011566,0.005632596,0.0006894242,0.0023861853,0.003045434,0.0019099215,0.013387911],"category_scores_gemma":[0.003348531,0.0009251279,0.0014330434,0.00506337,0.00034241486,0.0025375169,0.002387124,0.0016654672,0.019736962],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.004411415,0.0004739144,0.010684992,0.021362871,0.00080009375,0.0011499445,0.00077659456,0.0066795596,0.27264428,0.011469274,0.4543134,0.21523367],"study_design_scores_gemma":[0.00030388022,0.00022832233,0.0075248284,0.0007428761,0.00031273864,0.0009323839,0.00010215696,0.004231927,0.036615703,0.005868768,0.94297844,0.0001578857],"about_ca_topic_score_codex":0.0012542218,"about_ca_topic_score_gemma":0.0015429783,"teacher_disagreement_score":0.013387911,"about_ca_system_score_codex":0.0009856838,"about_ca_system_score_gemma":0.0016378079,"threshold_uncertainty_score":0.04478705},"labels":[],"label_agreement":null},{"id":"W4366492899","doi":"10.1093/database/baad023","title":"CARD*Shark: automated prioritization of literature curation for the Comprehensive Antibiotic Resistance Database","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Antibiotic Resistance in Bacteria","field":"Biochemistry, Genetics and Molecular Biology","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McMaster University","funders":"Canadian Institutes of Health Research; Cisco Systems Canada; Canada Foundation for Innovation; McMaster University; Cisco Systems","keywords":"Computer science; Database; Scope (computer science); Prioritization; Task (project management); Information retrieval; Triage; World Wide Web; Data science; Medicine; Business","score_opus":0.0199321420995732,"score_gpt":0.30137823407599623,"score_spread":0.281446091976423,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4366492899","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.010321583,0.0035624232,0.06148912,0.0018668345,0.0006503733,0.003664288,0.7747621,0.120467216,0.023216177],"genre_scores_gemma":[0.018944008,0.0023919265,0.32171094,0.00083683676,0.0002986874,0.0037284107,0.6357222,0.0076499353,0.008717063],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99446267,0.0010235785,0.0013394961,0.0010558537,0.0017777368,0.00034068478],"domain_scores_gemma":[0.97182995,0.011445156,0.0033371148,0.0041553066,0.00741807,0.0018143561],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.007104265,0.002295293,0.0027412937,0.027270488,0.0019972154,0.0068137175,0.0027497527,0.00109429,0.05431666],"category_scores_gemma":[0.038758744,0.0013139478,0.001865973,0.018559445,0.0006727092,0.0042034113,0.0068708677,0.0014084311,0.036794063],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007646186,0.00014622493,0.005149499,0.009658682,0.00037362037,0.0004957484,0.00046482024,0.0006895459,0.006178774,0.004305182,0.7893844,0.18238884],"study_design_scores_gemma":[0.000737078,0.0002234227,0.018950565,0.0019803094,0.00043141487,0.001075553,0.0005637378,0.012042278,0.01751095,0.007018618,0.93915737,0.000308654],"about_ca_topic_score_codex":0.0060989666,"about_ca_topic_score_gemma":0.01398694,"teacher_disagreement_score":0.9928957,"about_ca_system_score_codex":0.0018370958,"about_ca_system_score_gemma":0.010075385,"threshold_uncertainty_score":0.18170732},"labels":[],"label_agreement":null},{"id":"W4376224614","doi":"10.1093/database/baad030","title":"PYK-SubstitutionOME: an integrated database containing allosteric coupling, ligand affinity and mutational, structural, pathological, bioinformatic and computational information about pyruvate kinase isozymes","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Basic Energy Sciences; National Center for Research Resources; National Institute of General Medical Sciences; National Institute on Aging; Natural Sciences and Engineering Research Council of Canada; Office of Science; National Institutes of Health; W. M. Keck Foundation; U.S. Department of Energy","keywords":"Allosteric regulation; Computational biology; Function (biology); Computer science; Genome; Database; Biology; Bioinformatics; Biochemistry; Genetics; Gene; Enzyme","score_opus":0.015264263407754641,"score_gpt":0.2664963883285539,"score_spread":0.2512321249207992,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4376224614","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01080403,0.0015243043,0.007531543,0.00021068192,0.0000757685,0.00008589962,0.95553017,0.020981645,0.0032560055],"genre_scores_gemma":[0.017798495,0.0009380275,0.00801437,0.000102572085,0.000013420262,0.00016649536,0.971085,0.0011748171,0.0007067892],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9993623,0.00007327041,0.00010697476,0.00019980542,0.00018536963,0.0000722985],"domain_scores_gemma":[0.998703,0.00030308028,0.00025290385,0.0003421965,0.00017681939,0.00022206077],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00084665325,0.0023287737,0.0020212778,0.0029932957,0.0005831701,0.0014478961,0.0029654978,0.0019195371,0.021836407],"category_scores_gemma":[0.004010815,0.00084899476,0.001079134,0.0049613668,0.00044307925,0.0026112997,0.0025992352,0.0019399788,0.018762967],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0042467914,0.00039157746,0.018238552,0.0136660915,0.00067346904,0.0013636365,0.0002461483,0.012611142,0.034708794,0.011141337,0.8197581,0.08295429],"study_design_scores_gemma":[0.0012189542,0.0003610614,0.027945952,0.00056141766,0.0003806519,0.002291869,0.00017143947,0.024994256,0.026807863,0.012772599,0.9022242,0.00026972525],"about_ca_topic_score_codex":0.0035510308,"about_ca_topic_score_gemma":0.0047756713,"teacher_disagreement_score":0.021836407,"about_ca_system_score_codex":0.0013372347,"about_ca_system_score_gemma":0.002218693,"threshold_uncertainty_score":0.07305002},"labels":[],"label_agreement":null},{"id":"W4377862296","doi":"10.1093/database/baad038","title":"CIGAF—a database and interactive platform for insect-associated trichomycete fungi","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Infectious Diseases and Mycology","field":"Veterinary","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; University of Toronto","keywords":"Metadata; Biology; Data collection; World Wide Web; Field (mathematics); Ecology; Database; Computer science","score_opus":0.08022745347238294,"score_gpt":0.3457302355413059,"score_spread":0.26550278206892297,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4377862296","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012789083,0.0035284306,0.04628344,0.0006026206,0.00023386949,0.0006036396,0.7986835,0.121025845,0.016249564],"genre_scores_gemma":[0.041481715,0.0026320452,0.098685615,0.00050784077,0.00015543964,0.0010910888,0.8412367,0.009904857,0.004304731],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99864966,0.00018006591,0.00031912499,0.00035852296,0.0003550332,0.0001376451],"domain_scores_gemma":[0.9948219,0.0015780452,0.00077936216,0.0012496252,0.0008647115,0.0007064188],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025377858,0.0013768859,0.0014388943,0.011304413,0.0008566244,0.0038172582,0.0027959703,0.0010629927,0.023012685],"category_scores_gemma":[0.010789925,0.00067610765,0.0010772567,0.010882538,0.0004272203,0.0040050456,0.005254732,0.0012508174,0.015051583],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0014206729,0.00015520753,0.017528268,0.005468119,0.00039160682,0.0009022895,0.0014816158,0.002096557,0.008820414,0.012901739,0.7117151,0.23711847],"study_design_scores_gemma":[0.00025611927,0.000087871056,0.024177335,0.0007647559,0.00020189297,0.0010027387,0.0005046469,0.0056537213,0.0055619883,0.008338922,0.95322764,0.00022237812],"about_ca_topic_score_codex":0.00758624,"about_ca_topic_score_gemma":0.008846774,"teacher_disagreement_score":0.023012685,"about_ca_system_score_codex":0.00085733394,"about_ca_system_score_gemma":0.002273121,"threshold_uncertainty_score":0.07698512},"labels":[],"label_agreement":null},{"id":"W4385969415","doi":"10.1093/database/baad055","title":"Tissue-specific transcriptomes reveal potential mechanisms of microbiome heterogeneity in an ancient fish","year":2023,"lang":"en","type":"article","venue":"Database","topic":"Cancer-related molecular mechanisms research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Manitoba","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada; Manitoba Hydro","keywords":"Biology; Transcriptome; Vertebrate; Cecum; Microbiome; De novo transcriptome assembly; Sturgeon; Model organism; Gene; Zoology; Evolutionary biology; Genetics; Fish <Actinopterygii>; Gene expression; Ecology; Fishery","score_opus":0.023523564965304996,"score_gpt":0.30154293962552803,"score_spread":0.27801937466022303,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4385969415","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9951125,0.0003302956,0.0030199988,0.000048515154,0.0000070070705,0.000006193952,0.0007800873,0.000044830387,0.0006507166],"genre_scores_gemma":[0.99197704,0.00023138618,0.0042473283,0.00013588657,0.000009603586,0.00003466117,0.0022685116,0.000043608223,0.0010519455],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99986374,0.000010880272,0.000008983646,0.00006618767,0.000025364116,0.000024886342],"domain_scores_gemma":[0.99984646,0.000040101684,0.000045317043,0.000016359536,0.00003073067,0.000021065642],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00022121466,0.00018474828,0.00026057943,0.00056437234,0.00031500804,0.00036011753,0.00014196962,0.00030897162,0.0007908236],"category_scores_gemma":[0.00035341247,0.00024808245,0.00032687926,0.00046240294,0.00033178274,0.0003592195,0.0005481147,0.000310203,0.00024032622],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009125115,0.00000780611,0.021959886,0.000076311204,0.000029907447,0.00011717036,0.0004025272,0.00022279921,0.9729759,0.0001743577,0.000086919885,0.0038551579],"study_design_scores_gemma":[0.000008010255,0.00016345603,0.9619466,0.000029178605,0.00009267301,0.0004382323,0.00054923916,0.0024790324,0.030798512,0.0005337197,0.0029352277,0.000026033445],"about_ca_topic_score_codex":0.0010419164,"about_ca_topic_score_gemma":0.0031216922,"teacher_disagreement_score":0.0010419164,"about_ca_system_score_codex":0.00020446752,"about_ca_system_score_gemma":0.00021260843,"threshold_uncertainty_score":0.0026456118},"labels":[],"label_agreement":null},{"id":"W4387479018","doi":"10.1093/database/baad063","title":"ESOMIR: a curated database of biomarker genes and miRNAs associated with esophageal cancer","year":2023,"lang":"en","type":"article","venue":"Database","topic":"MicroRNA in disease regulation","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"Shanghai Jiao Tong University; Science and Technology Commission of Shanghai Municipality; King Saud University; National Natural Science Foundation of China; National Science Foundation","keywords":"microRNA; Database; Gene; Computational biology; Biomarker; Bioinformatics; Biology; Computer science; Genetics","score_opus":0.021802113412344958,"score_gpt":0.28589287277100256,"score_spread":0.2640907593586576,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4387479018","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.023858765,0.013256907,0.00953241,0.00033574458,0.00016337678,0.0003079093,0.93833995,0.009859729,0.0043453155],"genre_scores_gemma":[0.0234796,0.0038834675,0.018461285,0.00035456067,0.000045458855,0.00054922135,0.95138997,0.00053867244,0.0012977376],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9989261,0.00012030722,0.00027682207,0.00037045943,0.00021441068,0.0000918021],"domain_scores_gemma":[0.9989452,0.00035658857,0.00022729194,0.00014533565,0.00016658111,0.00015895555],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008993179,0.0016917685,0.0021964847,0.0067282547,0.000581612,0.0014610348,0.0018061505,0.0016959897,0.009942016],"category_scores_gemma":[0.0038334539,0.0005201339,0.0012235338,0.0059053665,0.00026792972,0.0010481806,0.0020239132,0.00077993446,0.006889152],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.006294859,0.00045830957,0.04978182,0.038666632,0.0018703658,0.0077032465,0.0010726235,0.008833827,0.08365865,0.0068348097,0.5211028,0.27372217],"study_design_scores_gemma":[0.0007124179,0.00036899635,0.06582217,0.0019020843,0.0013083704,0.005724641,0.00035363078,0.007877934,0.027029341,0.0044942973,0.884129,0.00027718532],"about_ca_topic_score_codex":0.0032942288,"about_ca_topic_score_gemma":0.0054979995,"teacher_disagreement_score":0.009942016,"about_ca_system_score_codex":0.00061053183,"about_ca_system_score_gemma":0.001981517,"threshold_uncertainty_score":0.03325939},"labels":[],"label_agreement":null},{"id":"W4389579251","doi":"10.1093/database/baad088","title":"Genotype and phenotype data standardization, utilization and integration in the big data era for agricultural sciences","year":2023,"lang":"en","type":"review","venue":"Database","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":32,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"National Research Council Canada; Saskatchewan Research Council (Canada)","funders":"Réseau de cancérologie Rossy; U.S. Department of Agriculture; National Institute of Food and Agriculture; National Science Foundation","keywords":"Data science; Raw data; Context (archaeology); Standardization; Computer science; Metadata; Annotation; Data collection; Biology; World Wide Web; Bioinformatics","score_opus":0.298928971651219,"score_gpt":0.3947060009767093,"score_spread":0.09577702932549031,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4389579251","genre_codex":"methods","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03971143,0.033306155,0.5928416,0.06804824,0.005048602,0.002004835,0.20821705,0.023033373,0.027788633],"genre_scores_gemma":[0.06490731,0.01971478,0.59485525,0.011566014,0.0015845929,0.0028510082,0.29547855,0.0060941493,0.0029482788],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9759942,0.008639233,0.0044787596,0.0042297575,0.005683137,0.00097480015],"domain_scores_gemma":[0.8444838,0.041785084,0.008656894,0.06683469,0.03236786,0.0058716694],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.05321373,0.0012889936,0.0020052763,0.009812907,0.0020260264,0.011686482,0.0048311567,0.0025660146,0.0033566325],"category_scores_gemma":[0.10564539,0.0009612353,0.0024966053,0.018683502,0.002592642,0.020450016,0.0119411815,0.0060349004,0.003344132],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00059860275,0.00024580385,0.047612246,0.009055054,0.0006364083,0.00078689645,0.005266772,0.005391853,0.018888192,0.15539795,0.23261781,0.52350247],"study_design_scores_gemma":[0.00005826111,0.0000706922,0.023101814,0.0037972436,0.00026493688,0.00034990188,0.002671994,0.0044360417,0.0074454625,0.12355142,0.8340174,0.00023474237],"about_ca_topic_score_codex":0.006570322,"about_ca_topic_score_gemma":0.00626992,"teacher_disagreement_score":0.05321373,"about_ca_system_score_codex":0.0032133637,"about_ca_system_score_gemma":0.012014067,"threshold_uncertainty_score":0.28142446},"labels":[],"label_agreement":null},{"id":"W4396685747","doi":"10.1093/database/baae031","title":"Pathway-based, reaction-specific annotation of disease variants for elucidation of molecular phenotypes","year":2024,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada Research Chairs; University of New Brunswick; University of Toronto; Ontario Institute for Cancer Research","funders":"National Institute of General Medical Sciences; National Institutes of Health; National Human Genome Research Institute; University of Toronto; European Commission; European Bioinformatics Institute","keywords":"Context (archaeology); Disease; Phenotype; Computational biology; Biology; Annotation; Genetics; Functional genomics; Clinical phenotype; Genomics; Gene; Genome; Medicine","score_opus":0.011007310499580822,"score_gpt":0.24289168076111675,"score_spread":0.23188437026153594,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4396685747","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.090625815,0.004857026,0.5530877,0.0009448489,0.00046376855,0.0007037097,0.2946375,0.04342871,0.0112509085],"genre_scores_gemma":[0.2682495,0.0040868735,0.4402732,0.00034061828,0.000105448154,0.0007046439,0.27787974,0.0036375418,0.00472241],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988544,0.00014717715,0.00017996908,0.00047283457,0.00027274244,0.00007290996],"domain_scores_gemma":[0.9975425,0.000818326,0.00042142376,0.0004895259,0.0005801572,0.00014809697],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015319036,0.0016218927,0.0009768696,0.0042068157,0.0007316916,0.0016375085,0.0013426533,0.0008778798,0.013148852],"category_scores_gemma":[0.004583914,0.0005285013,0.0029188334,0.004270825,0.0003850589,0.0016303174,0.0017960133,0.0013332379,0.0035829693],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0037724546,0.00046622104,0.13236143,0.016104281,0.002083551,0.0067900633,0.0022525568,0.1142376,0.26947963,0.07728425,0.10817532,0.26699266],"study_design_scores_gemma":[0.00031700425,0.0003558988,0.054940656,0.0013204745,0.0015462659,0.0038667375,0.0008733307,0.34662667,0.114382304,0.088648796,0.38673466,0.00038725958],"about_ca_topic_score_codex":0.006695502,"about_ca_topic_score_gemma":0.010148775,"teacher_disagreement_score":0.013148852,"about_ca_system_score_codex":0.0011632888,"about_ca_system_score_gemma":0.0021547945,"threshold_uncertainty_score":0.043987274},"labels":[],"label_agreement":null},{"id":"W4399484747","doi":"10.1093/database/baae014","title":"The Immunopeptidomics Ontology (ImPO)","year":2024,"lang":"en","type":"article","venue":"Database","topic":"vaccines and immunoinformatics approaches","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria","funders":"Fundação para a Ciência e a Tecnologia; Fundacja na rzecz Nauki Polskiej","keywords":"Computer science; Ontology; Information retrieval; World Wide Web; Software engineering; Philosophy; Epistemology","score_opus":0.011308519225832094,"score_gpt":0.24988501513391972,"score_spread":0.23857649590808763,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399484747","genre_codex":"methods","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.029541988,0.0076086386,0.7362913,0.006933928,0.0011742191,0.0024380754,0.1364797,0.020637866,0.058894224],"genre_scores_gemma":[0.09743248,0.00963244,0.6772314,0.003177437,0.00041326304,0.0022693775,0.20061742,0.0022939178,0.0069322987],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9967912,0.00060239126,0.00081630866,0.00055558723,0.0009853663,0.00024922664],"domain_scores_gemma":[0.9942542,0.002163484,0.0007624202,0.0011095974,0.0012206899,0.0004896214],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0045042024,0.00090708974,0.00074165076,0.008407394,0.0017246642,0.0038416258,0.002045356,0.0016808369,0.0020628201],"category_scores_gemma":[0.010027998,0.00065299263,0.0018234288,0.0075945156,0.0015559572,0.007418043,0.004098734,0.0020831868,0.0013946148],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00038758287,0.00027539075,0.015821852,0.007286006,0.0003638203,0.002785084,0.0058635306,0.0064991694,0.0191395,0.44813985,0.13868243,0.35475582],"study_design_scores_gemma":[0.000029336315,0.00005119316,0.0075675044,0.0015849136,0.00011326341,0.0016310859,0.0009451436,0.0081779305,0.003757762,0.06153811,0.91450846,0.000095231764],"about_ca_topic_score_codex":0.013834475,"about_ca_topic_score_gemma":0.01087035,"teacher_disagreement_score":0.013834475,"about_ca_system_score_codex":0.0034637214,"about_ca_system_score_gemma":0.010192632,"threshold_uncertainty_score":0.027507901},"labels":[],"label_agreement":null},{"id":"W4401535430","doi":"10.1093/database/baae073","title":"Functional implications of glycans and their curation: insights from the workshop held at the 16th Annual International Biocuration Conference in Padua, Italy","year":2024,"lang":"en","type":"article","venue":"Database","topic":"Glycosylation and Glycoproteins Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada's Michael Smith Genome Sciences Centre; University of British Columbia","funders":"National Institute of General Medical Sciences; National Heart, Lung, and Blood Institute; National Institutes of Health; National Human Genome Research Institute; Biotechnology and Biological Sciences Research Council; Novo Nordisk Fonden","keywords":"Data curation; Glycan; Function (biology); Glycosylation; Computer science; Data science; World Wide Web; Computational biology; Biology; Biochemistry; Genetics; Glycoprotein","score_opus":0.029365272496693136,"score_gpt":0.29121867497272247,"score_spread":0.26185340247602934,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4401535430","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.21967393,0.038440183,0.09359567,0.55228186,0.025190983,0.0018014256,0.008641513,0.0015618998,0.05881249],"genre_scores_gemma":[0.56356287,0.055242606,0.18791825,0.067472234,0.009358704,0.0032619792,0.030636434,0.0037112746,0.07883564],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.98180145,0.011300674,0.0007667321,0.0015045,0.003060592,0.0015660917],"domain_scores_gemma":[0.95320755,0.02696588,0.0014742453,0.0021840837,0.0116103785,0.0045578424],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.068847105,0.001274625,0.00067265664,0.0034692413,0.0058611445,0.012259548,0.0026737459,0.0027560256,0.004334349],"category_scores_gemma":[0.0543626,0.0005416018,0.0015004219,0.004090559,0.0034025738,0.006716748,0.012774386,0.0056586578,0.0016314773],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000732667,0.00053719024,0.012066827,0.0041541914,0.00014497021,0.0037662205,0.07855641,0.0032546413,0.009645704,0.02551797,0.53966147,0.32196173],"study_design_scores_gemma":[0.00006448653,0.0001905334,0.011929719,0.0032913277,0.00014840285,0.0007098408,0.036628004,0.003398989,0.009544858,0.01658752,0.9172556,0.00025074612],"about_ca_topic_score_codex":0.007956121,"about_ca_topic_score_gemma":0.016135536,"teacher_disagreement_score":0.9311529,"about_ca_system_score_codex":0.006170053,"about_ca_system_score_gemma":0.012318108,"threshold_uncertainty_score":0.36410266},"labels":[],"label_agreement":null},{"id":"W4401978887","doi":"10.1093/database/baae085","title":"Autoinhibited Protein Database: a curated database of autoinhibitory domains and their autoinhibition mechanisms","year":2024,"lang":"en","type":"article","venue":"Database","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"National Research Foundation of Korea","keywords":"Allosteric regulation; Database; Drug discovery; Mechanism (biology); Computational biology; Computer science; Biology; Bioinformatics; Biochemistry; Enzyme","score_opus":0.018850469998935308,"score_gpt":0.2772149938642937,"score_spread":0.2583645238653584,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4401978887","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008008903,0.011092946,0.009647765,0.00029824302,0.000110674424,0.00022961752,0.953113,0.01223649,0.0052622915],"genre_scores_gemma":[0.0065597985,0.003948526,0.011956886,0.00016997509,0.000024292869,0.00021457416,0.9752466,0.00072403386,0.0011553692],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99914074,0.00007632775,0.00022522916,0.00020812263,0.0002608708,0.000088682275],"domain_scores_gemma":[0.9985008,0.000348755,0.00033002804,0.00023342825,0.00029546558,0.00029149075],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011364435,0.0029267683,0.0031812012,0.007539623,0.0009081141,0.0029655981,0.0031908127,0.0023673163,0.01627966],"category_scores_gemma":[0.0039628875,0.0010129749,0.0013471076,0.009105717,0.00048662262,0.00244541,0.0028068346,0.0017340785,0.023829326],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0023225704,0.00036220317,0.009226594,0.031210778,0.0008527897,0.0021019264,0.0005459092,0.004734064,0.044705987,0.010503426,0.77576447,0.11766932],"study_design_scores_gemma":[0.00050674664,0.00017418768,0.016463496,0.0015987561,0.00043840418,0.0018652205,0.00017358075,0.0046887435,0.01143721,0.0058873175,0.9565895,0.00017693006],"about_ca_topic_score_codex":0.004625582,"about_ca_topic_score_gemma":0.0060008103,"teacher_disagreement_score":0.01627966,"about_ca_system_score_codex":0.0011678793,"about_ca_system_score_gemma":0.0043861223,"threshold_uncertainty_score":0.054460943},"labels":[],"label_agreement":null},{"id":"W4405227333","doi":"10.1093/database/baae116","title":"Probe my Pathway (PmP): a portal to explore the chemical coverage of the human Reactome","year":2024,"lang":"en","type":"article","venue":"Database","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada Research Chairs; University of Toronto","funders":"Structural Genomics Consortium; Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Computer science; Information retrieval; World Wide Web; Data science","score_opus":0.019083017541008565,"score_gpt":0.2575803154624792,"score_spread":0.23849729792147067,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405227333","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008319764,0.0026883192,0.016671829,0.0005246634,0.00007732732,0.00017858138,0.9123056,0.046570126,0.012663837],"genre_scores_gemma":[0.030725883,0.003406657,0.035529777,0.00066169444,0.000054825825,0.0006312617,0.9180339,0.006022111,0.004933842],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9996954,0.00004171994,0.00003922913,0.00010364262,0.00008436304,0.000035711375],"domain_scores_gemma":[0.9991273,0.00032135306,0.00013822052,0.00015579355,0.00009617129,0.0001612747],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007375754,0.0015762196,0.0010039136,0.004056884,0.0003608466,0.001581214,0.0013941838,0.0012706727,0.054873824],"category_scores_gemma":[0.0023694625,0.0007189059,0.0007594831,0.004446488,0.0002600989,0.0019209298,0.0027015437,0.000987959,0.029337326],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0028912125,0.00018106648,0.011508393,0.009237408,0.00045307155,0.0016125664,0.00052009546,0.0023917048,0.040424794,0.010824613,0.7895996,0.13035558],"study_design_scores_gemma":[0.0005220517,0.00023554593,0.01653547,0.00069872866,0.0001878423,0.0020745515,0.0002483589,0.005230377,0.024948021,0.015663031,0.9335283,0.00012767497],"about_ca_topic_score_codex":0.0017324126,"about_ca_topic_score_gemma":0.0028600881,"teacher_disagreement_score":0.054873824,"about_ca_system_score_codex":0.0006297796,"about_ca_system_score_gemma":0.0010603393,"threshold_uncertainty_score":0.18357122},"labels":[],"label_agreement":null},{"id":"W4407006479","doi":"10.1093/database/baae097","title":"Helping authors produce FAIR taxonomic data: evaluation of an author-driven phenotype data production prototype","year":2025,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canadian Museum of Nature; Agriculture and Agri-Food Canada; Université de Montréal; Ministry of Natural Resources and Forestry; University of Manitoba; University of Ottawa","funders":"National Science Foundation","keywords":"Computer science; Ontology; Workflow; Character (mathematics); Vocabulary; World Wide Web; Information retrieval; Data science; Artificial intelligence; Database","score_opus":0.15365448967522466,"score_gpt":0.40095387990632747,"score_spread":0.2472993902311028,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4407006479","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6097322,0.00021622158,0.3185634,0.00080378365,0.00027331678,0.00287163,0.0019594603,0.05621293,0.009367017],"genre_scores_gemma":[0.35090792,0.00020327259,0.6250633,0.00041686997,0.00004119101,0.0015337478,0.004665364,0.00505333,0.012115028],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.9950723,0.0021764534,0.00053512363,0.00079073355,0.00125074,0.00017466057],"domain_scores_gemma":[0.9543659,0.032527972,0.0006861633,0.006165725,0.004767669,0.0014866733],"candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.011042111,0.0009924959,0.0008557152,0.0009416524,0.0007178026,0.002930355,0.0044433502,0.0023104313,0.0067629353],"category_scores_gemma":[0.03745289,0.00065096724,0.00081152597,0.00067550415,0.0009467748,0.0037316794,0.002837311,0.001256116,0.0027161958],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0068076174,0.008674382,0.021788608,0.0032509442,0.00029175018,0.0077325907,0.052858334,0.022331633,0.20055524,0.009319038,0.042217202,0.62417275],"study_design_scores_gemma":[0.0035701953,0.013212218,0.021729646,0.00083318754,0.00064173917,0.0062659206,0.015342622,0.43678236,0.23401554,0.010531795,0.25617552,0.00089922966],"about_ca_topic_score_codex":0.0016179911,"about_ca_topic_score_gemma":0.0016568556,"teacher_disagreement_score":0.99555665,"about_ca_system_score_codex":0.0010215426,"about_ca_system_score_gemma":0.0013569716,"threshold_uncertainty_score":0.058396935},"labels":[],"label_agreement":null},{"id":"W4410539343","doi":"10.1093/database/baaf008","title":"An exploratory study combining Virtual Reality and Semantic Web for life science research using Graph2VR","year":2025,"lang":"en","type":"article","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"H2020 Health","keywords":"Computer science; Virtual reality; World Wide Web; Exploratory research; Semantic Web; Information retrieval; Web of science; Social Semantic Web; Data science; Human–computer interaction; MEDLINE; Sociology","score_opus":0.14193607316905618,"score_gpt":0.44667886393519246,"score_spread":0.3047427907661363,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4410539343","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.7209753,0.0015768682,0.2023653,0.0037661847,0.00023251133,0.003313334,0.02463509,0.0123540135,0.030781444],"genre_scores_gemma":[0.56185704,0.0008555563,0.4140719,0.000811745,0.00007263291,0.0021476182,0.013983979,0.002093436,0.004106104],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99526346,0.0033759912,0.0002002752,0.00044920007,0.00052650826,0.00018455759],"domain_scores_gemma":[0.9678908,0.026606888,0.0005083749,0.0029575804,0.0013523594,0.0006840153],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0096187005,0.00069506286,0.00048486388,0.0036248565,0.0010168629,0.0036310875,0.0016515715,0.0011148464,0.0062185926],"category_scores_gemma":[0.022461317,0.00041574694,0.0012175953,0.0036874902,0.0010212983,0.0038539208,0.00362178,0.0011054665,0.0012591154],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0039159455,0.0053337445,0.11262183,0.010624775,0.0009934258,0.0097982,0.094558455,0.034858964,0.043057144,0.060977,0.07321539,0.55004513],"study_design_scores_gemma":[0.0014091077,0.004583269,0.081579834,0.001873649,0.00089583796,0.005945739,0.059940048,0.15793121,0.04314957,0.079428986,0.5626083,0.00065452274],"about_ca_topic_score_codex":0.0045828917,"about_ca_topic_score_gemma":0.009628153,"teacher_disagreement_score":0.0096187005,"about_ca_system_score_codex":0.0010485747,"about_ca_system_score_gemma":0.0011344213,"threshold_uncertainty_score":0.050869167},"labels":[],"label_agreement":null},{"id":"W4414493406","doi":"10.1093/database/baaf053","title":"p53motifDB: integration of genomic information and tumour suppressor p53 binding motifs","year":2025,"lang":"en","type":"article","venue":"Database","topic":"Cancer-related Molecular Pathways","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Institute of General Medical Sciences; University at Albany; National Institutes of Health; Canadian Centre for Applied Research in Cancer Control","keywords":"Suppressor; Chromatin; Gene; Transcription factor; Genome; DNA; DNA binding site; Human genome; Promoter","score_opus":0.011861195864569407,"score_gpt":0.2605337110777503,"score_spread":0.2486725152131809,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4414493406","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03300132,0.004082967,0.038431335,0.00054300623,0.00016253136,0.00034267732,0.84543866,0.061052702,0.016944842],"genre_scores_gemma":[0.03954055,0.0017664565,0.027306264,0.00022475928,0.00005422053,0.00022187906,0.92490685,0.0022019502,0.003777074],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99974054,0.000031882395,0.000041023937,0.000076302,0.00007254029,0.000037661335],"domain_scores_gemma":[0.9995421,0.00009989321,0.00007200941,0.000073109084,0.00007857587,0.00013447306],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005587461,0.0012197858,0.0009578992,0.0038519737,0.0005159793,0.0013422072,0.0017604604,0.0008524132,0.022070903],"category_scores_gemma":[0.0016811899,0.00044160726,0.00049053394,0.0038782288,0.00020827613,0.0012826364,0.0017504325,0.00071163574,0.014245535],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0038765604,0.00042150295,0.017243944,0.0052782977,0.00034112626,0.0016459707,0.00065149245,0.0030120788,0.06373846,0.010437164,0.5903411,0.30301225],"study_design_scores_gemma":[0.0005148783,0.0003573904,0.029814692,0.0005008005,0.00026135423,0.0024814687,0.00046886955,0.011435125,0.041202683,0.010235599,0.902555,0.00017217587],"about_ca_topic_score_codex":0.0036533768,"about_ca_topic_score_gemma":0.0042105108,"teacher_disagreement_score":0.022070903,"about_ca_system_score_codex":0.00045662356,"about_ca_system_score_gemma":0.001063486,"threshold_uncertainty_score":0.07383454},"labels":[],"label_agreement":null},{"id":"W4414533220","doi":"10.1093/database/baaf044","title":"GExplore 1.5: a comprehensive <i>Caenorhabditis elegans</i> database for the analysis of gene function with a new user-friendly web interface","year":2025,"lang":"en","type":"article","venue":"Database","topic":"Genetics, Aging, and Longevity in Model Organisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Emily Carr University of Art and Design; Simon Fraser University","funders":"","keywords":"Interface (matter); Proteome; User interface; Function (biology); Gene; Online database; The Internet; Web application; Graphical user interface","score_opus":0.013398199947390752,"score_gpt":0.2579205502741165,"score_spread":0.24452235032672573,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4414533220","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0066931644,0.0025522816,0.05850715,0.00038399958,0.00021935148,0.00043429525,0.73153406,0.18678775,0.012888007],"genre_scores_gemma":[0.009769592,0.0017991698,0.07660647,0.0005335726,0.00004969496,0.0011540967,0.8808119,0.022790503,0.0064849136],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99932957,0.000059923834,0.000085270774,0.00018274735,0.000251729,0.00009072707],"domain_scores_gemma":[0.999413,0.00014657642,0.00007255665,0.00013909004,0.0001273002,0.00010155492],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001111099,0.0043442524,0.002822762,0.0040708454,0.0009531103,0.0028219782,0.0049166162,0.0019924045,0.07341093],"category_scores_gemma":[0.0017428892,0.0024939398,0.0017559838,0.003609765,0.0004960227,0.0032222883,0.0030617644,0.003073014,0.06571496],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012015653,0.00010226667,0.0016023491,0.004415663,0.00034077675,0.000741719,0.0002229706,0.000763156,0.081223205,0.0027511069,0.85699683,0.04963842],"study_design_scores_gemma":[0.0009856503,0.0001998862,0.014014683,0.0008598832,0.0002912899,0.0021549412,0.00010815381,0.006789219,0.060491107,0.0050225137,0.9086575,0.00042518295],"about_ca_topic_score_codex":0.005120439,"about_ca_topic_score_gemma":0.0069265147,"teacher_disagreement_score":0.07341093,"about_ca_system_score_codex":0.0009325814,"about_ca_system_score_gemma":0.0017929017,"threshold_uncertainty_score":0.24558401},"labels":[],"label_agreement":null},{"id":"W4416865968","doi":"10.1093/database/baag027","title":"Large-scale Manual Curation and Harmonization of Metadata from Metagenomic and Cancer Genomic Repositories: Challenges and Solutions","year":2025,"lang":"en","type":"preprint","venue":"Database","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute of Population and Public Health","funders":"National Cancer Institute; National Institutes of Health","keywords":"Metadata; Data curation; Harmonization; Geospatial metadata; Ontology; Standardization; Metadata repository; Digital curation","score_opus":0.041418253084737164,"score_gpt":0.30703858055245375,"score_spread":0.2656203274677166,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4416865968","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.13028455,0.012114221,0.76099837,0.046131935,0.0013599143,0.0048344047,0.019273799,0.012672507,0.012330377],"genre_scores_gemma":[0.094063975,0.002365181,0.8677826,0.0063005155,0.00035682597,0.0028265025,0.02059427,0.003656266,0.0020538496],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.8538506,0.07167219,0.029604154,0.015413058,0.026903903,0.002556139],"domain_scores_gemma":[0.5340558,0.17378949,0.048403278,0.17675652,0.0625301,0.0044648442],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.1868279,0.0017463745,0.002622961,0.017641174,0.007815648,0.013874441,0.007950565,0.002304157,0.0023452824],"category_scores_gemma":[0.2648692,0.002207272,0.0029731619,0.022422573,0.0050455746,0.012239569,0.021411825,0.004389629,0.0016723315],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00044817844,0.00065602985,0.10446231,0.009345648,0.0020798917,0.0015615881,0.03125447,0.0057486617,0.031675592,0.036162365,0.08165633,0.694949],"study_design_scores_gemma":[0.00029108085,0.0004137144,0.09541553,0.010297843,0.0023025984,0.0032017215,0.027486028,0.022908768,0.065528005,0.15689555,0.61417687,0.001082387],"about_ca_topic_score_codex":0.012063584,"about_ca_topic_score_gemma":0.024506206,"teacher_disagreement_score":0.8131721,"about_ca_system_score_codex":0.0045644194,"about_ca_system_score_gemma":0.0315631,"threshold_uncertainty_score":0.9880522},"labels":[],"label_agreement":null}]}