{"meta":{"query_hash":"2a78974ab50e","filters":{"venue":"NAR Genomics and Bioinformatics"},"cohort_total":54,"direct_labels_cover":0,"predictions_cover":54,"exported":54,"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/2a78974ab50e","api":"https://metacan.xera.ac/api/v1/cohort?venue=NAR+Genomics+and+Bioinformatics"},"results":[{"id":"W2964367788","doi":"10.1093/nargab/lqz003","title":"Open chromatin structure in PolyQ disease-related genes: a potential mechanism for CAG repeat expansion in the normal human population","year":2019,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":15,"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":"Azrieli Foundation; Israel Science Foundation","keywords":"Chromatin; Gene; Biology; Epigenetics; Genetics; Genome; ChIA-PET; Population; Chromatin remodeling","score_opus":0.00761182982789455,"score_gpt":0.2540048161247481,"score_spread":0.24639298629685358,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2964367788","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":[0.9934384,0.0006739132,0.0047216294,0.00011148348,0.000009199379,0.000013386893,0.00010144087,0.000034514833,0.0008960457],"genre_scores_gemma":[0.9990435,0.00010382375,0.0004550344,0.000037415495,0.000008725865,0.000006444894,0.00006595438,0.000004471926,0.0002744579],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99983037,0.000033381824,0.00001065421,0.00006450435,0.000037861068,0.00002315041],"domain_scores_gemma":[0.9997447,0.00007853688,0.00007008341,0.00003818536,0.000022111184,0.000046290956],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00031916922,0.000110957786,0.00017684967,0.00060377934,0.00017549741,0.00026927548,0.00022018439,0.00023230708,0.001677236],"category_scores_gemma":[0.0005658905,0.000089548594,0.00013995734,0.00033973373,0.00044621585,0.00021146314,0.00021208181,0.00016122033,0.00013391058],"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":[0.0005397208,0.00008734634,0.40739664,0.0000937009,0.00013607656,0.0016749679,0.0009998819,0.00042308908,0.56108046,0.005548319,0.00024804045,0.021771906],"study_design_scores_gemma":[0.00004593195,0.0004535937,0.9406605,0.000030822226,0.00013501245,0.007433136,0.00060148776,0.003942043,0.037263777,0.006369584,0.003036344,0.00002784328],"about_ca_topic_score_codex":0.00039209673,"about_ca_topic_score_gemma":0.0006872559,"teacher_disagreement_score":0.001677236,"about_ca_system_score_codex":0.000119989694,"about_ca_system_score_gemma":0.00011766849,"threshold_uncertainty_score":0.005610943},"labels":[],"label_agreement":null},{"id":"W3001111520","doi":"10.1093/nargab/lqaa002","title":"Bayesian correlation is a robust gene similarity measure for single-cell RNA-seq data","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":24,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ottawa Hospital; University of Ottawa","funders":"University of Bern; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Similarity (geometry); Bayesian probability; Correlation; Pearson product-moment correlation coefficient; Mathematics; Artificial intelligence; RNA-Seq; Similarity measure; Pattern recognition (psychology); Computational biology; Computer science; Gene; Statistics; Biology; Gene expression; Transcriptome; Genetics","score_opus":0.059689052934759304,"score_gpt":0.22642125919351536,"score_spread":0.16673220625875607,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3001111520","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":[0.019672398,0.000479341,0.97713655,0.000095569514,0.000039349594,0.000083931955,0.0007210238,0.0010720618,0.00069969415],"genre_scores_gemma":[0.39255756,0.00057060324,0.59840846,0.00039672168,0.00020954985,0.0006117687,0.004725908,0.00086873566,0.0016506432],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9896185,0.0027331184,0.0008831752,0.0027076316,0.0037170583,0.0003405981],"domain_scores_gemma":[0.9780739,0.012089618,0.0040583764,0.0025486124,0.002784282,0.0004452937],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009040083,0.0009569684,0.0020572434,0.0044075954,0.0009870803,0.0019239775,0.0014291867,0.0017631676,0.0016066732],"category_scores_gemma":[0.033849236,0.000627365,0.0018104308,0.0047694854,0.0016505443,0.0021908497,0.0019716704,0.0019734183,0.0012578544],"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":[0.0010781484,0.00034255398,0.06574888,0.0017528923,0.002133114,0.0006590595,0.0007382433,0.33164275,0.13294697,0.06540327,0.012600295,0.38495383],"study_design_scores_gemma":[0.000043187818,0.0003183054,0.033259846,0.0000997228,0.00017639586,0.0008671411,0.00011280704,0.85627294,0.033808593,0.06489047,0.00990236,0.00024829205],"about_ca_topic_score_codex":0.0021494443,"about_ca_topic_score_gemma":0.0025019539,"teacher_disagreement_score":0.009040083,"about_ca_system_score_codex":0.00141268,"about_ca_system_score_gemma":0.0016954102,"threshold_uncertainty_score":0.047809124},"labels":[],"label_agreement":null},{"id":"W3016947700","doi":"10.1093/nargab/lqaa024","title":"A minimum reporting standard for multiple sequence alignments","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":76,"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 Waterloo","funders":"Commonwealth Scientific and Industrial Research Organisation","keywords":"Completeness (order theory); Sequence (biology); Transparency (behavior); nobody; Computer science; Simple (philosophy); Data mining; Mathematics; Biology; Computer security; Genetics","score_opus":0.04466610776664429,"score_gpt":0.2708863985655486,"score_spread":0.2262202907989043,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3016947700","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"reporting","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"reporting","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0077478527,0.0012623585,0.9622335,0.0018184663,0.0006377232,0.001051068,0.0053450065,0.010707751,0.009196264],"genre_scores_gemma":[0.0426852,0.0007970653,0.93853647,0.00054221036,0.00042253896,0.002540367,0.011348321,0.0015867266,0.001541133],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.8351256,0.06949483,0.03781231,0.008328627,0.04636929,0.002869332],"domain_scores_gemma":[0.66672426,0.114209965,0.030830313,0.08505952,0.09933067,0.0038452493],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.09577065,0.0020214836,0.0022918864,0.010060301,0.003654592,0.010490462,0.006659635,0.0050332877,0.0058566537],"category_scores_gemma":[0.33994377,0.0015487204,0.002318906,0.009425648,0.0030033933,0.010428189,0.0068216706,0.005804354,0.0076500103],"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":[0.0010287772,0.00056822144,0.015502137,0.0035050905,0.00032973438,0.0006025633,0.002237392,0.02326021,0.027559653,0.40397662,0.112353645,0.409076],"study_design_scores_gemma":[0.0002953442,0.0011263394,0.006697479,0.003284046,0.00030576522,0.0023094947,0.0013351195,0.114852026,0.07189579,0.31955567,0.47770724,0.00063565397],"about_ca_topic_score_codex":0.0020220503,"about_ca_topic_score_gemma":0.0012731015,"teacher_disagreement_score":0.90422934,"about_ca_system_score_codex":0.0029298717,"about_ca_system_score_gemma":0.009625876,"threshold_uncertainty_score":0.50648963},"labels":[],"label_agreement":null},{"id":"W3032790130","doi":"10.1093/nargab/lqaa035","title":"Where are G-quadruplexes located in the human transcriptome?","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":36,"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 Sherbrooke","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; Université de Sherbrooke","keywords":"Transcriptome; Computational biology; Non-coding RNA; Biology; RNA splicing; Alternative splicing; RNA; Untranslated region; G-quadruplex; Nucleic acid structure; Gene; RNA-Seq; microRNA; Genetics; Gene expression; Messenger RNA; DNA","score_opus":0.017109350249607753,"score_gpt":0.22700013400142757,"score_spread":0.20989078375181983,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3032790130","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":[0.9459961,0.0068082386,0.036825277,0.0010950468,0.00005585098,0.000026894677,0.0047739926,0.00067246764,0.0037461282],"genre_scores_gemma":[0.97947925,0.0021744159,0.014463826,0.00025289116,0.00003194587,0.000025693009,0.0023468444,0.000058726248,0.0011663671],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.9998975,0.000022447357,0.0000053227427,0.000046920948,0.000012765698,0.000015047781],"domain_scores_gemma":[0.99983466,0.000063736035,0.00004456164,0.000017468923,0.000020854035,0.000018703655],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00011550942,0.00018593248,0.00035448567,0.00036125377,0.0002518765,0.0004825626,0.00017013987,0.00032156406,0.0017794365],"category_scores_gemma":[0.00047223282,0.00017569096,0.00035434397,0.00045053413,0.00025053168,0.00052154955,0.00014673025,0.00026572665,0.0008514845],"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":[0.0015855914,0.00010528168,0.12490758,0.0017722656,0.00022147672,0.0022240686,0.00072803587,0.034872368,0.62902015,0.01150957,0.0060468675,0.18700667],"study_design_scores_gemma":[0.00010060749,0.0011857112,0.41476467,0.000689329,0.0007375168,0.0075013083,0.0025775915,0.23361447,0.1912853,0.071366936,0.075954616,0.0002218568],"about_ca_topic_score_codex":0.0009082075,"about_ca_topic_score_gemma":0.0011783005,"teacher_disagreement_score":0.0017794365,"about_ca_system_score_codex":0.00017792708,"about_ca_system_score_gemma":0.00026457137,"threshold_uncertainty_score":0.005952835},"labels":[],"label_agreement":null},{"id":"W3037354611","doi":"10.1093/nargab/lqaa043","title":"Factorial study of the RNA-seq computational workflow identifies biases as technical gene signatures","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","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":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Sherbrooke","funders":"Fonds de recherche du Québec – Nature et technologies; Fonds de Recherche du Québec - Santé; Natural Sciences and Engineering Research Council of Canada","keywords":"Modularity (biology); RNA-Seq; Pipeline (software); Computational biology; Computer science; Workflow; RNA; Modular design; Software; Gene; Biology; Genetics; Transcriptome; Gene expression; Database; Programming language","score_opus":0.019293357317551538,"score_gpt":0.24295756390417217,"score_spread":0.22366420658662062,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3037354611","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":[0.4414167,0.0005571554,0.550408,0.0004519842,0.00014841341,0.000481176,0.0011922412,0.0023214412,0.0030228873],"genre_scores_gemma":[0.7379821,0.0002869794,0.2571118,0.00033413345,0.00003986912,0.0007720628,0.0014499541,0.0009880306,0.0010350997],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99201936,0.0028511127,0.0004623964,0.0023711252,0.0018994771,0.00039656216],"domain_scores_gemma":[0.98172045,0.011577326,0.0013956146,0.0028829046,0.002067507,0.00035625964],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01171695,0.0010361237,0.00075456063,0.0008890578,0.000750761,0.0022204362,0.0007859928,0.0007376763,0.0012585861],"category_scores_gemma":[0.02349586,0.00042974006,0.0012114501,0.0012377629,0.0008773283,0.0012752782,0.0010098842,0.0013656613,0.000709714],"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":[0.0016777673,0.0005172053,0.06742608,0.0009022974,0.00056822796,0.00017000189,0.00059625023,0.0286587,0.77447677,0.011851239,0.0015108209,0.11164466],"study_design_scores_gemma":[0.000072016366,0.00094539096,0.07495191,0.00010458503,0.00039499026,0.0002520207,0.00031908357,0.2796929,0.61566633,0.01789899,0.009509399,0.00019249876],"about_ca_topic_score_codex":0.0010549015,"about_ca_topic_score_gemma":0.0014949406,"teacher_disagreement_score":0.98828304,"about_ca_system_score_codex":0.0013237724,"about_ca_system_score_gemma":0.0018520217,"threshold_uncertainty_score":0.061965883},"labels":[],"label_agreement":null},{"id":"W3038449881","doi":"10.1093/nargab/lqaa046","title":"On the prediction of DNA-binding preferences of C2H2-ZF domains using structural models: application on human CTCF","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":14,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"BC Children's Hospital; University of British Columbia","funders":"European Regional Development Fund; Ministerio de Ciencia e Innovación","keywords":"Zinc finger; CTCF; DNA; Transcription factor; Computational biology; Binding site; Biology; DNA-binding protein; Chromatin; Genetics; Gene","score_opus":0.04841538685684868,"score_gpt":0.24961517097100394,"score_spread":0.20119978411415526,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3038449881","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":[0.8986335,0.00020284666,0.09774811,0.00008593033,0.000007962752,0.00004869624,0.0007424041,0.0013509027,0.0011796231],"genre_scores_gemma":[0.9325437,0.00016143452,0.0653353,0.000016841752,0.000005110499,0.000054633616,0.0014063307,0.00008660388,0.00039009398],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9999175,0.000029638482,0.000003819992,0.000019603463,0.00001779633,0.0000116128185],"domain_scores_gemma":[0.9996964,0.0002051523,0.000017559032,0.000018045923,0.000034891476,0.000028021012],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00031575825,0.0005509676,0.00047377145,0.0005869477,0.0002925076,0.00035099545,0.00038128064,0.000535442,0.0012747687],"category_scores_gemma":[0.0012227951,0.00022537872,0.0003993469,0.00044442381,0.00013113904,0.00028213303,0.00021956615,0.0002865026,0.0002800398],"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":[0.0006383882,0.00032698739,0.009472709,0.00015794173,0.00007514666,0.00020671946,0.000077784025,0.8937038,0.03985631,0.002235704,0.0005871037,0.05266137],"study_design_scores_gemma":[0.00002355746,0.000041666513,0.0007976149,0.000003056907,0.000006595021,0.00002332568,0.000017116135,0.9951461,0.0033049174,0.00045012563,0.00018168267,0.0000042587726],"about_ca_topic_score_codex":0.005874961,"about_ca_topic_score_gemma":0.004464462,"teacher_disagreement_score":0.005874961,"about_ca_system_score_codex":0.00035389274,"about_ca_system_score_gemma":0.000678506,"threshold_uncertainty_score":0.011681557},"labels":[],"label_agreement":null},{"id":"W3089247545","doi":"10.1093/nargab/lqaa066","title":"The expression of long non-coding RNAs is associated with H3Ac and H3K4me2 changes regulated by the HDA6-LDL1/2 histone modification complex in <i>Arabidopsis</i>","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Cancer-related molecular mechanisms research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":31,"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; Agriculture and Agri-Food Canada","funders":"Agriculture and Agri-Food Canada; National Taiwan University; Ministry of Science and Technology of the People's Republic of China","keywords":"Biology; Long non-coding RNA; Arabidopsis; Histone; Histone deacetylase; Regulation of gene expression; Genetics; RNA splicing; Transcription factor; Gene expression; Alternative splicing; Gene; RNA; Cell biology; Messenger RNA; Mutant","score_opus":0.01933077866174409,"score_gpt":0.253167448844162,"score_spread":0.23383667018241788,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3089247545","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":[0.9894907,0.0028295626,0.0031592841,0.00018936236,0.0000908399,0.000015462065,0.002689359,0.00020721897,0.0013283209],"genre_scores_gemma":[0.9895802,0.00057611236,0.002082842,0.0002699,0.000019662635,0.000024852796,0.0036451933,0.000083575644,0.0037176237],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9998567,0.0000057433654,0.000008825296,0.0000625739,0.000037586233,0.000028630417],"domain_scores_gemma":[0.9998989,0.000008191934,0.000032815646,0.000005955822,0.000016120886,0.000038013528],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00011967444,0.00030181676,0.00034968188,0.0003616111,0.000421639,0.00030984802,0.00021120945,0.00024058577,0.0011715987],"category_scores_gemma":[0.00007856395,0.00021274167,0.00055459485,0.00026787329,0.00022260622,0.00019026667,0.0002955984,0.00047300052,0.0004168657],"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":[0.00007627222,0.000006980564,0.0013261699,0.000043742846,0.00001031687,0.000047431764,0.000027886701,0.000028439974,0.99767524,0.000037525802,0.000112673384,0.0006074484],"study_design_scores_gemma":[0.000060815502,0.00014547507,0.34035727,0.000025830102,0.00010485493,0.0005369108,0.00028558358,0.0047741923,0.6454446,0.00021684385,0.007985955,0.00006169556],"about_ca_topic_score_codex":0.005535294,"about_ca_topic_score_gemma":0.0067651295,"teacher_disagreement_score":0.005535294,"about_ca_system_score_codex":0.00055752124,"about_ca_system_score_gemma":0.00022634889,"threshold_uncertainty_score":0.0110061765},"labels":[],"label_agreement":null},{"id":"W3097162993","doi":"10.1093/nargab/lqaa086","title":"aliFreeFoldMulti: alignment-free method to predict secondary structures of multiple RNA homologs","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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 Sherbrooke","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"RNA; Sequence (biology); Multiple sequence alignment; Set (abstract data type); Structural alignment; Algorithm; Computer science; Nucleic acid structure; Protein secondary structure; Sequence alignment; Nucleic acid secondary structure; Computational biology; Mathematics; Biology; Genetics; Peptide sequence","score_opus":0.012136323943378734,"score_gpt":0.22952791334956302,"score_spread":0.21739158940618428,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3097162993","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":[0.015904449,0.00066119555,0.93810046,0.000139317,0.00010071067,0.00015097485,0.001132629,0.041660886,0.0021493621],"genre_scores_gemma":[0.10148037,0.0003200964,0.8848027,0.00024060464,0.000042775806,0.00025296008,0.004801989,0.0036139004,0.004444552],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995766,0.00007797228,0.000030142895,0.00012419089,0.00015114227,0.000040015573],"domain_scores_gemma":[0.99952173,0.00020597996,0.000055445045,0.00007198762,0.000103816106,0.000041129373],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009075155,0.0014908051,0.0009948134,0.0015010021,0.0009472589,0.00072011334,0.0015743535,0.0013040998,0.008307171],"category_scores_gemma":[0.0018168689,0.00057673623,0.0011504112,0.0008783859,0.0003010856,0.001283652,0.0008633754,0.0011763911,0.0035260478],"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":[0.00093901507,0.0003065287,0.0068385857,0.0007168034,0.00039688993,0.0004316972,0.00023657308,0.11049453,0.055351317,0.00916458,0.04877533,0.76634824],"study_design_scores_gemma":[0.000074303156,0.00011778839,0.0012523296,0.00003252138,0.000042085067,0.000305758,0.000037840764,0.95084524,0.026024878,0.0059235576,0.015293762,0.000049861286],"about_ca_topic_score_codex":0.0032747425,"about_ca_topic_score_gemma":0.005765484,"teacher_disagreement_score":0.008307171,"about_ca_system_score_codex":0.00057941594,"about_ca_system_score_gemma":0.0013138452,"threshold_uncertainty_score":0.027790189},"labels":[],"label_agreement":null},{"id":"W3110037558","doi":"10.1093/nargab/lqaa098","title":"TOUCAN: a framework for fungal biosynthetic gene cluster discovery","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Microbial Natural Products and Biosynthesis","field":"Medicine","cited_by":36,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University; Université du Québec à Montréal","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Université du Québec à Montréal","keywords":"Biology; Computational biology; Scope (computer science); Computer science","score_opus":0.020215865460523626,"score_gpt":0.23932528590337646,"score_spread":0.21910942044285284,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3110037558","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":[0.006546342,0.0006519383,0.9758553,0.00029064855,0.000050008777,0.00016159867,0.0028655757,0.0124256285,0.0011530766],"genre_scores_gemma":[0.062960446,0.00044485662,0.9227157,0.00024331796,0.00004729167,0.00038993536,0.011260666,0.0005621911,0.0013756594],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99901223,0.00036509216,0.000058766087,0.00025810517,0.00025251045,0.000053343072],"domain_scores_gemma":[0.99881977,0.0007125761,0.0000698709,0.00011547283,0.00020241196,0.00007979857],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024166545,0.0012823055,0.0010549176,0.0021770888,0.0008001827,0.0013336481,0.0024970316,0.0011373272,0.0034901574],"category_scores_gemma":[0.004075956,0.0006162969,0.0021024353,0.001762988,0.0005600567,0.0011142656,0.0019850887,0.0017346456,0.0016054675],"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":[0.00076605816,0.00038610338,0.0068594497,0.0016941661,0.0007556996,0.00070774584,0.00029561817,0.49387738,0.012015362,0.049382303,0.040525053,0.39273506],"study_design_scores_gemma":[0.000033463915,0.00003980352,0.0003138442,0.000027162569,0.0000237899,0.000057952395,0.000023113176,0.97076553,0.0012428659,0.020161731,0.0072977333,0.000012955281],"about_ca_topic_score_codex":0.006455848,"about_ca_topic_score_gemma":0.014371173,"teacher_disagreement_score":0.006455848,"about_ca_system_score_codex":0.001143328,"about_ca_system_score_gemma":0.0022123577,"threshold_uncertainty_score":0.012836516},"labels":[],"label_agreement":null},{"id":"W3129861733","doi":"10.1093/nargab/lqab011","title":"Single-cell mapper (scMappR): using scRNA-seq to infer the cell-type specificities of differentially expressed genes","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","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":"Princess Margaret Cancer Centre; University of Toronto; SickKids Foundation; University Health Network; Canadian Institute for Advanced Research; Vector Institute","funders":"Canadian Institutes of Health Research; Hospital for Sick Children; Natural Sciences and Engineering Research Council of Canada; Ontario Ministry of Research, Innovation and Science","keywords":"RNA-Seq; RNA; Biology; Cell type; Computational biology; Gene expression; Gene; Population; Cell; Genetics; Transcriptome","score_opus":0.025780475905938335,"score_gpt":0.21534810033163332,"score_spread":0.18956762442569497,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3129861733","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":[0.025346356,0.0009202833,0.87547153,0.00029044691,0.00023744866,0.00023387128,0.013163939,0.08225913,0.0020770563],"genre_scores_gemma":[0.10399152,0.0008977132,0.8505668,0.00086009613,0.00010524954,0.001212886,0.023405401,0.016036447,0.0029237994],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983006,0.00027319347,0.00007871063,0.0007865889,0.00046159487,0.00009929282],"domain_scores_gemma":[0.997244,0.0015085163,0.0003518268,0.0004888397,0.00028729346,0.00011959281],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003997192,0.0020550545,0.001973313,0.0019449634,0.0011124384,0.002189891,0.0019782032,0.0015563004,0.006596443],"category_scores_gemma":[0.0074931025,0.0011300456,0.0021525957,0.0014760027,0.0010167384,0.0012475662,0.0021321652,0.0028271945,0.0061649354],"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":[0.001493981,0.00024989794,0.026008222,0.005122888,0.0022366382,0.0012697874,0.001476722,0.059113514,0.45310727,0.014388461,0.11519,0.32034254],"study_design_scores_gemma":[0.0002934826,0.0003447559,0.014037761,0.0002208166,0.0005047353,0.0013510889,0.00021159473,0.48961237,0.36257464,0.027799964,0.10261267,0.000436126],"about_ca_topic_score_codex":0.002099591,"about_ca_topic_score_gemma":0.0039325,"teacher_disagreement_score":0.006596443,"about_ca_system_score_codex":0.0006941632,"about_ca_system_score_gemma":0.0017221065,"threshold_uncertainty_score":0.022067308},"labels":[],"label_agreement":null},{"id":"W3158542203","doi":"10.1093/nargab/lqab035","title":"Wavelet Screening identifies regions highly enriched for differentially methylated loci for orofacial clefts","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Cleft Lip and Palate Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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":"Norges Forskningsråd; Réseau de cancérologie Rossy","keywords":"DNA methylation; Computational biology; Epigenetics; Differentially methylated regions; Biology; Methylation; Genetics; Bioinformatics; Computer science; Gene; Gene expression","score_opus":0.026889180161575744,"score_gpt":0.28434592654648333,"score_spread":0.25745674638490756,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3158542203","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":[0.9445405,0.00016887992,0.052791506,0.00025486978,0.000038286456,0.000030314355,0.00056609325,0.00051838765,0.0010911328],"genre_scores_gemma":[0.9732562,0.00007953163,0.025394628,0.00007384946,0.00001119533,0.000028400396,0.00071905204,0.000054989487,0.0003821547],"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99985814,0.00004932541,0.000007092093,0.000031777043,0.00002394593,0.00002958476],"domain_scores_gemma":[0.9993567,0.00042669472,0.00005517376,0.00006194774,0.00005451698,0.000044947676],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009176614,0.00035236267,0.0005032799,0.0004247611,0.00030625885,0.00041218568,0.0006485534,0.0004939295,0.0016743332],"category_scores_gemma":[0.0030971651,0.00021845858,0.0007161102,0.00039215785,0.000313438,0.00031646618,0.00053603825,0.0003775346,0.00019778042],"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":[0.00065978925,0.00017219692,0.089524604,0.00016709979,0.0003540878,0.00061828643,0.00012809188,0.835922,0.023359062,0.0038606753,0.0029781207,0.042255905],"study_design_scores_gemma":[0.00004771912,0.000043283737,0.0067371516,0.0000051848006,0.0000310519,0.000038467257,0.000030550636,0.9893217,0.0021592958,0.0012469349,0.0003303272,0.000008344579],"about_ca_topic_score_codex":0.007030213,"about_ca_topic_score_gemma":0.006621133,"teacher_disagreement_score":0.007030213,"about_ca_system_score_codex":0.000251348,"about_ca_system_score_gemma":0.00071537663,"threshold_uncertainty_score":0.0139786005},"labels":[],"label_agreement":null},{"id":"W3158876629","doi":"10.1093/nargab/lqab027","title":"On the use of direct-coupling analysis with a reduced alphabet of amino acids combined with super-secondary structure motifs for protein fold prediction","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Protein Structure and Dynamics","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":"BC Children's Hospital; University of British Columbia","funders":"Instituto de Salud Carlos III; Ministerio de Economía y Competitividad; Generalitat de Catalunya","keywords":"Protein secondary structure; Protein folding; Sequence (biology); Protein structure prediction; Protein structure; Structural motif; Alphabet; Computational biology; Protein sequencing; Amino acid; Chemistry; Biological system; Peptide sequence; Crystallography; Biology; Biochemistry","score_opus":0.008394857317764562,"score_gpt":0.19393503414601293,"score_spread":0.18554017682824836,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3158876629","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":[0.22383346,0.00024934794,0.77047604,0.00004860021,0.000015682752,0.00010053587,0.00028041922,0.0027785385,0.0022173617],"genre_scores_gemma":[0.42975673,0.00018151822,0.56726134,0.00005930871,0.000011940167,0.00009875574,0.000915082,0.00033657663,0.001378746],"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9997944,0.000052465188,0.000013422719,0.00006836848,0.000054460816,0.000016765252],"domain_scores_gemma":[0.9996687,0.00014458473,0.000048280683,0.000063929925,0.000039581697,0.000034972927],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00045174645,0.00059668615,0.0006047912,0.00095492,0.00030837982,0.0005057045,0.0006247906,0.00043521178,0.0017277787],"category_scores_gemma":[0.0010544333,0.00021395487,0.000674445,0.00080369785,0.0003257543,0.00044493418,0.00057342777,0.0005476935,0.0006574578],"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":[0.00064526976,0.00040214433,0.0063220616,0.0003242025,0.00018999576,0.00048010857,0.00017919025,0.46033654,0.20888337,0.009814251,0.001009111,0.31141376],"study_design_scores_gemma":[0.000012744507,0.00009883912,0.0007026791,0.000006241721,0.0000148074505,0.00012557414,0.000011696757,0.9850359,0.01118869,0.0021213184,0.0006705833,0.000011040522],"about_ca_topic_score_codex":0.0017069937,"about_ca_topic_score_gemma":0.0020432023,"teacher_disagreement_score":0.0017277787,"about_ca_system_score_codex":0.0002899327,"about_ca_system_score_gemma":0.00056123344,"threshold_uncertainty_score":0.005780041},"labels":[],"label_agreement":null},{"id":"W3164889407","doi":"10.1093/nargab/lqab037","title":"Detection of homozygous and hemizygous complete or partial exon deletions by whole-exome sequencing","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomic variations and chromosomal abnormalities","field":"Biochemistry, Genetics and Molecular Biology","cited_by":13,"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 Research Resources; National Institutes of Health; Fonds de Recherche en Santé Respiratoire; Hospital for Sick Children; Institut National de la Santé et de la Recherche Médicale; SCOR Corporate Foundation for Science; Institut des maladies génétiques Imagine; St. Giles Foundation; Université de Paris; Rockefeller University; Yale University; National Center for Advancing Translational Sciences; National Human Genome Research Institute; Agence Nationale de la Recherche; National Institute of Allergy and Infectious Diseases; Howard Hughes Medical Institute","keywords":"Exome sequencing; Exon; Genetics; Biology; Computational biology; Exome; Copy-number variation; Gene; Mutation; Genome","score_opus":0.012977036308436845,"score_gpt":0.20678007204484916,"score_spread":0.1938030357364123,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3164889407","genre_codex":"empirical","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":[0.630362,0.0014841461,0.3542119,0.00032591264,0.00009000492,0.00021802669,0.0059186863,0.00520983,0.002179476],"genre_scores_gemma":[0.69973004,0.0004814703,0.2841868,0.00035528504,0.0000373306,0.00019603105,0.012687224,0.0005082811,0.0018174043],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99829453,0.00040156132,0.00017526977,0.00062992057,0.00039619,0.0001025035],"domain_scores_gemma":[0.9977379,0.0013623426,0.00028694767,0.00031899117,0.00020922905,0.00008459812],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025804536,0.00079208985,0.00078455004,0.0016438392,0.00044068394,0.0008617415,0.0007935132,0.0007844316,0.0016963984],"category_scores_gemma":[0.00709076,0.00036574077,0.000803956,0.0009995169,0.0003731896,0.0004891838,0.0015938925,0.0005951075,0.00045253604],"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":[0.0024059638,0.00029986826,0.17634352,0.00094511197,0.0016245726,0.003086355,0.0006046372,0.060241893,0.35877636,0.003734919,0.011222357,0.3807146],"study_design_scores_gemma":[0.0003329696,0.00054534484,0.17133419,0.00015832859,0.00043294823,0.005390919,0.00036688795,0.4801867,0.30801874,0.010059632,0.02298473,0.00018872683],"about_ca_topic_score_codex":0.0015791068,"about_ca_topic_score_gemma":0.004369201,"teacher_disagreement_score":0.0025804536,"about_ca_system_score_codex":0.00027611104,"about_ca_system_score_gemma":0.00052885554,"threshold_uncertainty_score":0.013646901},"labels":[],"label_agreement":null},{"id":"W3172664492","doi":"10.1093/nargab/lqab042","title":"Retracted and Replaced: Known sequence features can explain half of all human gene ends","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":true,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Occupational Cancer Research Centre; University of Toronto","funders":"Canadian Institutes of Health Research","keywords":"Polyadenylation; Gene; Coding region; Genetics; Biology; Sequence (biology); Computational biology; Primary transcript; Context (archaeology); Regulatory sequence; Messenger RNA; Gene expression; Alternative splicing","score_opus":0.018737581049586446,"score_gpt":0.2740747268087642,"score_spread":0.25533714575917776,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3172664492","genre_codex":"empirical","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":[0.82124484,0.003912374,0.15350537,0.0013425617,0.0004832135,0.00008953122,0.005243917,0.0026133223,0.011564817],"genre_scores_gemma":[0.97494304,0.0006251943,0.013609475,0.000316024,0.00010023706,0.000035941983,0.0049886806,0.00035639046,0.0050251395],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.999448,0.00015851301,0.000031595722,0.00021184537,0.00009450812,0.000055562592],"domain_scores_gemma":[0.9980388,0.0010537093,0.00016748403,0.0004521159,0.00020402612,0.00008381277],"candidate_categories":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.00093664904,0.0007824397,0.000642267,0.0006436819,0.0004672474,0.0008781717,0.00071131205,0.0010478497,0.006050928],"category_scores_gemma":[0.004699748,0.00025415083,0.00093742396,0.0008346739,0.0004962417,0.0012977349,0.00061187276,0.0010889891,0.003968939],"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":[0.003625298,0.0003116336,0.21711412,0.00083814474,0.0005319662,0.003811791,0.00089247705,0.15917972,0.049460888,0.008695281,0.016021797,0.53951687],"study_design_scores_gemma":[0.00007863447,0.00056244305,0.08863071,0.00019672485,0.00032941962,0.004250268,0.00035767743,0.8258367,0.020690016,0.031183856,0.027776506,0.00010696806],"about_ca_topic_score_codex":0.0024849442,"about_ca_topic_score_gemma":0.003142513,"teacher_disagreement_score":0.99895215,"about_ca_system_score_codex":0.00037727374,"about_ca_system_score_gemma":0.00045176773,"threshold_uncertainty_score":0.020242333},"labels":[],"label_agreement":null},{"id":"W3217246491","doi":"10.1093/nargab/lqab105","title":"RNA-Scoop: interactive visualization of transcripts in single-cell transcriptomes","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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; University of British Columbia","funders":"National Human Genome Research Institute; National Institutes of Health; Genome British Columbia; Canada's Michael Smith Genome Sciences Centre; Genome Canada","keywords":"Transcriptome; RNA; Computational biology; Visualization; Biology; Cell; Single-cell analysis; Computer science; Genetics; Gene expression; Gene; Data mining","score_opus":0.013069874968820455,"score_gpt":0.228204744840605,"score_spread":0.21513486987178454,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3217246491","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":[0.027942473,0.00069086015,0.53637874,0.00094052183,0.00044563744,0.00020358675,0.040012497,0.38701865,0.006367055],"genre_scores_gemma":[0.2622591,0.001444174,0.631652,0.0011154226,0.00026675174,0.0017401914,0.04744452,0.047040723,0.007037144],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99959093,0.00006379012,0.000033020835,0.000092211405,0.00015807236,0.00006189848],"domain_scores_gemma":[0.99882394,0.0006699109,0.00006866314,0.00011770749,0.00016831883,0.00015136182],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011126879,0.0012353525,0.0007973045,0.0018018228,0.0009192143,0.0015645727,0.001568837,0.0009279342,0.019422898],"category_scores_gemma":[0.0021314158,0.00053477037,0.0010265387,0.0013614732,0.00049820385,0.0014988933,0.0022104597,0.002062299,0.0026068613],"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":[0.0023278927,0.00026056496,0.007996872,0.0033254158,0.0005516602,0.002033794,0.0029861117,0.02205985,0.30375785,0.015859777,0.3843224,0.2545178],"study_design_scores_gemma":[0.00067686324,0.00021708231,0.015897121,0.0006213018,0.00017062825,0.0020134633,0.0011447101,0.40102658,0.22471161,0.06676086,0.28618664,0.0005732107],"about_ca_topic_score_codex":0.0030456681,"about_ca_topic_score_gemma":0.004678515,"teacher_disagreement_score":0.019422898,"about_ca_system_score_codex":0.0005335886,"about_ca_system_score_gemma":0.000994684,"threshold_uncertainty_score":0.06497604},"labels":[],"label_agreement":null},{"id":"W4206004774","doi":"10.1093/nargab/lqab122","title":"A comparison on predicting functional impact of genomic variants","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Rare Diseases","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":"University of Calgary","funders":"National Key Research and Development Program of China","keywords":"Single-nucleotide polymorphism; Missense mutation; Genomics; Human genome; Functional genomics; SNP","score_opus":0.014672094106870068,"score_gpt":0.2587007337804651,"score_spread":0.24402863967359503,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4206004774","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":[0.8602415,0.038072634,0.07182051,0.0026749456,0.0007237916,0.00028447356,0.010571033,0.005167252,0.010443793],"genre_scores_gemma":[0.9193673,0.0039491216,0.05324776,0.00062377605,0.00019688219,0.0001412391,0.020381385,0.0003441144,0.0017485576],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9935051,0.002705821,0.0007204809,0.0012823389,0.0015013363,0.00028495968],"domain_scores_gemma":[0.9711603,0.0237345,0.000738184,0.0016326406,0.0021721278,0.0005622724],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01557363,0.0017114418,0.0015693068,0.0053082383,0.00058685796,0.0018833373,0.0015345212,0.0015294155,0.0014314185],"category_scores_gemma":[0.023373498,0.00029659606,0.001952807,0.0023869276,0.00050919777,0.0015203955,0.0013562862,0.0013194613,0.0006018497],"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":[0.0041042683,0.00079131185,0.33251947,0.0020195625,0.005334178,0.0006775485,0.00025962072,0.2315205,0.005893688,0.003329105,0.019243253,0.3943075],"study_design_scores_gemma":[0.00034141634,0.0010274604,0.08681041,0.00031723615,0.0009494914,0.0010405573,0.00023171473,0.88959086,0.006709996,0.004034313,0.008788117,0.00015844987],"about_ca_topic_score_codex":0.008055055,"about_ca_topic_score_gemma":0.008830499,"teacher_disagreement_score":0.01557363,"about_ca_system_score_codex":0.0011166629,"about_ca_system_score_gemma":0.0014656397,"threshold_uncertainty_score":0.082362235},"labels":[],"label_agreement":null},{"id":"W4206771039","doi":"10.1093/nargab/lqab123","title":"FILER: a framework for harmonizing and querying large-scale functional genomics knowledge","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":20,"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 on Aging; Weston Brain Institute","keywords":"Scale (ratio); Genomics; Computer science; Functional genomics; Data science; Computational biology; Biology; Genome; Geography; Genetics; Cartography; Gene","score_opus":0.02949434952622089,"score_gpt":0.25344907683266604,"score_spread":0.22395472730644517,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4206771039","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":[0.0018138466,0.0009748029,0.5910475,0.0008975827,0.00019201452,0.00065359904,0.03889904,0.36214602,0.0033755596],"genre_scores_gemma":[0.036874417,0.0017447073,0.6829357,0.0015772575,0.0001888734,0.002619819,0.2218936,0.048291095,0.0038745322],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9903515,0.0018719608,0.0015725594,0.0023828917,0.0030845234,0.0007365821],"domain_scores_gemma":[0.9865991,0.005749542,0.00082410546,0.004342023,0.0014003442,0.0010849726],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.020140205,0.0048873876,0.004479834,0.009011571,0.0022202071,0.010077059,0.017049951,0.0048307693,0.020011805],"category_scores_gemma":[0.030280296,0.0038758602,0.008381544,0.0080145905,0.0036719071,0.014109958,0.018366568,0.0057823816,0.016664438],"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":[0.0031374071,0.0006025703,0.0051406487,0.006565049,0.0013965378,0.0028315259,0.0026103754,0.07110447,0.020629622,0.11030425,0.52409124,0.25158632],"study_design_scores_gemma":[0.0015135541,0.00039503607,0.0034614815,0.0016244795,0.00040421245,0.0017587331,0.0010576049,0.29714546,0.022439016,0.23387155,0.4352413,0.0010874475],"about_ca_topic_score_codex":0.01677063,"about_ca_topic_score_gemma":0.013802885,"teacher_disagreement_score":0.020140205,"about_ca_system_score_codex":0.0035168899,"about_ca_system_score_gemma":0.004984847,"threshold_uncertainty_score":0.106512845},"labels":[],"label_agreement":null},{"id":"W4211146659","doi":"10.1093/nargab/lqab125","title":"An improved molecular inversion probe based targeted sequencing approach for low variant allele frequency","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Cancer Genomics and Diagnostics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":14,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Toronto General Hospital; University of Toronto; Princess Margaret Cancer Centre; University Health Network","funders":"Rising Tide Foundation; Applebaum Foundation; Princess Margaret Cancer Foundation","keywords":"Allele frequency; Allele; Genetics; Inversion (geology); Computational biology; Biology; Paleontology; Gene","score_opus":0.007255798986068817,"score_gpt":0.20408456254668625,"score_spread":0.19682876356061743,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4211146659","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":[0.20455754,0.0013479142,0.78372145,0.00061075576,0.00016866403,0.00024346453,0.0027931123,0.0044356864,0.0021215195],"genre_scores_gemma":[0.45437393,0.00058839505,0.53699625,0.00086031057,0.000103858656,0.00030214444,0.0032809894,0.00061702845,0.0028771744],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99862325,0.00030694788,0.00006940857,0.00049904664,0.00037626893,0.0001250123],"domain_scores_gemma":[0.99890125,0.0005037397,0.00014854583,0.00014774787,0.0002208121,0.00007795215],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015800302,0.00095450255,0.00089864933,0.0010914047,0.00030500704,0.0008017107,0.0009970336,0.0010772578,0.002009379],"category_scores_gemma":[0.0030225024,0.00054892665,0.0008586466,0.0006125365,0.00041025542,0.00059998315,0.001008137,0.0015062288,0.0015467951],"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":[0.00033149018,0.000108022905,0.0069966414,0.00029174727,0.00016490245,0.0003812417,0.0001345434,0.0076921596,0.90724444,0.0015223004,0.0016190577,0.07351345],"study_design_scores_gemma":[0.00007635337,0.00059803383,0.014694888,0.00004236133,0.00026573704,0.0020313875,0.000058956986,0.25885466,0.7067597,0.0038965328,0.0125583485,0.00016310168],"about_ca_topic_score_codex":0.0013280545,"about_ca_topic_score_gemma":0.0025125113,"teacher_disagreement_score":0.002009379,"about_ca_system_score_codex":0.00034579638,"about_ca_system_score_gemma":0.00068717333,"threshold_uncertainty_score":0.008356094},"labels":[],"label_agreement":null},{"id":"W4214852607","doi":"10.1093/nargab/lqac010","title":"G-quadruplex occurrence and conservation: more than just a question of guanine–cytosine content","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"DNA and Nucleic Acid Chemistry","field":"Biochemistry, Genetics and Molecular Biology","cited_by":18,"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 Sherbrooke","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Guanine; Cytosine; G-quadruplex; Content (measure theory); Chemistry; DNA; Mathematics; Biochemistry; Gene; Mathematical analysis","score_opus":0.020899956561881126,"score_gpt":0.2418839903462985,"score_spread":0.2209840337844174,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4214852607","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":[0.9911582,0.0012283901,0.0057935547,0.00009524803,0.000008074476,0.0000055759306,0.00045128187,0.000101090736,0.0011586149],"genre_scores_gemma":[0.99789596,0.00011209034,0.0014514446,0.000023867138,0.000007771757,0.0000027125097,0.00034473315,0.000016271413,0.00014521486],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.9996474,0.00007981535,0.000022040404,0.00014598783,0.00006846189,0.000036203164],"domain_scores_gemma":[0.9981943,0.0009007055,0.00043297056,0.00013010514,0.00019846197,0.0001434105],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005286013,0.00017788909,0.00028736898,0.0010587601,0.00019296205,0.00079320284,0.00020042065,0.0003298667,0.0014441643],"category_scores_gemma":[0.0019946452,0.00010970048,0.000173108,0.0006767861,0.00044317474,0.0007058091,0.00027525445,0.000308903,0.00050366914],"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":[0.00078843883,0.00006990385,0.45237437,0.0005837358,0.0003518547,0.00024843507,0.0006275222,0.0049553136,0.49292642,0.0017060636,0.00060157105,0.044766456],"study_design_scores_gemma":[0.000019100491,0.0005248535,0.8508533,0.00013422973,0.00027124822,0.0012252349,0.001001149,0.034599293,0.099640995,0.0067880587,0.004867412,0.00007523597],"about_ca_topic_score_codex":0.0005298841,"about_ca_topic_score_gemma":0.0006583281,"teacher_disagreement_score":0.0014441643,"about_ca_system_score_codex":0.00015981209,"about_ca_system_score_gemma":0.00013205849,"threshold_uncertainty_score":0.004831195},"labels":[],"label_agreement":null},{"id":"W4220964301","doi":"10.1093/nargab/lqac018","title":"Data-driven identification of inherent features of eukaryotic stress-responsive genes","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","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 Toronto","funders":"National Cancer Institute; National Institutes of Health; Boehringer Ingelheim Fonds","keywords":"Gene; Computational biology; Biology; Saccharomyces cerevisiae; Yeast; Gene expression; Genetics","score_opus":0.020233198738345302,"score_gpt":0.2752859193507044,"score_spread":0.2550527206123591,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4220964301","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":[0.90229714,0.0006914685,0.08235062,0.0002837716,0.000031733165,0.00008953896,0.012500647,0.0010696974,0.0006853681],"genre_scores_gemma":[0.9473526,0.0001695419,0.0333717,0.00011062497,0.000020576497,0.000119635566,0.018539064,0.00009473849,0.00022154312],"study_design_codex":"observational","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9993247,0.00017536081,0.00005258713,0.00023049206,0.00016545704,0.000051427276],"domain_scores_gemma":[0.997285,0.0017428239,0.00027891312,0.00019570948,0.00039630322,0.00010134208],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013606539,0.00044612482,0.00066397287,0.0011935459,0.0003604634,0.00079099013,0.0004559118,0.00039544,0.0004813867],"category_scores_gemma":[0.003446684,0.00018723904,0.00091362983,0.0012235297,0.00038883876,0.00042393646,0.00041303626,0.000607644,0.00022582075],"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":[0.001982074,0.00058446795,0.30271775,0.0013089344,0.00090884184,0.0007393935,0.00023413425,0.30118376,0.29120326,0.004932446,0.0040076966,0.09019719],"study_design_scores_gemma":[0.00003603898,0.0002569325,0.062344626,0.000033600412,0.00012627975,0.00021344417,0.00009918013,0.89691746,0.0315113,0.005984891,0.0024277004,0.00004863673],"about_ca_topic_score_codex":0.0017521966,"about_ca_topic_score_gemma":0.0033199578,"teacher_disagreement_score":0.0017521966,"about_ca_system_score_codex":0.0005869799,"about_ca_system_score_gemma":0.0008449563,"threshold_uncertainty_score":0.00719589},"labels":[],"label_agreement":null},{"id":"W4221050649","doi":"10.1093/nargab/lqac026","title":"Construction of a chromosome-level Japanese stickleback species genome using ultra-dense linkage analysis with single-cell sperm sequencing","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"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":"Institute of Genetics; Japan Society for the Promotion of Science; Research Organization of Information and Systems","keywords":"Genome; Biology; Genetics; Linkage (software); Contig; Genetic linkage; Genotyping; Genomics; Chromosome; Sperm; Computational biology; Stickleback; Genotype; Gene; Fish <Actinopterygii>","score_opus":0.025442406299582006,"score_gpt":0.20445496936941632,"score_spread":0.17901256306983432,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4221050649","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":[0.60446435,0.0013428318,0.384261,0.00018216738,0.0000910925,0.00032025718,0.005253435,0.001737094,0.0023477757],"genre_scores_gemma":[0.502461,0.00093102065,0.47918114,0.000149726,0.000026166008,0.0002497701,0.012337475,0.00038071338,0.004283091],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99980754,0.000023791436,0.000015250992,0.00007890172,0.00005658331,0.000017974895],"domain_scores_gemma":[0.99974006,0.00006103646,0.000053622054,0.000047678615,0.000067594,0.000030061206],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004153215,0.00031976466,0.00047167475,0.00088765105,0.00038975282,0.00045844333,0.00040854578,0.00043252343,0.0014173621],"category_scores_gemma":[0.00036362925,0.00030940052,0.0006585308,0.0006612842,0.0001853139,0.00029575403,0.000578589,0.0006155038,0.00069551903],"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":[0.000061734536,0.000021999931,0.001762615,0.00016120603,0.00004318796,0.00011389673,0.00012186726,0.0008941716,0.98076206,0.00042415143,0.00016917787,0.015464036],"study_design_scores_gemma":[0.00007438927,0.0005283017,0.095954046,0.00009527153,0.0004312831,0.0010946841,0.00041793575,0.034128282,0.82563883,0.0015611512,0.03996852,0.00010723799],"about_ca_topic_score_codex":0.0027215134,"about_ca_topic_score_gemma":0.005453659,"teacher_disagreement_score":0.0027215134,"about_ca_system_score_codex":0.000294521,"about_ca_system_score_gemma":0.00038734576,"threshold_uncertainty_score":0.0054112673},"labels":[],"label_agreement":null},{"id":"W4289712796","doi":"10.1093/nargab/lqac057","title":"Mining bacterial NGS data vastly expands the complete genomes of temperate phages","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Bacteriophages and microbial interactions","field":"Environmental Science","cited_by":19,"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 Guelph; Simon Fraser University; University of British Columbia","funders":"National Institutes of Health; National Key Research and Development Program of China; Hebei Provincial Key Research Projects; U.S. National Library of Medicine; National Natural Science Foundation of China","keywords":"Bacterial genome size; Genome; Temperate climate; Biology; Computational biology; Evolutionary biology; Data science; Computer science; Genetics; Ecology; Gene","score_opus":0.03181369653655028,"score_gpt":0.23701824864955845,"score_spread":0.20520455211300817,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4289712796","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":[0.6041881,0.0141786765,0.10841118,0.0014493784,0.00031970564,0.0002650296,0.25465167,0.0050087655,0.0115276715],"genre_scores_gemma":[0.36702207,0.00550806,0.15432206,0.00085693086,0.00010575903,0.0005464797,0.4681684,0.00069580605,0.0027744353],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99840814,0.0002702296,0.0001291755,0.00058039156,0.00048060704,0.0001315031],"domain_scores_gemma":[0.997982,0.0009169186,0.0002211355,0.00035226485,0.00038928812,0.0001383254],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014433161,0.0012753104,0.0012353096,0.0034486777,0.001100235,0.0013766565,0.0008331025,0.0009592427,0.001883478],"category_scores_gemma":[0.0038801208,0.00046732224,0.001303469,0.0040014256,0.00041427684,0.0013589517,0.0017509235,0.0011574642,0.001675735],"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":[0.0019795345,0.00082910765,0.2040492,0.010296935,0.0013821282,0.0027586299,0.0016520831,0.03949452,0.302993,0.006267819,0.052324053,0.37597296],"study_design_scores_gemma":[0.00021837876,0.0005646966,0.26283848,0.0018206573,0.0010508194,0.0028921673,0.0032809186,0.14403293,0.10444684,0.031482548,0.44706345,0.0003082017],"about_ca_topic_score_codex":0.003383288,"about_ca_topic_score_gemma":0.0065394132,"teacher_disagreement_score":0.0034486777,"about_ca_system_score_codex":0.0005071544,"about_ca_system_score_gemma":0.0016031051,"threshold_uncertainty_score":0.0076330304},"labels":[],"label_agreement":null},{"id":"W4292615690","doi":"10.1093/nargab/lqac058","title":"A computational approach to rapidly design peptides that detect SARS-CoV-2 surface protein S","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"SARS-CoV-2 and COVID-19 Research","field":"Medicine","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Agriculture and Agri-Food Canada; Ottawa Hospital; University of Regina; Carleton University; University of Ottawa","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Coronavirus disease 2019 (COVID-19); 2019-20 coronavirus outbreak; Computational biology; Pandemic; Peptide; Coronavirus; Surface plasmon resonance; Biology; Virology; Computer science; Medicine; Biochemistry; Disease; Nanotechnology; Infectious disease (medical specialty); Pathology; Materials science","score_opus":0.06575502002795491,"score_gpt":0.30127007006995987,"score_spread":0.23551505004200496,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4292615690","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":[0.37888917,0.00039474515,0.6098813,0.00058234046,0.00011401885,0.00027024254,0.00050429243,0.0023122828,0.007051594],"genre_scores_gemma":[0.5835348,0.0002568231,0.41246045,0.00037292176,0.000027980039,0.0003518391,0.0009582068,0.00015787398,0.0018792007],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99979955,0.000057091504,0.000012670889,0.000053982163,0.000051497856,0.000025222176],"domain_scores_gemma":[0.9995926,0.0002527029,0.000040968647,0.000028133773,0.00005682338,0.00002880155],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006159349,0.00091474655,0.0005832815,0.00039765015,0.00041770464,0.0007308157,0.0007367883,0.0007075894,0.0016100383],"category_scores_gemma":[0.0014894082,0.00033833148,0.00057125936,0.00030299797,0.00037859223,0.00053640263,0.0005392632,0.00075517205,0.00027991168],"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":[0.0004478715,0.0005431324,0.007182652,0.0003140365,0.00019758115,0.00019006577,0.000097332966,0.88127375,0.031725768,0.00604818,0.0016039927,0.0703756],"study_design_scores_gemma":[0.000033998538,0.00014124847,0.00018199245,0.000003902629,0.000020821913,0.000018798191,0.000016013753,0.9938793,0.0037305926,0.0012669609,0.0007020761,0.0000042763068],"about_ca_topic_score_codex":0.0016794965,"about_ca_topic_score_gemma":0.0028898325,"teacher_disagreement_score":0.0016794965,"about_ca_system_score_codex":0.00042548552,"about_ca_system_score_gemma":0.0014058775,"threshold_uncertainty_score":0.005386114},"labels":[],"label_agreement":null},{"id":"W4310786235","doi":"10.1093/nargab/lqac092","title":"PBSIM3: a simulator for all types of PacBio and ONT long reads","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":106,"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":"Institute of Genetics; Japan Society for the Promotion of Science","keywords":"Computer science; Nanopore sequencing; High fidelity; Fidelity; Throughput; DNA sequencing; Operating system; Engineering; Wireless; Telecommunications","score_opus":0.010942076505552106,"score_gpt":0.23069063892170819,"score_spread":0.2197485624161561,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4310786235","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":[0.09118356,0.0009122098,0.8099864,0.00082391984,0.0006811034,0.00054096116,0.011405939,0.056025214,0.028440678],"genre_scores_gemma":[0.37459585,0.0012429535,0.5618132,0.001058316,0.000120607685,0.002731434,0.023004455,0.014495127,0.020938052],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9996321,0.000070780625,0.000027869864,0.000061026953,0.00015188573,0.000056176286],"domain_scores_gemma":[0.99899215,0.00047560802,0.00006448113,0.000113205366,0.00024113654,0.00011348684],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009497161,0.0010736719,0.0007606391,0.00038317524,0.00071497896,0.0008389238,0.0030004294,0.0015594193,0.010453923],"category_scores_gemma":[0.003072535,0.0007802433,0.0012911507,0.00065486605,0.00046165116,0.0011793773,0.0012132562,0.0020672115,0.002228819],"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":[0.00060811837,0.0003129188,0.007851405,0.00082189986,0.0002997571,0.00039972935,0.0005186273,0.84199405,0.033484466,0.028560326,0.03364049,0.05150836],"study_design_scores_gemma":[0.00007584555,0.000050367435,0.00030208932,0.00001892381,0.000025278365,0.000040992938,0.000025497042,0.9626763,0.0104834195,0.0028158065,0.023453636,0.0000318228],"about_ca_topic_score_codex":0.008628036,"about_ca_topic_score_gemma":0.0071923677,"teacher_disagreement_score":0.010453923,"about_ca_system_score_codex":0.0011740699,"about_ca_system_score_gemma":0.0027254452,"threshold_uncertainty_score":0.034971833},"labels":[],"label_agreement":null},{"id":"W4317624614","doi":"10.1093/nargab/lqad006","title":"Model-driven experimental design workflow expands understanding of regulatory role of Nac in <i>Escherichia coli</i>","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","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":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Queen's University","funders":"National Institutes of Health; Ministry of Science and ICT, South Korea; Novo Nordisk Fonden; National Institute of General Medical Sciences; Danmarks Tekniske Universitet; Novo Nordisk; National Research Foundation of Korea; Ulsan National Institute of Science and Technology; National Research Foundation","keywords":"Escherichia coli; Workflow; Computer science; Computational biology; Business; Chemistry; Biochemical engineering; Process management; Biology; Engineering; Biochemistry; Gene; Database","score_opus":0.0216666012447547,"score_gpt":0.22283707848553883,"score_spread":0.20117047724078413,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4317624614","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":[0.18414593,0.00027480457,0.79174,0.0003453519,0.00012738035,0.0005308233,0.004511142,0.014577018,0.00374756],"genre_scores_gemma":[0.49523866,0.0004238756,0.4927424,0.00019259356,0.000025560372,0.0017916949,0.0066075935,0.0014656546,0.001512034],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99924374,0.00015966923,0.00005105423,0.0002794961,0.00020936965,0.000056551617],"domain_scores_gemma":[0.999084,0.00034182955,0.000092183094,0.00023385788,0.00017972004,0.00006835913],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016076036,0.0010050659,0.00080806087,0.00036832478,0.0004929044,0.0012055156,0.0010294674,0.0006080742,0.0027012315],"category_scores_gemma":[0.0016985735,0.00062154565,0.0011495814,0.0003266207,0.0004155518,0.00060028,0.00078528724,0.0010021806,0.00091025286],"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":[0.0005984387,0.00023596344,0.004320972,0.0005813622,0.00012007434,0.00017442662,0.00008458596,0.11981784,0.8443503,0.004713759,0.0018239022,0.023178436],"study_design_scores_gemma":[0.00010810672,0.00032657423,0.0032144664,0.00003142145,0.0000793278,0.00007850004,0.00005175662,0.55784184,0.4216676,0.0045703757,0.011936846,0.00009313056],"about_ca_topic_score_codex":0.0023073163,"about_ca_topic_score_gemma":0.003060078,"teacher_disagreement_score":0.0027012315,"about_ca_system_score_codex":0.0011566525,"about_ca_system_score_gemma":0.0018557365,"threshold_uncertainty_score":0.009036481},"labels":[],"label_agreement":null},{"id":"W4317743512","doi":"10.1093/nargab/lqad003","title":"Differential Expression Enrichment Tool (DEET): an interactive atlas of human differential gene expression","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","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":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canadian Institute for Advanced Research; Vector Institute; SickKids Foundation; University of Toronto","funders":"National Institutes of Health; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; Centre for Applied Genomics; Genome Canada","keywords":"Computational biology; Expression (computer science); Pipeline (software); Gene expression; Gene; Computer science; Differential (mechanical device); Biology; Gene expression profiling; DEET; Information retrieval; Data mining; Genetics","score_opus":0.009721680046971246,"score_gpt":0.24657784619213813,"score_spread":0.23685616614516689,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4317743512","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":[0.031711653,0.007058828,0.5385583,0.0006443396,0.0006298068,0.0006337366,0.3127338,0.09017926,0.017850282],"genre_scores_gemma":[0.08794935,0.003408163,0.69378763,0.00067589385,0.0001689903,0.004612388,0.19077723,0.011421182,0.0071990467],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99816614,0.00047548584,0.00016249523,0.00048390822,0.00056154904,0.0001504846],"domain_scores_gemma":[0.9977285,0.0013120195,0.00026136133,0.00031611364,0.00027618502,0.000105886626],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024457253,0.0017185458,0.001726275,0.0066323075,0.00092447933,0.0014262729,0.0019916582,0.0006563969,0.024505408],"category_scores_gemma":[0.005026546,0.0008597596,0.002052912,0.0061183316,0.00045754397,0.0009176845,0.0025303904,0.001749454,0.00616386],"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":[0.0023763927,0.00027546665,0.03512955,0.011128345,0.0027130796,0.0013854485,0.0014053817,0.017626343,0.19275674,0.024311503,0.38092047,0.32997116],"study_design_scores_gemma":[0.00067882484,0.00045009612,0.065540396,0.0007946498,0.0014980857,0.002454117,0.00036674165,0.06834607,0.10769669,0.036925077,0.7148952,0.0003540599],"about_ca_topic_score_codex":0.0018754898,"about_ca_topic_score_gemma":0.005017678,"teacher_disagreement_score":0.024505408,"about_ca_system_score_codex":0.0008527592,"about_ca_system_score_gemma":0.0020135303,"threshold_uncertainty_score":0.08197874},"labels":[],"label_agreement":null},{"id":"W4323076270","doi":"10.1093/nargab/lqad017","title":"BaM-seq and TBaM-seq, highly multiplexed and targeted RNA-seq protocols for rapid, low-cost library generation from bacterial samples","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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; Natural Sciences and Engineering Research Council of Canada; Howard Hughes Medical Institute; National Institutes of Health; National Science Foundation","keywords":"RNA-Seq; Computational biology; DNA sequencing; Transcriptome; RNA; Genomic library; Deep sequencing; Biology; Computer science; Gene; Gene expression; Genetics; Genome; Base sequence","score_opus":0.027829999735565897,"score_gpt":0.2409403088289274,"score_spread":0.21311030909336148,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4323076270","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":[0.030955033,0.006768635,0.93435144,0.00052636466,0.00074322923,0.0012463578,0.010237673,0.009131054,0.006040239],"genre_scores_gemma":[0.04635984,0.004844915,0.91910625,0.00081658823,0.00019240401,0.0033300358,0.016889567,0.0012552135,0.0072051776],"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","domain_scores_codex":[0.9969868,0.0007174875,0.00026320608,0.0006557885,0.0011582836,0.00021834437],"domain_scores_gemma":[0.9990746,0.0002725147,0.00018320583,0.00018946502,0.00019518627,0.000085068794],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025360882,0.0016649115,0.0014059137,0.0017896343,0.0011415662,0.0017298588,0.0018867303,0.0012408072,0.005066273],"category_scores_gemma":[0.002272081,0.0012738258,0.0012409525,0.0017048726,0.00079711876,0.0013363843,0.0016373339,0.002826435,0.005618275],"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":[0.00015822225,0.000073115574,0.00041138474,0.0008721653,0.00006517475,0.000096836346,0.00008908111,0.0007487473,0.95813763,0.0023551383,0.0039966297,0.032995783],"study_design_scores_gemma":[0.000037526614,0.00020050023,0.0017808292,0.000102399834,0.00006665095,0.0005156941,0.00005756492,0.0078093978,0.900325,0.0023674748,0.0866412,0.00009566886],"about_ca_topic_score_codex":0.0006317373,"about_ca_topic_score_gemma":0.0021615936,"teacher_disagreement_score":0.005066273,"about_ca_system_score_codex":0.0007779725,"about_ca_system_score_gemma":0.0015118164,"threshold_uncertainty_score":0.016948402},"labels":[],"label_agreement":null},{"id":"W4362562485","doi":"10.1093/nargab/lqad031","title":"Known sequence features explain half of all human gene ends","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","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":"Occupational Cancer Research Centre; University of Toronto","funders":"Canadian Institutes of Health Research; University of Toronto","keywords":"Polyadenylation; Gene; Biology; Genetics; Coding region; Computational biology; Sequence (biology); Primary transcript; Context (archaeology); RNA; Messenger RNA; Alternative splicing","score_opus":0.020650263060473267,"score_gpt":0.2755685505550601,"score_spread":0.2549182874945868,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4362562485","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":[0.9087055,0.0028192224,0.07639153,0.0003151555,0.000050310504,0.000046268717,0.0055164937,0.0009914769,0.0051641795],"genre_scores_gemma":[0.9899412,0.00036504617,0.0042361584,0.00007107676,0.0000205086,0.00001942801,0.004053901,0.00008543221,0.0012072338],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99963427,0.000091052265,0.000017268794,0.00015485927,0.000054174667,0.000048267375],"domain_scores_gemma":[0.99923646,0.0005084538,0.000057568,0.00010484177,0.00005625042,0.000036543694],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004982805,0.00069087785,0.00051795086,0.0008042292,0.00033917403,0.00054189074,0.0002859298,0.0006894121,0.0035907454],"category_scores_gemma":[0.0014423941,0.00022592802,0.0009075967,0.00061962515,0.0002784384,0.0004355085,0.00035840127,0.0004353682,0.0018211701],"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":[0.0023929079,0.00027403928,0.37306845,0.00082402525,0.0008913514,0.0023158689,0.0004182681,0.18994455,0.0981767,0.0056418204,0.008859016,0.31719306],"study_design_scores_gemma":[0.00006844312,0.0003380558,0.20550717,0.00013708879,0.00048561752,0.0038849905,0.00023669093,0.7234491,0.032215975,0.020267485,0.013343404,0.000065979795],"about_ca_topic_score_codex":0.0017216252,"about_ca_topic_score_gemma":0.0028122747,"teacher_disagreement_score":0.0035907454,"about_ca_system_score_codex":0.00023699312,"about_ca_system_score_gemma":0.00029169413,"threshold_uncertainty_score":0.012012184},"labels":[],"label_agreement":null},{"id":"W4366818125","doi":"10.1093/nargab/lqad038","title":"SUP: a probabilistic framework to propagate genome sequence uncertainty, with applications","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","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":"Public Health Agency of Canada; Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Western University","keywords":"Resampling; Computer science; Propagation of uncertainty; Probabilistic logic; Sequence (biology); Representation (politics); Variance (accounting); Algorithm; Data mining; Artificial intelligence; Biology","score_opus":0.015766099384214038,"score_gpt":0.2448559491589844,"score_spread":0.22908984977477037,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4366818125","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":[0.00034036938,0.00009164295,0.9985133,0.00012422218,0.00003059055,0.00001428918,0.000089675304,0.00044966055,0.00034632706],"genre_scores_gemma":[0.0571858,0.0007241594,0.9354717,0.00034833938,0.00048299055,0.00033453095,0.00080936623,0.0008584799,0.0037846263],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99623567,0.0018459989,0.00018848952,0.0006315481,0.00094785483,0.00015047599],"domain_scores_gemma":[0.99141425,0.005719369,0.00070530886,0.0009208221,0.00095915364,0.00028105054],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009087081,0.0020780968,0.0013346882,0.0029979998,0.001541628,0.0034087591,0.0033652978,0.002293016,0.0067789364],"category_scores_gemma":[0.018417839,0.0016080047,0.0034119908,0.0020712751,0.0026888035,0.0035421131,0.0042756964,0.0041915923,0.0021229985],"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":[0.0001479498,0.00006847778,0.001825824,0.0002524196,0.00022176285,0.00032856295,0.00027378977,0.4899486,0.0029235105,0.38715684,0.007855558,0.10899664],"study_design_scores_gemma":[0.000016661283,0.000035870806,0.000175932,0.000028414815,0.000025902993,0.000088455046,0.000018715356,0.82884526,0.00059961056,0.16071406,0.009414793,0.000036267204],"about_ca_topic_score_codex":0.008254163,"about_ca_topic_score_gemma":0.006766155,"teacher_disagreement_score":0.009087081,"about_ca_system_score_codex":0.0017397362,"about_ca_system_score_gemma":0.00302451,"threshold_uncertainty_score":0.048057675},"labels":[],"label_agreement":null},{"id":"W4378713713","doi":"10.1093/nargab/lqad052","title":"Refining the genomic determinants underlying escape from X-chromosome inactivation","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genetic and Clinical Aspects of Sex Determination and Chromosomal Abnormalities","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":"BC Children's Hospital; University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"X-inactivation; Biology; Genetics; Gene; Transgene; X chromosome; Transcription (linguistics); Transcription factor; Phenotype; Chromosome; Context (archaeology)","score_opus":0.022748884197983422,"score_gpt":0.25523385560712253,"score_spread":0.2324849714091391,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4378713713","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":[0.99345857,0.00046616772,0.005393956,0.000029357081,0.0000031559425,0.000012636093,0.00018031117,0.000042617256,0.0004132212],"genre_scores_gemma":[0.9954875,0.00025484985,0.0031325952,0.000016106465,0.0000021812414,0.000011381812,0.00041841893,0.000023094519,0.00065382384],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9998423,0.000030459223,0.000011786643,0.00003414077,0.000053414893,0.00002788184],"domain_scores_gemma":[0.9998487,0.000051315565,0.000052759817,0.000011566893,0.000015207931,0.000020321564],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00023185016,0.00024502186,0.00020473849,0.00032222323,0.00008629187,0.00026190525,0.00015904303,0.00018293732,0.0012696947],"category_scores_gemma":[0.00023019945,0.00013803686,0.00019722957,0.00014975441,0.00024020113,0.0001303352,0.00031435073,0.00034888054,0.00028061197],"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":[0.000047321737,0.000008543373,0.0011293375,0.000016759319,0.0000035796895,0.000038733717,0.000007825094,0.00010070466,0.9976826,0.00006680356,0.0000051593956,0.00089263555],"study_design_scores_gemma":[0.000014139559,0.0003229494,0.027039595,0.000011577654,0.000031561496,0.0007691821,0.000060437476,0.0017498508,0.9682673,0.00017898512,0.0015487983,0.0000057210605],"about_ca_topic_score_codex":0.00024929052,"about_ca_topic_score_gemma":0.0003581383,"teacher_disagreement_score":0.0012696947,"about_ca_system_score_codex":0.00013130221,"about_ca_system_score_gemma":0.00017858009,"threshold_uncertainty_score":0.004247546},"labels":[],"label_agreement":null},{"id":"W4383216982","doi":"10.1093/nargab/lqad065","title":"GCPBayes pipeline: a tool for exploring pleiotropy at the gene level","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genetic Mapping and Diversity in Plants and Animals","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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 Cancer Institute; Ligue Contre le Cancer; Ovarian Cancer Research Fund; National Institutes of Health; Cancer Research UK; Government of Canada; Institut National de la Santé et de la Recherche Médicale; Fondation du cancer du sein du Québec; Canadian Institutes of Health Research; Gray Foundation; Genome Canada; European Commission; Breast Cancer Research Foundation","keywords":"Pleiotropy; Pipeline (software); Context (archaeology); Computer science; Set (abstract data type); Genome-wide association study; Computational biology; Gene; Genome; Data mining; Biology; Phenotype; Genetics; Single-nucleotide polymorphism","score_opus":0.061373062267999635,"score_gpt":0.24300302344086586,"score_spread":0.18162996117286623,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4383216982","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":[0.0048178826,0.00063325797,0.3908231,0.00039851322,0.00039704534,0.00031815696,0.08597822,0.5136858,0.0029479351],"genre_scores_gemma":[0.042466115,0.0010259547,0.68775016,0.0012919422,0.00018385233,0.0031827558,0.12846239,0.12842728,0.0072095487],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990521,0.00019418151,0.00008252297,0.00035696055,0.00021763668,0.00009667715],"domain_scores_gemma":[0.99747807,0.0016475779,0.00015755046,0.00035825212,0.00023578359,0.00012279254],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0033803664,0.0034776027,0.0021038044,0.0037677789,0.0013517508,0.0020719178,0.0028271824,0.0013762396,0.07607163],"category_scores_gemma":[0.007893314,0.0018599604,0.003929852,0.002139788,0.0006681141,0.0022801557,0.0033038494,0.0032672675,0.024365881],"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":[0.002098874,0.00032036804,0.01322924,0.005959467,0.002445476,0.0012900445,0.0016026003,0.009106901,0.050368246,0.012569876,0.69091034,0.21009849],"study_design_scores_gemma":[0.0021550886,0.00048495227,0.029252999,0.0009791473,0.0012098725,0.0021730282,0.0004633952,0.11982757,0.06345431,0.07552043,0.70350415,0.0009750359],"about_ca_topic_score_codex":0.003968278,"about_ca_topic_score_gemma":0.0054629156,"teacher_disagreement_score":0.07607163,"about_ca_system_score_codex":0.0006891398,"about_ca_system_score_gemma":0.0023208417,"threshold_uncertainty_score":0.2544849},"labels":[],"label_agreement":null},{"id":"W4391994538","doi":"10.1093/nargab/lqae018","title":"CDBProm: the Comprehensive Directory of Bacterial Promoters","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Izaak Walton Killam Health Centre; Dalhousie University","funders":"Dirección General de Asuntos del Personal Académico, Universidad Nacional Autónoma de México; Research Nova Scotia; Universidad Nacional Autónoma de México; Dalhousie Medical Research Foundation; Dalhousie University; Canadian Institutes of Health Research; Li Ka Shing Foundation","keywords":"Promoter; Genome; Bacterial genome size; Biology; Computational biology; Genetics; Gene; Bacterial artificial chromosome; DNA sequencing; In silico; Directory; Computer science; Gene expression","score_opus":0.013056275883971543,"score_gpt":0.23510266022786458,"score_spread":0.22204638434389304,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4391994538","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":[0.047625963,0.0040871208,0.07473957,0.0005922453,0.00025905008,0.00036933107,0.6672465,0.19753957,0.007540611],"genre_scores_gemma":[0.06198782,0.0015588583,0.087143086,0.00021737683,0.00008074037,0.00068005355,0.83755916,0.008012318,0.0027606098],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.999012,0.00013549539,0.00008553643,0.00027770078,0.00034571893,0.0001435329],"domain_scores_gemma":[0.9982078,0.00040334061,0.000305601,0.00037353387,0.0003566587,0.00035311855],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010346506,0.0016406289,0.0013033802,0.003652806,0.001037596,0.0015750359,0.0014539687,0.0010476101,0.0071026906],"category_scores_gemma":[0.0032681606,0.001058017,0.0009026152,0.0042024734,0.0003903398,0.0016982016,0.0022890107,0.0015939813,0.015029049],"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":[0.0046234624,0.0004972205,0.044239316,0.005797208,0.0002258406,0.0012488861,0.00076941936,0.014386958,0.09313759,0.010724452,0.581399,0.24295066],"study_design_scores_gemma":[0.0004924415,0.00062210206,0.033649135,0.00075018086,0.00017554141,0.0014507527,0.00037959672,0.047851596,0.0678697,0.01517339,0.8312212,0.0003643086],"about_ca_topic_score_codex":0.002467646,"about_ca_topic_score_gemma":0.0038030162,"teacher_disagreement_score":0.0071026906,"about_ca_system_score_codex":0.0007684511,"about_ca_system_score_gemma":0.001976519,"threshold_uncertainty_score":0.023760915},"labels":[],"label_agreement":null},{"id":"W4392844094","doi":"10.1093/nargab/lqae005","title":"Functional domain annotation by structural similarity","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":19,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Canadian Institutes of Health Research","keywords":"Annotation; Computational biology; Structural similarity; In silico; Protein domain; Domain (mathematical analysis); Similarity (geometry); Structural alignment; Sequence alignment; Proteome; UniProt; Biology; Sequence (biology); Protein sequencing; Benchmark (surveying); Computer science; Bioinformatics; Genetics; Peptide sequence; Artificial intelligence; Geography","score_opus":0.007423367596102924,"score_gpt":0.21458169780913128,"score_spread":0.20715833021302835,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4392844094","genre_codex":"empirical","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":[0.5840777,0.002904861,0.36405775,0.00036721476,0.00016973208,0.0005996197,0.0186364,0.01594919,0.013237568],"genre_scores_gemma":[0.74054986,0.000587378,0.22913536,0.00009899146,0.000046644593,0.00024287966,0.02606397,0.00063762226,0.002637304],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99887985,0.00029171517,0.00015998018,0.00026723096,0.00029218433,0.0001090061],"domain_scores_gemma":[0.99796593,0.0007853809,0.000286793,0.00030878247,0.0005451075,0.00010798818],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012466611,0.00088965177,0.00084721495,0.006222958,0.0005914202,0.0011054034,0.00087181455,0.0006089984,0.008372777],"category_scores_gemma":[0.003549899,0.0002593179,0.0009425549,0.0029960219,0.00030671555,0.001242818,0.00088581245,0.00055966986,0.003463812],"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":[0.002649112,0.0005647654,0.048616573,0.002986327,0.0003916032,0.0008653697,0.00049529306,0.01102564,0.57114357,0.0059189135,0.010268138,0.34507465],"study_design_scores_gemma":[0.00025896615,0.0012061009,0.10049057,0.00047389528,0.0004678122,0.0043435395,0.000868884,0.35720366,0.42955184,0.01804806,0.08683831,0.00024837858],"about_ca_topic_score_codex":0.00086469715,"about_ca_topic_score_gemma":0.000907174,"teacher_disagreement_score":0.008372777,"about_ca_system_score_codex":0.00043922913,"about_ca_system_score_gemma":0.00049609906,"threshold_uncertainty_score":0.028009772},"labels":[],"label_agreement":null},{"id":"W4399154467","doi":"10.1093/nargab/lqae060","title":"G-quadruplex propensity in <i>H. neanderthalensis</i>, <i>H. sapiens</i> and Denisovans mitochondrial genomes","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","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":"Vermont Agency of Natural Resources; Institut National Du Cancer; Grantová Agentura České Republiky; Agence Nationale de la Recherche; CNIB","keywords":"Homo sapiens; Mitochondrial DNA; Genome; Genetics; Biology; DNA; Evolutionary biology; Gene; Geography","score_opus":0.011667398392851707,"score_gpt":0.21640499014123454,"score_spread":0.20473759174838282,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399154467","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":[0.99532557,0.00055147236,0.00079962914,0.000027097904,0.0000068963077,0.000010563215,0.0019592252,0.00003195097,0.0012874657],"genre_scores_gemma":[0.993321,0.0001545432,0.0011561975,0.00002414543,0.000004737074,0.000010854849,0.0045485524,0.000012889186,0.000766909],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.99989164,0.000010665995,0.00000946619,0.000052620973,0.000021519332,0.00001403157],"domain_scores_gemma":[0.99974126,0.00005834453,0.000093265415,0.00001546861,0.00005435306,0.000037306825],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000121810604,0.00019143733,0.00024043233,0.0008763253,0.00030149665,0.00040215364,0.00015541865,0.00026421592,0.001802909],"category_scores_gemma":[0.00038506254,0.000096862175,0.00027083707,0.00059053843,0.0001768807,0.0001322832,0.00024981634,0.00019959663,0.00040011815],"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":[0.00061032997,0.00003819271,0.065958664,0.00044054535,0.00018208858,0.00030492677,0.0009540259,0.0009162374,0.9142074,0.00074442104,0.00033761357,0.01530563],"study_design_scores_gemma":[0.000029096183,0.00034676152,0.88419944,0.00007702391,0.00023668357,0.0017161758,0.0009289564,0.0036468005,0.09158349,0.0010698623,0.016117562,0.000048176345],"about_ca_topic_score_codex":0.0025064442,"about_ca_topic_score_gemma":0.004041844,"teacher_disagreement_score":0.0025064442,"about_ca_system_score_codex":0.00028746296,"about_ca_system_score_gemma":0.00015956587,"threshold_uncertainty_score":0.006031394},"labels":[],"label_agreement":null},{"id":"W4399656493","doi":"10.1093/nargab/lqae068","title":"Structure-based learning to predict and model protein–DNA interactions and transcription-factor co-operativity in <i>cis</i>-regulatory elements","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","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":"BC Children's Hospital; University of British Columbia","funders":"Agencia Estatal de Investigación; Ministerio de Ciencia e Innovación; Generalitat de Catalunya; Human Frontier Science Program","keywords":"Transcription factor; DNA; Computational biology; Transcription (linguistics); Genetics; Biology; Gene","score_opus":0.009518222244531825,"score_gpt":0.2376803225427711,"score_spread":0.22816210029823927,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399656493","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":[0.42988962,0.0005310968,0.5648219,0.00034087844,0.000024133606,0.00007435227,0.00055368774,0.0019865884,0.0017777594],"genre_scores_gemma":[0.8967908,0.00021679795,0.099583566,0.00010101348,0.000026120953,0.00014770639,0.0010537406,0.00010186576,0.0019785063],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9998424,0.000039061815,0.000008029128,0.00006198536,0.00002879875,0.000019735508],"domain_scores_gemma":[0.9995468,0.0002988247,0.00005716178,0.000031900418,0.000040184175,0.000025096877],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004070596,0.0006381907,0.0007041519,0.00054936163,0.0003274597,0.0005404956,0.0008469913,0.0011108133,0.0011411227],"category_scores_gemma":[0.0010421461,0.00054518593,0.0006962002,0.00041576457,0.0005428073,0.000601313,0.00038323773,0.0008643693,0.00043487086],"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":[0.00003573526,0.00007702697,0.0011811258,0.00002045807,0.000026148784,0.000026440295,0.000013356231,0.98373574,0.002869956,0.0008414817,0.00021639283,0.010956079],"study_design_scores_gemma":[0.0000018716202,0.0000059934523,0.00008336507,4.836805e-7,0.0000016600275,0.0000030986344,0.0000012261913,0.9987804,0.00037387988,0.0007014558,0.00004553766,9.503117e-7],"about_ca_topic_score_codex":0.0069339676,"about_ca_topic_score_gemma":0.007966621,"teacher_disagreement_score":0.0069339676,"about_ca_system_score_codex":0.00082798523,"about_ca_system_score_gemma":0.0007441726,"threshold_uncertainty_score":0.01378721},"labels":[],"label_agreement":null},{"id":"W4400562029","doi":"10.1093/nargab/lqae079","title":"Optimizing hybrid ensemble feature selection strategies for transcriptomic biomarker discovery in complex diseases","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","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":"Université Laval","funders":"Natural Sciences and Engineering Research Council of Canada; Agence Nationale de la Recherche","keywords":"Feature selection; Robustness (evolution); Computer science; Biomarker discovery; Feature (linguistics); Data mining; Machine learning; Artificial intelligence; Computational biology; Biology; Gene","score_opus":0.01872548788231583,"score_gpt":0.2641647738759884,"score_spread":0.2454392859936726,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4400562029","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":[0.11746831,0.0005664103,0.8804254,0.00022913281,0.000031733263,0.000068624955,0.00013290487,0.000616456,0.0004610336],"genre_scores_gemma":[0.8121673,0.00021125564,0.1858978,0.00017230083,0.000050344413,0.00016645083,0.0005851173,0.00006226318,0.0006871865],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990005,0.000538011,0.000048446072,0.00016825982,0.00014737048,0.00009750937],"domain_scores_gemma":[0.998007,0.0013360723,0.00009683416,0.00018498606,0.00030585175,0.000069277216],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0040621143,0.001025594,0.0011912589,0.00082723016,0.00035221825,0.0006517311,0.0009240853,0.0006406787,0.00062743766],"category_scores_gemma":[0.0053928094,0.00026799666,0.0010623751,0.0006812396,0.0003271324,0.00095063256,0.00068067305,0.0007893366,0.00022834726],"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":[0.0003373849,0.00018928853,0.0080948565,0.00005578319,0.00043420304,0.00011987588,0.0000706888,0.8202017,0.009328529,0.0014296467,0.0010682495,0.15866981],"study_design_scores_gemma":[0.000008172438,0.00006194203,0.0007720752,0.0000034369455,0.00002348793,0.000017073271,0.00001051622,0.9961123,0.0014613608,0.001363194,0.00016024904,0.00000627933],"about_ca_topic_score_codex":0.0035266527,"about_ca_topic_score_gemma":0.005161383,"teacher_disagreement_score":0.0040621143,"about_ca_system_score_codex":0.00049441494,"about_ca_system_score_gemma":0.0007354257,"threshold_uncertainty_score":0.021482766},"labels":[],"label_agreement":null},{"id":"W4400738267","doi":"10.1093/nargab/lqae084","title":"<b> s </b>CIRCLE—An interactive visual exploration tool for single cell RNA-Seq data","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","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":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Bayerisches Staatsministerium für Wissenschaft und Kunst","keywords":"Metadata; Visualization; Computer science; Data visualization; Focus (optics); RNA-Seq; Software; Interactive visualization; Dimensionality reduction; Data mining; Gene; Artificial intelligence; World Wide Web; Transcriptome; Gene expression; Biology; Programming language","score_opus":0.03484221898688085,"score_gpt":0.2773704812826971,"score_spread":0.24252826229581625,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4400738267","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":[0.008054894,0.00053559954,0.27602094,0.001148193,0.00064396154,0.0003396296,0.063610666,0.63144064,0.018205473],"genre_scores_gemma":[0.06870565,0.001051179,0.6460989,0.0029471223,0.00037889119,0.002173455,0.09857206,0.1455009,0.034571752],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99939966,0.00010418208,0.000054554795,0.0001163466,0.00023447027,0.000090781214],"domain_scores_gemma":[0.99791414,0.0011997818,0.00013275474,0.00020158954,0.00031444582,0.00023735614],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019370025,0.0017660604,0.00091125915,0.0033784504,0.0009349982,0.0024772035,0.0020926371,0.0013300843,0.1390212],"category_scores_gemma":[0.004079141,0.00089272275,0.0014482795,0.0015683329,0.00065839273,0.002555955,0.0030417135,0.0020274268,0.040059276],"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":[0.0005601547,0.00011412979,0.002061608,0.00147259,0.00013628254,0.00075353973,0.0008348135,0.0015631958,0.03785883,0.0048389332,0.80881244,0.1409935],"study_design_scores_gemma":[0.00045662222,0.00019014925,0.0073492266,0.00061131193,0.00008407331,0.0012930161,0.00034979638,0.04921483,0.07626272,0.01670535,0.84703964,0.00044328076],"about_ca_topic_score_codex":0.0035161986,"about_ca_topic_score_gemma":0.0057241237,"teacher_disagreement_score":0.1390212,"about_ca_system_score_codex":0.00058358966,"about_ca_system_score_gemma":0.0010754563,"threshold_uncertainty_score":0.46507227},"labels":[],"label_agreement":null},{"id":"W4401385491","doi":"10.1093/nargab/lqae098","title":"Evaluating cell type deconvolution in FFPE breast tissue: application to benign breast disease","year":2024,"lang":"en","type":"review","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"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 Cancer Institute; National Institutes of Health; Mayo Clinic","keywords":"Deconvolution; Transcriptome; Computational biology; Cell type; RNA; Computer science; Cell; Pattern recognition (psychology); Artificial intelligence; Biology; Gene expression; Gene; Algorithm; Genetics","score_opus":0.030019536453562142,"score_gpt":0.3274310347500039,"score_spread":0.29741149829644176,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4401385491","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00086904963,0.99118257,0.005620904,0.00038130322,0.00020413018,0.000015765409,0.00007411997,0.000044074735,0.0016081214],"genre_scores_gemma":[0.008536041,0.9831655,0.005721737,0.0005177065,0.00020600315,0.000035372723,0.00020343528,0.000030982224,0.0015831555],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9996537,0.00008001904,0.00003115456,0.00008581197,0.00012650972,0.000022794879],"domain_scores_gemma":[0.99926466,0.00040187142,0.000065040374,0.000025530097,0.00020948827,0.000033386557],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016339871,0.0008999299,0.0011570383,0.0019943025,0.00015811416,0.0009906522,0.00087456347,0.0011021722,0.0014954894],"category_scores_gemma":[0.0019968124,0.00032034406,0.0007375171,0.0012679428,0.00045985705,0.0007606488,0.0006164907,0.0012324713,0.0012015406],"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":[0.000078194105,0.000027430815,0.00047729115,0.007385355,0.00017428945,0.00013967525,0.000041806816,0.0013693471,0.0046161376,0.0028033971,0.008149654,0.97473735],"study_design_scores_gemma":[0.000056061243,0.00040994183,0.005368592,0.005309742,0.000704196,0.004268414,0.00012472307,0.0044972347,0.021606045,0.008839966,0.9486786,0.00013654848],"about_ca_topic_score_codex":0.001228769,"about_ca_topic_score_gemma":0.0019551783,"teacher_disagreement_score":0.0019943025,"about_ca_system_score_codex":0.0006564886,"about_ca_system_score_gemma":0.00090643665,"threshold_uncertainty_score":0.008641422},"labels":[],"label_agreement":null},{"id":"W4402853960","doi":"10.1093/nargab/lqae122","title":"eQTL-Detect: nextflow-based pipeline for eQTL detection in modular format with sharable and parallelizable scripts","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genetic Mapping and Diversity in Plants and Animals","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":"University of Alberta; Agriculture and Agri-Food Canada","funders":"European Commission","keywords":"Pipeline (software); Expression quantitative trait loci; Computer science; Workflow; Modular design; Scripting language; Parallelizable manifold; Orchestration; Data mining; Database; Programming language; Biology; Gene","score_opus":0.00964774568446424,"score_gpt":0.20494182402514696,"score_spread":0.19529407834068271,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4402853960","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":[0.009516222,0.00012094372,0.74754846,0.00031312072,0.0001744146,0.0005426875,0.009506526,0.23019546,0.0020822154],"genre_scores_gemma":[0.077728234,0.00021738112,0.84656674,0.0008669544,0.00009356654,0.0027220408,0.04215153,0.024497291,0.005156263],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9987657,0.00019290377,0.00013663371,0.0004691352,0.00028191967,0.00015369279],"domain_scores_gemma":[0.9980221,0.00084646186,0.00018169098,0.00039576791,0.00037120286,0.00018285215],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0043732333,0.0017893405,0.0008577973,0.0013718237,0.0011238514,0.0018841488,0.002340611,0.0010244567,0.0150794815],"category_scores_gemma":[0.0056720455,0.0013629755,0.0017918233,0.0007975674,0.00090745196,0.0026323162,0.0022618468,0.0031079813,0.006375938],"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":[0.005440908,0.0011228756,0.01688087,0.0021826506,0.00082196784,0.0014832112,0.0021979702,0.030456439,0.3413619,0.03703087,0.21478407,0.34623626],"study_design_scores_gemma":[0.0009843284,0.0004912264,0.0061457083,0.00018923543,0.0002376579,0.0007663645,0.00015257119,0.36503077,0.37961575,0.043295797,0.20256367,0.0005270007],"about_ca_topic_score_codex":0.0025656645,"about_ca_topic_score_gemma":0.0021247256,"teacher_disagreement_score":0.0150794815,"about_ca_system_score_codex":0.00116745,"about_ca_system_score_gemma":0.0024847798,"threshold_uncertainty_score":0.050445855},"labels":[],"label_agreement":null},{"id":"W4404333429","doi":"10.1093/nargab/lqae145","title":"Refining SARS-CoV-2 intra-host variation by leveraging large-scale sequencing data","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"SARS-CoV-2 and COVID-19 Research","field":"Medicine","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":"Université de Montréal; Mila - Quebec Artificial Intelligence Institute; Montreal Heart Institute","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Fonds de Recherche du Québec - Santé; Institut de Valorisation des Données; Canada Foundation for Innovation","keywords":"Workflow; Computational biology; Host (biology); DNA sequencing; Computer science; Cluster analysis; Visualization; Biology; Data mining; Genetics; Artificial intelligence; DNA; Database","score_opus":0.06431370519734829,"score_gpt":0.3304385491763903,"score_spread":0.266124843979042,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4404333429","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":[0.5615893,0.0010556111,0.4216435,0.00056181935,0.00012013118,0.00031277345,0.0059279986,0.005167118,0.0036217074],"genre_scores_gemma":[0.55348617,0.00045172556,0.43649867,0.00018415802,0.000042862983,0.00015670928,0.007021572,0.0008231733,0.0013349463],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988398,0.00024289319,0.0001122927,0.0004016035,0.00031067996,0.00009277963],"domain_scores_gemma":[0.9980338,0.00070080446,0.00033098008,0.0003020545,0.0005191008,0.0001132442],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0022915965,0.00074539153,0.0005699839,0.0024270758,0.00077490095,0.0017430709,0.0004582548,0.00066366285,0.0013023664],"category_scores_gemma":[0.005055727,0.00034625625,0.0009277893,0.0018707955,0.0003367391,0.0009033004,0.0010892561,0.000974689,0.0007709794],"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":[0.0008124709,0.00032788527,0.11840267,0.0006755296,0.0004259965,0.0009872953,0.0020919119,0.04112023,0.607427,0.0047289226,0.004109609,0.21889053],"study_design_scores_gemma":[0.00008999068,0.00045290895,0.16208579,0.00021825143,0.00033641385,0.0014763554,0.0016769135,0.43230134,0.3369024,0.024477115,0.039700106,0.00028238687],"about_ca_topic_score_codex":0.0032906078,"about_ca_topic_score_gemma":0.01058241,"teacher_disagreement_score":0.0032906078,"about_ca_system_score_codex":0.0005305892,"about_ca_system_score_gemma":0.0011543069,"threshold_uncertainty_score":0.012119234},"labels":[],"label_agreement":null},{"id":"W4405282922","doi":"10.1093/nargab/lqae161","title":"Water-mediated ribonucleotide–amino acid pairs and higher-order structures at the RNA–protein interface: analysis of the crystal structure database and a topological classification","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","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":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Interface (matter); Topology (electrical circuits); Amino acid; Database; Order (exchange); Ribonucleotide; Computer science; Crystal structure; RNA; Computational biology; Data mining; Biology; Chemistry; Crystallography; Mathematics; Biochemistry; Nucleotide; Combinatorics; Operating system; Gene","score_opus":0.0112042506949473,"score_gpt":0.2291308502527272,"score_spread":0.2179265995577799,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405282922","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":[0.92254573,0.0020030683,0.008782905,0.00021783642,0.000025059135,0.00009577242,0.061591726,0.0027042467,0.0020335636],"genre_scores_gemma":[0.7301756,0.0011902192,0.028407343,0.00007074957,0.000018958854,0.00015210878,0.23901428,0.000342539,0.0006281975],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9993801,0.00012786873,0.000057859314,0.00019810286,0.00015665534,0.000079381716],"domain_scores_gemma":[0.99894637,0.00038868893,0.0002370526,0.00017620555,0.00012660903,0.00012511277],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010057867,0.0006799725,0.00077490933,0.003912697,0.0005591726,0.0010526903,0.00067493203,0.00040286966,0.0023379105],"category_scores_gemma":[0.0025137998,0.00020359774,0.00082799036,0.004993764,0.00035162224,0.00095175003,0.00087562867,0.000594148,0.0008657768],"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":[0.0033770273,0.0019250404,0.42436668,0.0060945572,0.0015089402,0.0014825141,0.0011409164,0.08792114,0.13909024,0.019896513,0.13011985,0.18307662],"study_design_scores_gemma":[0.00036429087,0.0006370865,0.21952665,0.00023625887,0.00048911094,0.0009619779,0.00091845193,0.6588827,0.035404,0.019786008,0.06262767,0.00016589473],"about_ca_topic_score_codex":0.0030152032,"about_ca_topic_score_gemma":0.006520015,"teacher_disagreement_score":0.003912697,"about_ca_system_score_codex":0.00046233975,"about_ca_system_score_gemma":0.00069670595,"threshold_uncertainty_score":0.007821083},"labels":[],"label_agreement":null},{"id":"W4405620286","doi":"10.1093/nargab/lqae176","title":"SARS-CoV-2 Illumina GeNome Assembly Line (SIGNAL), a Snakemate workflow for rapid and bulk analysis of Illumina sequencing of SARS-CoV-2 genomes","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"SARS-CoV-2 and COVID-19 Research","field":"Medicine","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":"Ontario Institute for Cancer Research; Canada Research Chairs; Vector Institute; University of Saskatchewan; University of Manitoba; Public Health Agency of Canada; University Health Network; Health Sciences Centre; Sunnybrook Health Science Centre; University of Toronto; Dalhousie University; Perimeter Institute; McMaster University","funders":"Genome Canada; Canadian Institutes of Health Research; Michael G. DeGroote Institute for Infectious Disease Research, McMaster University; University of Toronto; Natural Sciences and Engineering Research Council of Canada; McMaster University","keywords":"Workflow; Illumina dye sequencing; Whole genome sequencing; DNA sequencing; Genome; Personal genomics; Computer science; Computational biology; Biology; Genetics; Database; DNA; Gene","score_opus":0.057473568328843525,"score_gpt":0.3294186607319,"score_spread":0.2719450924030565,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405620286","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":[0.030804222,0.0010673333,0.79417175,0.0006077948,0.00044593556,0.0013001182,0.04981844,0.10395129,0.017833168],"genre_scores_gemma":[0.041983765,0.00056089234,0.82935566,0.0009090495,0.00010559191,0.0020809083,0.0969687,0.016738152,0.01129738],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9980901,0.00033691412,0.00016242646,0.0006810917,0.0005820988,0.00014738074],"domain_scores_gemma":[0.99859434,0.00036371555,0.00017117274,0.00028720437,0.00043554822,0.00014801607],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0032611566,0.0013196995,0.00089168985,0.0013055762,0.0011032715,0.0017353154,0.0012071341,0.00072219013,0.011250144],"category_scores_gemma":[0.0036747777,0.0012475562,0.001628083,0.0011571805,0.00047940528,0.0009885634,0.0018540518,0.0024184885,0.013656097],"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":[0.0023372856,0.00021061122,0.009707727,0.0018345773,0.00053879817,0.00068296795,0.0016816687,0.007421254,0.4579982,0.010913758,0.29630226,0.21037082],"study_design_scores_gemma":[0.0004013343,0.0006863163,0.012775945,0.00030123693,0.00020286618,0.0010186123,0.00025187057,0.07052416,0.30006993,0.0098184105,0.60350734,0.00044191635],"about_ca_topic_score_codex":0.0029264244,"about_ca_topic_score_gemma":0.0070025683,"teacher_disagreement_score":0.011250144,"about_ca_system_score_codex":0.0005369656,"about_ca_system_score_gemma":0.0018083337,"threshold_uncertainty_score":0.037635446},"labels":[],"label_agreement":null},{"id":"W4408163552","doi":"10.1093/nargab/lqaf011","title":"iModEst: disentangling -omic impacts on gene expression variation across genes and tissues","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Chromatin Dynamics","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":"Canadian Institute for Advanced Research; SickKids Foundation; Vector Institute; University of Toronto","funders":"","keywords":"Gene; Biology; Variation (astronomy); Genetics; Gene expression; Computational biology; Genetic variation; Evolutionary biology","score_opus":0.00522141975189719,"score_gpt":0.24582336438267025,"score_spread":0.24060194463077306,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408163552","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":[0.37262893,0.0019624748,0.59062225,0.0006338563,0.000073810734,0.00008274526,0.0231681,0.0062088696,0.0046189185],"genre_scores_gemma":[0.8598328,0.0007441342,0.116779126,0.00021672624,0.000054418837,0.00020207293,0.019534089,0.0009867186,0.00164995],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992053,0.00028371933,0.000033232798,0.00026687328,0.00015987884,0.0000511021],"domain_scores_gemma":[0.998133,0.0013261231,0.00013338214,0.0002646695,0.00008595445,0.000056881872],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002179487,0.00092945754,0.0010264096,0.0017159275,0.00039243122,0.001497647,0.0007714171,0.00037578648,0.0030477117],"category_scores_gemma":[0.003991315,0.0003126316,0.0017987586,0.0015474763,0.00041302305,0.0005890289,0.0014853609,0.0007679588,0.000497839],"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":[0.001209423,0.00019233522,0.2842077,0.0010837843,0.006087729,0.0013090955,0.0004857575,0.38543376,0.08390923,0.02591694,0.00987324,0.20029104],"study_design_scores_gemma":[0.00005167333,0.00021514791,0.0679785,0.000064462256,0.0007228297,0.0004338988,0.00014989096,0.8700792,0.01527005,0.02930398,0.015677264,0.00005304407],"about_ca_topic_score_codex":0.0033798814,"about_ca_topic_score_gemma":0.00398536,"teacher_disagreement_score":0.0033798814,"about_ca_system_score_codex":0.0006202485,"about_ca_system_score_gemma":0.00071017136,"threshold_uncertainty_score":0.011526346},"labels":[],"label_agreement":null},{"id":"W4408249210","doi":"10.1093/nargab/lqaf016","title":"BacTermFinder: a comprehensive and general bacterial terminator finder using a CNN ensemble","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","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":"Memorial University of Newfoundland","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada","keywords":"Terminator (solar); False positive paradox; Computer science; Bacterial transcription; Bacterial genome size; Convolutional neural network; Artificial intelligence; Computational biology; Bacterial protein; Machine learning; Biology; Data mining; RNA; Genome; Gene; Genetics; RNA polymerase; Physics","score_opus":0.012491023769565504,"score_gpt":0.2440232268859487,"score_spread":0.2315322031163832,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408249210","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":[0.08074184,0.0032785318,0.7610054,0.00036168128,0.00042759685,0.00038907304,0.017512864,0.12938622,0.0068968236],"genre_scores_gemma":[0.20638263,0.0014430925,0.72146624,0.0007880926,0.000121720754,0.0007844884,0.05409346,0.004536648,0.010383565],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99957794,0.00003027136,0.000023390867,0.00018722878,0.00011501745,0.00006626606],"domain_scores_gemma":[0.9996792,0.00011624426,0.00004835326,0.000047546124,0.000081894985,0.000026795784],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00093508593,0.0026765435,0.0010981509,0.0015051758,0.0005833508,0.0009669993,0.00198389,0.0014468038,0.004486693],"category_scores_gemma":[0.001955223,0.0009744935,0.0016939418,0.00062908104,0.00024705168,0.0015664055,0.0011858416,0.0018537474,0.0028795952],"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":[0.0013686246,0.00037021353,0.012761348,0.0016133647,0.00066249334,0.00052003714,0.00019200912,0.14956018,0.1952424,0.004507742,0.04860563,0.58459604],"study_design_scores_gemma":[0.00007511042,0.00018914408,0.0018342495,0.00008080929,0.000112882575,0.00019185132,0.000036333637,0.91963017,0.059844475,0.003521418,0.014412485,0.000070994945],"about_ca_topic_score_codex":0.006730518,"about_ca_topic_score_gemma":0.010457607,"teacher_disagreement_score":0.006730518,"about_ca_system_score_codex":0.0011408818,"about_ca_system_score_gemma":0.001223454,"threshold_uncertainty_score":0.015009463},"labels":[],"label_agreement":null},{"id":"W4409905545","doi":"10.1093/nargab/lqaf048","title":"Discovering governing equations of biological systems through representation learning and sparse model discovery","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Gene Regulatory Network Analysis","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"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 Waterloo","funders":"","keywords":"Autoencoder; Computer science; Identification (biology); Representation (politics); Gene regulatory network; Artificial intelligence; Machine learning; Dynamical systems theory; Biological network; Complex system; Biological data; Nonlinear system; PageRank; Systems biology; Key (lock); Theoretical computer science; Deep learning; Data mining; Computational biology; Gene; Biology; Bioinformatics; Gene expression","score_opus":0.02124096170560071,"score_gpt":0.26378734613985466,"score_spread":0.24254638443425394,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4409905545","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":[0.019947357,0.00012432132,0.97882926,0.00024677932,0.00000933255,0.000020069305,0.00009696338,0.00030460456,0.0004212777],"genre_scores_gemma":[0.56877345,0.000509691,0.42714947,0.00025092298,0.00009164801,0.00022838212,0.0008219013,0.00014435442,0.002030259],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99970526,0.000112848815,0.000013522751,0.00007531653,0.000066838016,0.000026250438],"domain_scores_gemma":[0.99820185,0.0013348609,0.00017147268,0.00013918978,0.00009816619,0.00005444536],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008577793,0.0007008698,0.00081685215,0.00063764415,0.0003559695,0.0008801303,0.0009282664,0.0009550565,0.0008272991],"category_scores_gemma":[0.003336823,0.00058524,0.0008040227,0.00041951987,0.0009834835,0.0010697911,0.0010622309,0.0020972327,0.00025832836],"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":[0.000015500193,0.000024894845,0.0005663115,0.000039298728,0.00003095331,0.000052812975,0.00003017371,0.97016114,0.0019645286,0.01212758,0.00041692553,0.014569926],"study_design_scores_gemma":[0.0000010805358,0.0000024928365,0.00003235091,8.441054e-7,8.630472e-7,0.0000035552266,0.000001337318,0.9949818,0.0001465765,0.0047683227,0.00005939998,0.0000013882228],"about_ca_topic_score_codex":0.0036841056,"about_ca_topic_score_gemma":0.0049014757,"teacher_disagreement_score":0.0036841056,"about_ca_system_score_codex":0.0006251146,"about_ca_system_score_gemma":0.00081127934,"threshold_uncertainty_score":0.0073252916},"labels":[],"label_agreement":null},{"id":"W4411754998","doi":"10.1093/nargab/lqaf087","title":"MOLGENIS VIP: an end-to-end DNA variant interpretation pipeline for research and diagnostics configurable to support rapid implementation of new methods","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"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":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek","keywords":"Scalability; Pipeline (software); Computer science; Data science; DNA sequencing; Genome; Variety (cybernetics); Software; Protocol (science); Computational biology; Data mining; Biology; Medicine; Artificial intelligence; Genetics; Database; Gene","score_opus":0.04216451958384585,"score_gpt":0.39957096283162724,"score_spread":0.35740644324778137,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4411754998","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":[0.013583393,0.0012013111,0.4916122,0.0008569147,0.0006433602,0.0008150819,0.07874364,0.40482324,0.0077208863],"genre_scores_gemma":[0.05228916,0.0006845821,0.7283022,0.0015340149,0.000250436,0.0015245889,0.16633676,0.043561004,0.005517303],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9972645,0.0004527934,0.00022211533,0.0011542444,0.0006894731,0.00021685625],"domain_scores_gemma":[0.99526227,0.0019412538,0.0006315327,0.00097019883,0.00065500883,0.00053966005],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0059494926,0.002691091,0.0017375024,0.0033004985,0.0013711244,0.0035767846,0.0031159932,0.0019667954,0.022115104],"category_scores_gemma":[0.01128068,0.0023060213,0.0023133943,0.0026233979,0.0010975782,0.002326436,0.0049423333,0.0040323553,0.020624962],"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":[0.0061941235,0.0003324404,0.025018195,0.004232705,0.0015916507,0.0020955352,0.0021025112,0.012478947,0.1346776,0.0111962855,0.49565688,0.3044231],"study_design_scores_gemma":[0.0017294582,0.0009344955,0.035405222,0.0009780632,0.00060106214,0.0034526596,0.0006093277,0.12760726,0.1884452,0.053057764,0.58622944,0.00095009466],"about_ca_topic_score_codex":0.0023587672,"about_ca_topic_score_gemma":0.003013035,"teacher_disagreement_score":0.022115104,"about_ca_system_score_codex":0.001037132,"about_ca_system_score_gemma":0.00302635,"threshold_uncertainty_score":0.07398242},"labels":[],"label_agreement":null},{"id":"W4412492594","doi":"10.1093/nargab/lqaf099","title":"Modeling RNA duplex dynamics with Gibbs sampling enhances base-pair prediction accuracy and reveals structural activity profiles","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","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":"Institute for Research in Immunology and Cancer","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Google","keywords":"RNA; Base pair; Duplex (building); Nucleic acid secondary structure; Nucleic acid structure; Biological system; Statistical physics; Algorithm; Computational biology; Computer science; Physics; Biology; Genetics; DNA","score_opus":0.011571702968631912,"score_gpt":0.24118674235263773,"score_spread":0.22961503938400582,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4412492594","genre_codex":"empirical","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":[0.776251,0.0002682504,0.21958028,0.0001426051,0.000029439538,0.000036618712,0.00027868955,0.0016411095,0.0017720856],"genre_scores_gemma":[0.96636283,0.000059880036,0.032837063,0.00003556121,0.000008130495,0.000024676638,0.00027169363,0.00008429496,0.00031588264],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99980515,0.00006527856,0.0000079263955,0.000053133786,0.000045992052,0.00002245083],"domain_scores_gemma":[0.9988796,0.0007895301,0.00006498202,0.00010428488,0.00009961577,0.000061904655],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000871249,0.00044292386,0.00044691397,0.0004657182,0.00026496287,0.0005314595,0.00055441144,0.00058188726,0.0007460759],"category_scores_gemma":[0.002509146,0.0002077333,0.00032549075,0.0002802965,0.00035573982,0.0006637581,0.00032265636,0.00046215512,0.00018728938],"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":[0.0001098433,0.000053593474,0.004181675,0.000031703155,0.000025008087,0.000028887147,0.000034106524,0.975093,0.0052612196,0.002129564,0.0002859307,0.012765466],"study_design_scores_gemma":[0.0000022273045,0.000007978613,0.00012690957,7.4256974e-7,0.0000013420272,0.0000026915507,0.0000021274907,0.9983865,0.0009343079,0.000487767,0.000045274315,0.0000021307212],"about_ca_topic_score_codex":0.0058824294,"about_ca_topic_score_gemma":0.005808964,"teacher_disagreement_score":0.0058824294,"about_ca_system_score_codex":0.00054061465,"about_ca_system_score_gemma":0.0006816848,"threshold_uncertainty_score":0.011696398},"labels":[],"label_agreement":null},{"id":"W4413342468","doi":"10.1093/nargab/lqaf108","title":"Explicit Scale Simulation for analysis of RNA-sequencing count data with ALDEx2","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","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":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; National Institutes of Health","keywords":"Computer science; Normalization (sociology); Scale (ratio); Sample size determination; Data mining; Sample (material); Cutoff; False positive paradox; Statistics; Machine learning; Mathematics","score_opus":0.026348781228629416,"score_gpt":0.27803068670699643,"score_spread":0.251681905478367,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413342468","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":[0.19132322,0.00025790167,0.79748756,0.0007809584,0.00016740996,0.00012232835,0.0011317072,0.00550612,0.003222809],"genre_scores_gemma":[0.6949153,0.00016986199,0.2979,0.0004899587,0.000064083026,0.00056670903,0.0017222008,0.0015584923,0.0026133596],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99863297,0.0007783763,0.000053891054,0.00019281008,0.000255888,0.000086117245],"domain_scores_gemma":[0.9877434,0.010539219,0.0003511666,0.00056901906,0.00056760816,0.00022967666],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00589844,0.0008697066,0.00094715954,0.0006177568,0.0008305249,0.0016482625,0.001730008,0.0014165201,0.0038427904],"category_scores_gemma":[0.016747357,0.0006114781,0.001343529,0.00075028074,0.0013144647,0.0012220447,0.0015663399,0.002122942,0.0006582406],"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":[0.00015981086,0.000052625746,0.004711994,0.00009710172,0.00006563617,0.00012599952,0.00016904765,0.96881783,0.0026635132,0.016620664,0.0010717816,0.005444075],"study_design_scores_gemma":[0.0000063480575,0.0000050042577,0.00014460381,0.0000027067354,0.0000022275233,0.000006070648,0.000008570298,0.9960632,0.00050849357,0.0029690445,0.0002782111,0.0000054823076],"about_ca_topic_score_codex":0.007877944,"about_ca_topic_score_gemma":0.0056916196,"teacher_disagreement_score":0.007877944,"about_ca_system_score_codex":0.001402272,"about_ca_system_score_gemma":0.000995912,"threshold_uncertainty_score":0.03119433},"labels":[],"label_agreement":null},{"id":"W4414887972","doi":"10.1093/nargab/lqaf132","title":"VIRUS-MVP: a framework for comprehensive surveillance of viral mutations and their functional impacts","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Viral Infections and Outbreaks Research","field":"Medicine","cited_by":1,"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; University of British Columbia; University of Calgary; Simon Fraser University","funders":"Canadian Institutes of Health Research; Simon Fraser University; Michael Smith Health Research BC; Genome Canada","keywords":"Genomics; Functional genomics; Annotation; Resource (disambiguation); Mutation; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Genome; Modular design","score_opus":0.032538701256875954,"score_gpt":0.3202955777649885,"score_spread":0.28775687650811255,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4414887972","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":[0.0029442809,0.0009325063,0.8256953,0.0006412515,0.00021781113,0.00051744847,0.017179482,0.14738639,0.0044855545],"genre_scores_gemma":[0.04936736,0.002557166,0.8594089,0.00091691857,0.00018482447,0.0018872891,0.058475442,0.022514218,0.004687863],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9982134,0.00042891628,0.00025292672,0.0004593005,0.0004674056,0.00017797502],"domain_scores_gemma":[0.99793494,0.0008498899,0.00019400324,0.00046228693,0.000258678,0.00030024146],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0058688847,0.0025796683,0.0018436629,0.0038699128,0.0013306988,0.0048005288,0.0039417245,0.0024864185,0.012846707],"category_scores_gemma":[0.007498861,0.0018090002,0.0036228455,0.0020054518,0.0013330703,0.0040599215,0.006721331,0.003744483,0.008313132],"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":[0.002409315,0.000604372,0.018263975,0.007435571,0.0014665603,0.0031682611,0.00353933,0.08295561,0.05981154,0.20190792,0.28160134,0.33683622],"study_design_scores_gemma":[0.00037019607,0.00021979836,0.0045048227,0.001101344,0.00024771973,0.001232863,0.0004421967,0.252959,0.016887302,0.22989053,0.49166608,0.0004781476],"about_ca_topic_score_codex":0.008106047,"about_ca_topic_score_gemma":0.009989901,"teacher_disagreement_score":0.012846707,"about_ca_system_score_codex":0.0016585009,"about_ca_system_score_gemma":0.003928344,"threshold_uncertainty_score":0.0429765},"labels":[],"label_agreement":null},{"id":"W4416380425","doi":"10.1093/nargab/lqaf148","title":"Genomic islands in <i>Pseudomonas</i> encode modular hotspots of defence and anti-defence systems","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Vibrio bacteria research studies","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":"McMaster University","funders":"Medical Research Council; Wellcome Trust; UK Research and Innovation; European Molecular Biology Organization","keywords":"Pathogenicity island; Virulence; Lytic cycle; ENCODE; Genomic island; Arms race; Genome","score_opus":0.0066160139177305925,"score_gpt":0.2331474672512933,"score_spread":0.2265314533335627,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4416380425","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":[0.97062534,0.0022495908,0.016712582,0.000265592,0.000040907482,0.00007593716,0.0040081823,0.00072815944,0.0052936375],"genre_scores_gemma":[0.9815694,0.0004950206,0.009728426,0.00011429822,0.000028098595,0.000023629662,0.0060379175,0.00009087542,0.0019123188],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.9998795,0.000014073494,0.00000720635,0.00004421785,0.0000314728,0.000023536142],"domain_scores_gemma":[0.99958354,0.00011288771,0.00013328229,0.00004186764,0.000050186052,0.00007825389],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00014874953,0.00028847824,0.00019581523,0.0005702002,0.0002233565,0.00044916983,0.00020318059,0.00027012403,0.0020132095],"category_scores_gemma":[0.00046826692,0.00014296334,0.00021987286,0.00057190546,0.00017852288,0.00027969346,0.00042923752,0.00030794824,0.00067875755],"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":[0.00030678997,0.000040090355,0.021037335,0.00034445085,0.000047688365,0.0005353087,0.00018718101,0.00062480423,0.9439156,0.0010694489,0.00080407877,0.03108727],"study_design_scores_gemma":[0.000056467936,0.00034476758,0.6038354,0.00017851575,0.0001932741,0.0037766693,0.0007591287,0.0073602125,0.34226996,0.003542301,0.037585102,0.00009810153],"about_ca_topic_score_codex":0.0013660906,"about_ca_topic_score_gemma":0.0020994742,"teacher_disagreement_score":0.0020132095,"about_ca_system_score_codex":0.00027767627,"about_ca_system_score_gemma":0.0002570351,"threshold_uncertainty_score":0.006734848},"labels":[],"label_agreement":null},{"id":"W4417268800","doi":"10.1093/nargab/lqaf170","title":"<tt>SAFARI</tt> : pangenome alignment of ancient DNA using purine/pyrimidine encodings","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","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":"Université Laval","funders":"National Human Genome Research Institute; Novo Nordisk Fonden; Novo Nordisk","keywords":"Spurious relationship; Ancient DNA; Extant taxon; Genome; DNA; Divergence (linguistics)","score_opus":0.012432450812789271,"score_gpt":0.23299770210798124,"score_spread":0.22056525129519197,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417268800","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":[0.015022926,0.0006943512,0.55829394,0.00047100722,0.0006873802,0.0003064692,0.028260468,0.38954294,0.0067205573],"genre_scores_gemma":[0.05219991,0.0005156575,0.81224304,0.00063034816,0.00014969154,0.00061123644,0.0717519,0.053994592,0.007903533],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9993368,0.00008449935,0.00008218453,0.00026518488,0.0001867537,0.000044469296],"domain_scores_gemma":[0.99920887,0.00026473074,0.00014259324,0.00015048713,0.0001733496,0.00006008368],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012531935,0.001996079,0.00096425123,0.0020318988,0.00084207486,0.0014873212,0.0026503883,0.0018399636,0.03229816],"category_scores_gemma":[0.0056495755,0.0009987943,0.0011402147,0.0022713817,0.0005959438,0.0018323506,0.0018629592,0.0021581212,0.013885329],"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":[0.0014189298,0.00018374289,0.004741402,0.002801321,0.00047954053,0.000962782,0.0011662574,0.011511697,0.15912962,0.01927501,0.46548802,0.33284166],"study_design_scores_gemma":[0.00048357664,0.0005159275,0.006785372,0.00025987776,0.00014543328,0.0013398627,0.00020319893,0.20981207,0.23426205,0.023682615,0.5221419,0.00036813668],"about_ca_topic_score_codex":0.0018272785,"about_ca_topic_score_gemma":0.005717763,"teacher_disagreement_score":0.03229816,"about_ca_system_score_codex":0.0006565809,"about_ca_system_score_gemma":0.00079466304,"threshold_uncertainty_score":0.10804808},"labels":[],"label_agreement":null},{"id":"W4417268808","doi":"10.1093/nargab/lqaf175","title":"Genome sequence assembly and annotation of <i>MATA</i> and <i>MATB</i> strains of <i>Yarrowia lipolytica</i>","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Microbial Metabolic Engineering and Bioproduction","field":"Biochemistry, Genetics and Molecular Biology","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; National Institutes of Health; National Cancer Institute; Agricultural Research Service; University of Alberta; U.S. Department of Agriculture","keywords":"Genome; Genome project; Whole genome sequencing; Gene; DNA sequencing; Gene Annotation; Sequence assembly; Reference genome; Yeast","score_opus":0.006117237744006283,"score_gpt":0.22040193155462012,"score_spread":0.21428469381061385,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4417268808","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":[0.42569363,0.0075294026,0.11239347,0.0008345304,0.00053799665,0.0014796666,0.4391745,0.00670284,0.005653877],"genre_scores_gemma":[0.13202989,0.0032019634,0.120540984,0.00022913546,0.000092489085,0.0009275665,0.73698956,0.0012541874,0.0047342177],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9994708,0.000064883396,0.00006519495,0.00020134292,0.00012924158,0.00006851955],"domain_scores_gemma":[0.9994324,0.00007083109,0.00015083177,0.00005385766,0.00020504025,0.00008707896],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00045580423,0.0012476476,0.0010884264,0.0019025147,0.0007801678,0.0007752399,0.0007681747,0.0006530791,0.0020829996],"category_scores_gemma":[0.00096815213,0.0005258261,0.0012944019,0.0039587924,0.00024621078,0.00035269858,0.00041664636,0.0011817012,0.0033732923],"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":[0.0011259494,0.00012525215,0.002198424,0.0013439229,0.00007206289,0.0004245639,0.00024821388,0.0011862663,0.96989924,0.0003781391,0.004483286,0.01851472],"study_design_scores_gemma":[0.00035357822,0.000739344,0.13386394,0.000665224,0.0007688607,0.0012571521,0.00055125344,0.020464603,0.6930162,0.0009315048,0.14716129,0.00022700573],"about_ca_topic_score_codex":0.007280446,"about_ca_topic_score_gemma":0.0059348517,"teacher_disagreement_score":0.007280446,"about_ca_system_score_codex":0.0006975257,"about_ca_system_score_gemma":0.0019073143,"threshold_uncertainty_score":0.0144761205},"labels":[],"label_agreement":null},{"id":"W7117139568","doi":"10.1093/nargab/lqaf189","title":"Life at the extremes: maximally divergent microbes with similar genomic signatures linked to extreme environments","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","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":"University of Guelph; Western University; University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada","keywords":"Taxonomic rank; Genome; Extreme environment; Metagenomics; Genomics; Biological classification; Adaptation (eye); Phylogenomics; Trait","score_opus":0.01003646980641636,"score_gpt":0.19776717148554032,"score_spread":0.18773070167912395,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W7117139568","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":[0.97944653,0.0007022096,0.01816801,0.000104909624,0.00000860814,0.000018417595,0.00063339784,0.00021857892,0.00069942453],"genre_scores_gemma":[0.97611564,0.00029665304,0.021884155,0.0000899962,0.000012359153,0.000018741808,0.0014158897,0.00003504494,0.00013154851],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.999681,0.000057799673,0.000022289265,0.00013838721,0.000060847862,0.00003968371],"domain_scores_gemma":[0.99949753,0.00013490916,0.00021381764,0.00004982456,0.00003247691,0.00007136263],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003979146,0.00035201636,0.0005996256,0.00064402376,0.00041883878,0.0011112368,0.00031155391,0.0004910476,0.000807004],"category_scores_gemma":[0.001054782,0.00024058025,0.0004044244,0.0009104902,0.00044454035,0.00058714807,0.0009810751,0.00060337404,0.00039645273],"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":[0.0015919403,0.00016721348,0.37035555,0.0006553519,0.00033748525,0.0019300035,0.0006873998,0.011020631,0.51104254,0.0025080699,0.0010449388,0.098658904],"study_design_scores_gemma":[0.00005783237,0.0006153357,0.7325054,0.00020252701,0.00040681835,0.005427054,0.001695154,0.121196315,0.1150351,0.014006085,0.008687134,0.0001651926],"about_ca_topic_score_codex":0.00032622018,"about_ca_topic_score_gemma":0.0007114479,"teacher_disagreement_score":0.0011112368,"about_ca_system_score_codex":0.0001537259,"about_ca_system_score_gemma":0.00021691002,"threshold_uncertainty_score":0.0026997328},"labels":[],"label_agreement":null},{"id":"W7117923635","doi":"10.1093/nargab/lqaf208","title":"G-quadruplex structures as modulators of alternative promoter usage","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"DNA and Nucleic Acid Chemistry","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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":"Vermont Agency of Natural Resources; Institut National Du Cancer; Agence Nationale de la Recherche; CNIB","keywords":"Promoter; Gene isoform; Gene; Transcription (linguistics); Transcription factor; Regulation of gene expression; DNA; Gene expression","score_opus":0.004621207016402525,"score_gpt":0.23251254045158906,"score_spread":0.22789133343518653,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W7117923635","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":[0.98993987,0.002210228,0.005983541,0.000067941655,0.000019834322,0.000022868606,0.00062928715,0.00021485741,0.0009116617],"genre_scores_gemma":[0.9942122,0.00057608145,0.0038946045,0.00006891278,0.000010457702,0.000021942027,0.00072273443,0.000029673463,0.0004632338],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9997714,0.00004429136,0.0000140909615,0.0000765727,0.000059960505,0.00003370364],"domain_scores_gemma":[0.9996619,0.00011598468,0.00011695592,0.000024891071,0.00003734558,0.000042984666],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00028121352,0.00020919442,0.000281958,0.0003496422,0.00014138654,0.0003968441,0.0001603474,0.00022079343,0.0008722479],"category_scores_gemma":[0.0005116996,0.00020393463,0.0001979034,0.0002577163,0.00021492894,0.00021584856,0.00024041283,0.00037357726,0.00027585923],"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":[0.00021574009,0.000016198448,0.005923923,0.000110212066,0.000024781462,0.00007797822,0.0000397585,0.0006484278,0.9867964,0.00030940675,0.00010746134,0.005729635],"study_design_scores_gemma":[0.00003169131,0.00046805537,0.049745373,0.000029154808,0.00008526034,0.0004716247,0.00009795808,0.012572617,0.92937887,0.0009909143,0.0060947263,0.000033757464],"about_ca_topic_score_codex":0.00025977226,"about_ca_topic_score_gemma":0.00067735073,"teacher_disagreement_score":0.0008722479,"about_ca_system_score_codex":0.0002469702,"about_ca_system_score_gemma":0.00017573418,"threshold_uncertainty_score":0.002918005},"labels":[],"label_agreement":null}]}