{"meta":{"query_hash":"d9fbe9c57fec","filters":{"venue":"Computer applications in the biosciences"},"cohort_total":19,"direct_labels_cover":0,"predictions_cover":19,"exported":19,"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/d9fbe9c57fec","api":"https://metacan.xera.ac/api/v1/cohort?venue=Computer+applications+in+the+biosciences"},"results":[{"id":"W2096149709","doi":"10.1093/bioinformatics/bti1037","title":"Reversal distance for partially ordered genomes","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Genome Rearrangement Algorithms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":27,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Ottawa","funders":"","keywords":"Genome; Position (finance); Chromosome; Computer science; Representation (politics); Order (exchange); Gene; Mutation; Genetics; Computational biology; Biology; Algorithm","score_opus":0.014496985183509515,"score_gpt":0.27133429649502283,"score_spread":0.2568373113115133,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2096149709","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":[0.18995765,0.0006311771,0.7974959,0.0009434182,0.000056353118,0.00016784687,0.0015663591,0.0015523972,0.007628985],"genre_scores_gemma":[0.69490874,0.00042821997,0.29470038,0.00028180378,0.00007179704,0.00019876825,0.004676421,0.00032190155,0.0044119675],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99886584,0.00028623946,0.00007995704,0.00034831982,0.000334651,0.000085161],"domain_scores_gemma":[0.99233586,0.0054230737,0.0006555734,0.00074249075,0.00060153403,0.00024144305],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015478845,0.00051805674,0.0007455262,0.0018808282,0.00063244934,0.0013168707,0.0016273941,0.00097226456,0.0047497707],"category_scores_gemma":[0.014152646,0.00035013538,0.0008449237,0.0012538227,0.0011353473,0.002365075,0.0013968326,0.0014058006,0.00070231996],"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":[0.0004084943,0.00012078742,0.007507913,0.00041574964,0.000067639776,0.00035701648,0.00042000916,0.50579005,0.004643327,0.3123192,0.0041239117,0.16382588],"study_design_scores_gemma":[0.00003979311,0.00012124881,0.0012127571,0.000033782926,0.000017321998,0.00029259516,0.000108780376,0.6292209,0.0030758753,0.3615016,0.00434765,0.000027704664],"about_ca_topic_score_codex":0.0021534734,"about_ca_topic_score_gemma":0.0017188577,"teacher_disagreement_score":0.0047497707,"about_ca_system_score_codex":0.0014782477,"about_ca_system_score_gemma":0.0011315403,"threshold_uncertainty_score":0.015889585},"labels":[],"label_agreement":null},{"id":"W2099398646","doi":"10.1093/bioinformatics/bti170","title":"BIAS: Bioinformatics Integrated Application Software","year":2004,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University","funders":"","keywords":"Computer science; Software; Object (grammar); Data mining; Relational database; Bioinformatics; Data science; Software engineering; Artificial intelligence; Programming language; Biology","score_opus":0.013024605203504892,"score_gpt":0.24295220211584717,"score_spread":0.22992759691234227,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2099398646","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":[0.0028910276,0.0013639821,0.59076095,0.0015073534,0.000542207,0.0008653155,0.011125227,0.37098798,0.019955881],"genre_scores_gemma":[0.03440148,0.002237308,0.8087072,0.0033542947,0.00057678623,0.0033091658,0.039025266,0.08139926,0.026989209],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9963431,0.0006993397,0.00031442748,0.0005711744,0.001785149,0.00028682486],"domain_scores_gemma":[0.9922943,0.0033409111,0.0004924015,0.0010695682,0.0020689687,0.00073376775],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004996551,0.0013461162,0.0013168262,0.0023159678,0.0007913494,0.002986923,0.003957835,0.0016106501,0.038688123],"category_scores_gemma":[0.01753323,0.0015002971,0.0012684006,0.0032553074,0.00080868445,0.0041372203,0.00485039,0.0038819052,0.03777235],"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":[0.0008560835,0.00015044435,0.003193071,0.0020376726,0.00019713925,0.000337828,0.0004045243,0.0059169196,0.014393809,0.035978787,0.53348124,0.40305254],"study_design_scores_gemma":[0.00059378747,0.00018078263,0.003159872,0.0004650318,0.0001360139,0.000870515,0.00007145916,0.0486313,0.013485052,0.06502773,0.8671632,0.0002153166],"about_ca_topic_score_codex":0.0021769518,"about_ca_topic_score_gemma":0.0016189996,"teacher_disagreement_score":0.038688123,"about_ca_system_score_codex":0.001000049,"about_ca_system_score_gemma":0.0026248868,"threshold_uncertainty_score":0.12942469},"labels":[],"label_agreement":null},{"id":"W2103184043","doi":"10.1093/bioinformatics/bti321","title":"MtbRegList, a database dedicated to the analysis of transcriptional regulation in Mycobacterium tuberculosis","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Tuberculosis Research and Epidemiology","field":"Medicine","cited_by":49,"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 Québécois de la Recherche sur la Nature et les Technologies; U.S. National Library of Medicine; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Mycobacterium tuberculosis; Tuberculosis; Database; Computational biology; Biology; Computer science; Medicine","score_opus":0.034468796006127594,"score_gpt":0.3397371658698042,"score_spread":0.30526836986367656,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2103184043","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":[0.03591984,0.017640634,0.029424926,0.0010594998,0.00033094554,0.00037734496,0.8199856,0.07776483,0.017496364],"genre_scores_gemma":[0.03931412,0.0037701065,0.03442223,0.00041573576,0.000120995115,0.00037539884,0.914216,0.0028872425,0.0044783377],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9991804,0.000117012234,0.00014266396,0.00022661057,0.00023493783,0.000098366865],"domain_scores_gemma":[0.9983523,0.0006016979,0.0003436686,0.00024154097,0.00019816542,0.00026264825],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007618313,0.0015729272,0.002028554,0.004630342,0.0009530249,0.002071919,0.0019866393,0.0015817928,0.027740698],"category_scores_gemma":[0.0039343215,0.0006935192,0.0007682014,0.0055643655,0.00042472343,0.0013541623,0.0012771824,0.0010535677,0.021916453],"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":[0.0071104006,0.00040997504,0.018148731,0.016759371,0.00031949783,0.0022665872,0.00082465645,0.00431577,0.18689719,0.00500456,0.4714833,0.28646],"study_design_scores_gemma":[0.00050626684,0.00037836327,0.027790833,0.00089195854,0.0003983189,0.0021233365,0.00020973684,0.00567876,0.042277843,0.0034566126,0.9161503,0.00013783833],"about_ca_topic_score_codex":0.0024802112,"about_ca_topic_score_gemma":0.0034505036,"teacher_disagreement_score":0.027740698,"about_ca_system_score_codex":0.0006768923,"about_ca_system_score_gemma":0.001941754,"threshold_uncertainty_score":0.09280187},"labels":[],"label_agreement":null},{"id":"W2107500667","doi":"10.1093/bioinformatics/bti397","title":"A robust neural networks approach for spatial and intensity-dependent normalization of cDNA microarray data","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":39,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université Laval","funders":"","keywords":"Normalization (sociology); Sampling bias; Computer science; Microarray; Computational biology; Pattern recognition (psychology); Biology; Biological system; Statistics; Sample size determination; Artificial intelligence; Mathematics; Gene expression; Gene; Genetics","score_opus":0.04246975923685588,"score_gpt":0.2754001957445745,"score_spread":0.23293043650771866,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2107500667","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":[0.0038201204,0.00020927591,0.9945496,0.000087751265,0.000027673874,0.000028766317,0.00005987987,0.00085344637,0.0003635767],"genre_scores_gemma":[0.17974035,0.00051908987,0.8142499,0.000267134,0.00014413946,0.00047060417,0.0007650503,0.0003629382,0.0034808377],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99867225,0.0003158302,0.000080264836,0.00044453182,0.00039838682,0.00008876978],"domain_scores_gemma":[0.99830073,0.0008242769,0.00022047531,0.00012944003,0.00049040647,0.0000346589],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019683095,0.0013925226,0.0009170889,0.001054588,0.00044333297,0.00095999654,0.001685769,0.0011725405,0.0014674056],"category_scores_gemma":[0.0051234616,0.00053108606,0.0012258039,0.0013875325,0.00069489825,0.001060403,0.0010763782,0.0020816568,0.00064115634],"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":[0.000120652585,0.00006474015,0.0005921836,0.00013558067,0.0001493155,0.00007809986,0.000054242842,0.7243297,0.017254226,0.0044707246,0.001596413,0.25115407],"study_design_scores_gemma":[0.0000024494714,0.000013747596,0.00013910465,0.000005552577,0.000009372659,0.000012361084,0.0000034575448,0.9948021,0.0032281817,0.0014345401,0.0003402659,0.000008812022],"about_ca_topic_score_codex":0.0074982997,"about_ca_topic_score_gemma":0.0053708996,"teacher_disagreement_score":0.0074982997,"about_ca_system_score_codex":0.0013746709,"about_ca_system_score_gemma":0.0011203202,"threshold_uncertainty_score":0.014909327},"labels":[],"label_agreement":null},{"id":"W2108149347","doi":"10.1093/bioinformatics/bti330","title":"Recoverable one-dimensional encoding of three-dimensional protein structures","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":34,"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; Ministry of Education, Culture, Sports, Science and Technology","keywords":"Protein structure; Protein secondary structure; Sequence (biology); Protein structure prediction; Computer science; Protein tertiary structure; Set (abstract data type); Structural bioinformatics; Structural motif; Biological system; Algorithm; Crystallography; Biology; Chemistry; Genetics","score_opus":0.013849647384956375,"score_gpt":0.25164552140407603,"score_spread":0.23779587401911967,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2108149347","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":[0.4222305,0.00046722539,0.5718785,0.00025179356,0.000065746885,0.00003472583,0.0013395944,0.0014415471,0.002290311],"genre_scores_gemma":[0.6692373,0.00050800334,0.3270039,0.000048620204,0.00001582778,0.000048204358,0.0023174186,0.00014232456,0.0006784367],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99985874,0.000036282374,0.000011020182,0.000029367176,0.000049342565,0.000015301644],"domain_scores_gemma":[0.9994122,0.00017499458,0.00009663528,0.00020175299,0.00008637913,0.000028093253],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00024534774,0.00039346173,0.00047057445,0.0004507893,0.00025562692,0.0005874968,0.0005366957,0.0004255514,0.0007973721],"category_scores_gemma":[0.0016342641,0.0003061815,0.0005082791,0.0004922126,0.00038223778,0.0005814393,0.0005435738,0.0007620539,0.00037917992],"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":[0.0007195585,0.00017666421,0.006546287,0.00045246052,0.00010664533,0.00078121567,0.00034744892,0.45936412,0.3262179,0.0576054,0.003061034,0.14462136],"study_design_scores_gemma":[0.00001762429,0.00007269927,0.0014342836,0.00001804389,0.000017229559,0.00022278013,0.000046109482,0.9257908,0.056403518,0.013294408,0.0026472008,0.000035424204],"about_ca_topic_score_codex":0.0006917891,"about_ca_topic_score_gemma":0.00074307906,"teacher_disagreement_score":0.0007973721,"about_ca_system_score_codex":0.00038607977,"about_ca_system_score_gemma":0.00041409698,"threshold_uncertainty_score":0.00280118},"labels":[],"label_agreement":null},{"id":"W2108926360","doi":"10.1093/bioinformatics/bti1050","title":"Search for folding nuclei in native protein structures","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Executable; Computer science; Protein folding; Folding (DSP implementation); Simple (philosophy); Algorithm; Graph; Theoretical computer science; Topology (electrical circuits); Physics; Mathematics; Combinatorics; Engineering","score_opus":0.014176281649470424,"score_gpt":0.296746002949353,"score_spread":0.28256972129988256,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2108926360","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":[0.47399464,0.0038014732,0.5057297,0.0010310633,0.00012027232,0.00022586835,0.0049247975,0.0062706997,0.0039015461],"genre_scores_gemma":[0.7613398,0.00087511446,0.22420125,0.00021662544,0.000071993774,0.00013781058,0.011181927,0.0004506894,0.0015248198],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994265,0.00012567166,0.000037706533,0.00026750754,0.00010137775,0.00004138122],"domain_scores_gemma":[0.9971052,0.0016025318,0.00044840615,0.00034570627,0.0003049884,0.00019322032],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001115654,0.0009267635,0.0012342123,0.0016625902,0.0008500808,0.0011028202,0.0015740807,0.00163451,0.0026626003],"category_scores_gemma":[0.004373744,0.00037312933,0.00077216985,0.0012000758,0.001099497,0.0013038657,0.00076484133,0.00082349236,0.0013914759],"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":[0.0010025826,0.00041197694,0.06019296,0.002024365,0.0003028788,0.0010452357,0.00047046933,0.6054974,0.030603237,0.029042725,0.021821147,0.24758504],"study_design_scores_gemma":[0.000050699346,0.00010639639,0.003434665,0.00007109779,0.000047121797,0.0003333614,0.00017712761,0.93893605,0.014287227,0.036993127,0.0055381428,0.000025090336],"about_ca_topic_score_codex":0.0016453242,"about_ca_topic_score_gemma":0.0022402883,"teacher_disagreement_score":0.0026626003,"about_ca_system_score_codex":0.00054912944,"about_ca_system_score_gemma":0.0008684245,"threshold_uncertainty_score":0.0089072585},"labels":[],"label_agreement":null},{"id":"W2114368649","doi":"10.1093/bioinformatics/bti1045","title":"GenXHC: a probabilistic generative model for cross-hybridization compensation in high-density genome-wide microarray data","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":18,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Generative model; Computer science; Probabilistic logic; Generative grammar; Genome; Compensation (psychology); Computational biology; Statistical model; Microarray; Data mining; Artificial intelligence; Biology; Genetics; Gene; Gene expression","score_opus":0.04346260838480358,"score_gpt":0.31504882174478693,"score_spread":0.27158621335998334,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2114368649","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":[0.00544987,0.000086298984,0.99351716,0.0001324119,0.00001399627,0.000035276666,0.00007393866,0.0005529206,0.00013816748],"genre_scores_gemma":[0.25463998,0.0002152901,0.7388949,0.00062441267,0.00012321344,0.0005179847,0.0015579129,0.0005052006,0.002921069],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983277,0.00081362505,0.0000576724,0.00039896328,0.00030357807,0.00009843066],"domain_scores_gemma":[0.99574536,0.0033054948,0.00023721239,0.00030817828,0.00030692676,0.00009688201],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005630623,0.0008629601,0.0014022486,0.0010156309,0.0007801946,0.00092056405,0.003267631,0.0017504811,0.0018012406],"category_scores_gemma":[0.007864643,0.0011409308,0.0016549875,0.001059531,0.0016619388,0.0010304713,0.0017411237,0.0024560601,0.00046729954],"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":[0.00010447788,0.000044701446,0.0016290012,0.00005452171,0.00011637023,0.0000735011,0.00009971716,0.90729046,0.003148962,0.012020962,0.0018945195,0.07352285],"study_design_scores_gemma":[0.000007455482,0.0000062601016,0.00014962046,0.0000022579413,0.0000049853084,0.000017601616,0.000003301876,0.9952996,0.00046790473,0.0038010143,0.00023457568,0.0000054769926],"about_ca_topic_score_codex":0.010297651,"about_ca_topic_score_gemma":0.012716366,"teacher_disagreement_score":0.010297651,"about_ca_system_score_codex":0.0015746707,"about_ca_system_score_gemma":0.0015521329,"threshold_uncertainty_score":0.029777944},"labels":[],"label_agreement":null},{"id":"W2117003089","doi":"10.1093/bioinformatics/bti280","title":"Computational identification of human mitochondrial proteins based on homology to yeast mitochondrially targeted proteins","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Machine Learning in Bioinformatics","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":"Hospital for Sick Children","funders":"Canadian Institutes of Health Research; Hospital for Sick Children","keywords":"Biology; Human genome; Computational biology; Saccharomyces cerevisiae; Homology (biology); Genetics; Human proteins; Gene; Genome; Human mitochondrial genetics; Mitochondrion; UniGene; Expressed sequence tag","score_opus":0.008637087028099883,"score_gpt":0.2773267700318632,"score_spread":0.2686896830037633,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2117003089","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":[0.95595825,0.0013053624,0.0071399123,0.00037897716,0.000045687808,0.00010083265,0.031836472,0.001636532,0.0015979273],"genre_scores_gemma":[0.7853436,0.000856073,0.052894108,0.00020038923,0.00004144093,0.00019415835,0.15952489,0.00015274875,0.00079258584],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9997601,0.00003944863,0.000024825284,0.000113222726,0.000040420004,0.000022013539],"domain_scores_gemma":[0.99929404,0.0004170679,0.00008822043,0.000048201735,0.000063740896,0.00008864185],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005642726,0.00081730523,0.000823627,0.0029739507,0.0005032109,0.0007834284,0.0008059799,0.000857193,0.0026684604],"category_scores_gemma":[0.0018827997,0.00024516627,0.0010410844,0.0020468603,0.00027379527,0.00048344702,0.0005475325,0.00043906833,0.00083045685],"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":[0.010295708,0.0021482178,0.272319,0.005503327,0.0022716979,0.005163497,0.00042365346,0.39876127,0.06568046,0.0048948186,0.065500505,0.16703781],"study_design_scores_gemma":[0.0006867567,0.0005844517,0.080885,0.0001286102,0.00044249644,0.0022160828,0.00031341572,0.88758713,0.012146503,0.0040298277,0.010930919,0.000048846505],"about_ca_topic_score_codex":0.0031450223,"about_ca_topic_score_gemma":0.004820281,"teacher_disagreement_score":0.0031450223,"about_ca_system_score_codex":0.0006180203,"about_ca_system_score_gemma":0.00078313827,"threshold_uncertainty_score":0.008926928},"labels":[],"label_agreement":null},{"id":"W2120815702","doi":"10.1093/bioinformatics/bti1040","title":"ExonHunter: a comprehensive approach to gene finding","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":52,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Annotation; Hidden Markov model; Gene prediction; Probabilistic logic; Set (abstract data type); Sequence (biology); Markov chain; Extension (predicate logic); Data mining; Artificial intelligence; Machine learning; Gene; Genome; Genetics; Biology","score_opus":0.02161260978141782,"score_gpt":0.26534300760899315,"score_spread":0.24373039782757533,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2120815702","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":[0.0058747274,0.0006467616,0.930936,0.00042742016,0.00005867263,0.00012839823,0.0032015236,0.05733917,0.001387335],"genre_scores_gemma":[0.022794204,0.0003913733,0.9659255,0.00025841763,0.000055282955,0.0001508408,0.0057971887,0.0026102592,0.0020168768],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99799913,0.00025939252,0.00014263725,0.0008147415,0.00070760323,0.00007643725],"domain_scores_gemma":[0.99632984,0.0019852086,0.00027907788,0.00077113067,0.00038869592,0.0002460281],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029981993,0.0012892595,0.0016035524,0.0028260523,0.0011813112,0.0015935936,0.0035258182,0.0011711987,0.0067255185],"category_scores_gemma":[0.00621643,0.0010449308,0.001059846,0.0026202756,0.001106519,0.0032463863,0.002622889,0.0018723595,0.0038775583],"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":[0.00073455746,0.0002254607,0.006559113,0.0011409639,0.00032790515,0.00081581384,0.0005198571,0.03307757,0.07208929,0.022052871,0.09101016,0.7714464],"study_design_scores_gemma":[0.0002482864,0.00023323882,0.004276796,0.00021091555,0.00021727066,0.0032606255,0.00018151721,0.5741929,0.16180986,0.08773313,0.16733614,0.00029920338],"about_ca_topic_score_codex":0.0017815654,"about_ca_topic_score_gemma":0.0035592539,"teacher_disagreement_score":0.0067255185,"about_ca_system_score_codex":0.0006077321,"about_ca_system_score_gemma":0.002019184,"threshold_uncertainty_score":0.022499084},"labels":[],"label_agreement":null},{"id":"W2121105117","doi":"10.1093/bioinformatics/bti406","title":"Modularized learning of genetic interaction networks from biological annotations and mRNA expression data","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":59,"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":"","keywords":"Expression (computer science); Computer science; Messenger RNA; Computational biology; Annotation; Artificial intelligence; Biology; Genetics; Gene; Programming language","score_opus":0.02539486015560128,"score_gpt":0.27828838538913875,"score_spread":0.2528935252335375,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2121105117","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":[0.052362926,0.0001219978,0.94662976,0.00008536001,0.000006349697,0.000026265301,0.00013410457,0.0004029857,0.00023032217],"genre_scores_gemma":[0.5608545,0.0001961318,0.43657768,0.00008297432,0.000045963,0.00012226679,0.0012203935,0.000069357986,0.00083072344],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9993161,0.00021648842,0.000031211828,0.0002716886,0.00011097216,0.000053552714],"domain_scores_gemma":[0.9968741,0.0020268806,0.0004130265,0.00027367158,0.00032210955,0.000090118825],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001662868,0.000825772,0.00081278244,0.0014179803,0.0003546461,0.0006432299,0.0012517327,0.00081508246,0.0009151759],"category_scores_gemma":[0.006562518,0.00041881725,0.0008341014,0.0010923289,0.00071724807,0.0011895167,0.0009953794,0.00094442954,0.00021974565],"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":[0.00028676176,0.00013403248,0.007601779,0.00017380029,0.00017295768,0.00012763175,0.000135201,0.7337431,0.014252833,0.00573721,0.00072956056,0.23690517],"study_design_scores_gemma":[0.000009891095,0.00002859573,0.0011601567,0.000005483405,0.00002058444,0.000029481827,0.000008574261,0.98677546,0.002493604,0.009260777,0.00019908122,0.000008266116],"about_ca_topic_score_codex":0.002544547,"about_ca_topic_score_gemma":0.003912961,"teacher_disagreement_score":0.002544547,"about_ca_system_score_codex":0.0008293533,"about_ca_system_score_gemma":0.0009196151,"threshold_uncertainty_score":0.0087941885},"labels":[],"label_agreement":null},{"id":"W2121194346","doi":"10.1093/bioinformatics/bti427","title":"NodMutDB: a database for genes and mutants involved in symbiosis","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Legume Nitrogen Fixing Symbiosis","field":"Agricultural and Biological Sciences","cited_by":22,"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":"Functional genomics; Mutant; Gene; Computational biology; Function (biology); Biology; Schema (genetic algorithms); Symbiosis; Genomics; Database; Computer science; Genetics; Genome; Information retrieval","score_opus":0.028458525071656963,"score_gpt":0.25355197669410756,"score_spread":0.2250934516224506,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2121194346","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03212758,0.0072659245,0.052989535,0.00083116884,0.00022063218,0.00032083367,0.8492156,0.04135675,0.015671933],"genre_scores_gemma":[0.028723344,0.002194116,0.04164556,0.00025771942,0.000040359297,0.0002473906,0.92295307,0.0017739692,0.0021645413],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9993523,0.00007775765,0.00019009893,0.00014936154,0.0001830372,0.00004748124],"domain_scores_gemma":[0.9989436,0.00024241534,0.00026907653,0.0001764085,0.00018714549,0.00018118777],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009787346,0.0013228984,0.0011948934,0.0035190901,0.00056362135,0.0012798321,0.0027294985,0.0009768128,0.015629176],"category_scores_gemma":[0.0022379074,0.0004092207,0.0006824912,0.00467804,0.00027572172,0.0014099315,0.0017296828,0.00069667905,0.011142691],"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":[0.0036258262,0.00043035686,0.02176936,0.013091009,0.0005617455,0.0024437197,0.0008215094,0.0049661743,0.13921021,0.019648166,0.50523174,0.2882002],"study_design_scores_gemma":[0.00039600706,0.00023617093,0.019078605,0.00071503705,0.00037831138,0.0024036916,0.00026346848,0.0060499497,0.038808253,0.007753486,0.92378604,0.0001310205],"about_ca_topic_score_codex":0.0023570398,"about_ca_topic_score_gemma":0.0025567487,"teacher_disagreement_score":0.015629176,"about_ca_system_score_codex":0.0004945339,"about_ca_system_score_gemma":0.0014334448,"threshold_uncertainty_score":0.052284777},"labels":[],"label_agreement":null},{"id":"W2127325419","doi":"10.1093/bioinformatics/bti439","title":"pFind: a novel database-searching software system for automated peptide and protein identification via tandem mass spectrometry","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Advanced Proteomics Techniques and Applications","field":"Chemistry","cited_by":181,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"","keywords":"Computer science; Usability; Software; XML; Database; Tandem mass spectrometry; Proteomics; Identification (biology); Database search engine; Tandem mass tag; Data mining; Mass spectrometry; Information retrieval; Quantitative proteomics; Programming language; Operating system; Chemistry; Search engine; Chromatography; Biology","score_opus":0.01994045347924747,"score_gpt":0.2977135505692231,"score_spread":0.27777309708997566,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2127325419","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":[0.016407717,0.0015704869,0.64110506,0.00027508417,0.00020431634,0.00040474546,0.00975425,0.32551742,0.0047609103],"genre_scores_gemma":[0.07653266,0.0012616748,0.8487932,0.0005388641,0.00014010843,0.0010927125,0.0519185,0.0113493735,0.008372771],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99925274,0.00007325456,0.00009726611,0.00021552795,0.0002895006,0.000071647475],"domain_scores_gemma":[0.9993037,0.00022086693,0.00010987475,0.0001061236,0.00016568741,0.000093724964],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019186465,0.0021233894,0.0015887076,0.0029056973,0.0007903682,0.0012628686,0.003590851,0.00092269655,0.012388707],"category_scores_gemma":[0.0023363181,0.0008154958,0.00077152037,0.0017451306,0.00046867906,0.002616434,0.0017921431,0.0015843532,0.008506986],"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":[0.0026689703,0.00039041982,0.0048109805,0.0016370005,0.00041365428,0.0010828098,0.00028950657,0.0054268567,0.14159025,0.0063037793,0.1917128,0.643673],"study_design_scores_gemma":[0.0015840429,0.0009403478,0.011397525,0.00027652635,0.0004490405,0.00833821,0.00012405685,0.22147642,0.36621743,0.021531582,0.3670295,0.00063536735],"about_ca_topic_score_codex":0.00091663765,"about_ca_topic_score_gemma":0.0006111868,"teacher_disagreement_score":0.012388707,"about_ca_system_score_codex":0.000649458,"about_ca_system_score_gemma":0.0010920251,"threshold_uncertainty_score":0.04144436},"labels":[],"label_agreement":null},{"id":"W2144014661","doi":"10.1093/bioinformatics/bti273","title":"Online Predicted Human Interaction Database","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":620,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research","funders":"","keywords":"Drosophila melanogaster; Gene ontology; Computer science; Ontology; Visualization; Computational biology; Artificial intelligence; Biology; Gene; Gene expression; Genetics","score_opus":0.017661479102434197,"score_gpt":0.2957764760332367,"score_spread":0.2781149969308025,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2144014661","genre_codex":"dataset","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.003122154,0.0018869409,0.005088547,0.00031501378,0.000059712707,0.00010861044,0.9754524,0.004020611,0.009945975],"genre_scores_gemma":[0.007028454,0.000765364,0.007972166,0.00016131284,0.000021901515,0.00017309892,0.982249,0.00022150988,0.0014072085],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99924505,0.00015292126,0.00014043757,0.00023164289,0.00015869756,0.000071343784],"domain_scores_gemma":[0.9981639,0.0007539121,0.00022181541,0.0003063199,0.0003064204,0.00024771906],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011203191,0.0018415861,0.0017333993,0.0053563076,0.0011703026,0.0017359961,0.002217209,0.0017789106,0.0909721],"category_scores_gemma":[0.0051218187,0.00061341317,0.0008651608,0.0063646347,0.00026937795,0.0014797878,0.00201953,0.0010169143,0.051874798],"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":[0.001680178,0.00016953876,0.0091334805,0.006037294,0.0003286838,0.0011712709,0.00022339626,0.0023166894,0.0046317424,0.008627586,0.9025046,0.063175544],"study_design_scores_gemma":[0.0004206976,0.00013046063,0.014618329,0.0007157409,0.00030835846,0.0019094342,0.00017766572,0.0047135376,0.0038271137,0.0117261885,0.96135867,0.00009387797],"about_ca_topic_score_codex":0.004263578,"about_ca_topic_score_gemma":0.0061050276,"teacher_disagreement_score":0.0909721,"about_ca_system_score_codex":0.0008108322,"about_ca_system_score_gemma":0.0027788247,"threshold_uncertainty_score":0.30433196},"labels":[],"label_agreement":null},{"id":"W2156336405","doi":"10.1093/bioinformatics/bti1041","title":"Multi-way clustering of microarray data using probabilistic sparse matrix factorization","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":97,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; University of Toronto; Microsoft","keywords":"Cluster analysis; Probabilistic logic; Computer science; Data mining; Sparse matrix; Non-negative matrix factorization; Noise (video); Hierarchical clustering; Matrix decomposition; Generative model; Matrix (chemical analysis); Algorithm; Artificial intelligence; Generative grammar; Eigenvalues and eigenvectors","score_opus":0.07690284190076663,"score_gpt":0.3453961809586526,"score_spread":0.26849333905788597,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2156336405","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":[0.0032633604,0.00007917463,0.9961971,0.0001317537,0.000007897762,0.000021254358,0.00003586483,0.00016371183,0.000099891586],"genre_scores_gemma":[0.20006748,0.00036397614,0.79741895,0.00025075424,0.00015014388,0.00024689492,0.0006729265,0.00009950744,0.00072938023],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99668604,0.0011582501,0.00014186311,0.0009299441,0.00094420987,0.00013970662],"domain_scores_gemma":[0.99417275,0.003508756,0.00068463077,0.0008046814,0.0006703891,0.0001587296],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038103668,0.0011096436,0.0016740958,0.0015423953,0.00095208286,0.0010497791,0.0024570907,0.0019419952,0.00091412576],"category_scores_gemma":[0.0100388685,0.00071299105,0.0018393305,0.0023560028,0.0019517462,0.0018057456,0.0014450385,0.0019474364,0.00072368386],"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":[0.00015191597,0.00011666021,0.0023021104,0.0003539464,0.00018071548,0.00017631268,0.00039164306,0.76343733,0.016164636,0.02891185,0.0031488335,0.18466406],"study_design_scores_gemma":[0.000014890252,0.00002938566,0.0004404435,0.000011298926,0.000015603615,0.00012708845,0.000021078333,0.96630806,0.002931476,0.029168308,0.0009042711,0.000028135273],"about_ca_topic_score_codex":0.0029785282,"about_ca_topic_score_gemma":0.0028888967,"teacher_disagreement_score":0.0038103668,"about_ca_system_score_codex":0.0011033413,"about_ca_system_score_gemma":0.0012118258,"threshold_uncertainty_score":0.020151377},"labels":[],"label_agreement":null},{"id":"W2159605359","doi":"10.1093/bioinformatics/bti483","title":"Ensemble dependence model for classification and prediction of cancer and normal gene expression data","year":2005,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":39,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Ensemble forecasting; Ensemble learning; Computational biology; Cancer; Expression (computer science); Gene; Gene expression; Artificial intelligence; Genetics; Biology; Programming language","score_opus":0.06174760510571803,"score_gpt":0.32438642513740557,"score_spread":0.26263882003168754,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2159605359","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":[0.1882049,0.00045990435,0.8084621,0.00045892972,0.00003945182,0.00006384546,0.00061258365,0.0007954108,0.0009029428],"genre_scores_gemma":[0.9258962,0.00025699325,0.07031258,0.000101243015,0.000064051324,0.00014657553,0.0016123162,0.00004662847,0.0015633552],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988341,0.00052764814,0.0000640526,0.0002492059,0.0002321136,0.00009289738],"domain_scores_gemma":[0.99657047,0.0022797922,0.00024800387,0.00031334907,0.0004681076,0.00012022881],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003008715,0.0008189053,0.0011293293,0.0015813434,0.00040719507,0.00060518854,0.0011346408,0.00094011583,0.0010786328],"category_scores_gemma":[0.0066742366,0.00025927936,0.0008047218,0.0010027536,0.00045832997,0.001099089,0.00059536565,0.001193455,0.000401629],"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":[0.0004997808,0.00026871852,0.02395286,0.00006672247,0.00021191579,0.000146643,0.00012410466,0.8314516,0.0035918683,0.0058901077,0.00193843,0.1318573],"study_design_scores_gemma":[0.0000023870384,0.000015866892,0.0007648977,0.0000017520916,0.000006016325,0.000013402447,0.0000025182476,0.9972976,0.00024349197,0.0015612225,0.00008751035,0.0000033634592],"about_ca_topic_score_codex":0.0059966454,"about_ca_topic_score_gemma":0.004367492,"teacher_disagreement_score":0.0059966454,"about_ca_system_score_codex":0.00086804637,"about_ca_system_score_gemma":0.00063482026,"threshold_uncertainty_score":0.015911758},"labels":[],"label_agreement":null},{"id":"W2403880286","doi":"10.1093/bioinformatics/bts490","title":"ECCB 2012: The 11th European Conference on Computational Biology","year":2012,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Genetics, Bioinformatics, and Biomedical Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"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":"University at Buffalo; Swiss Institute of Bioinformatics; Schweizerische Akademie der Medizinischen Wissenschaften; Université de Genève; Eidgenössische Technische Hochschule Zürich; Universität Zürich; Branco Weiss Fellowship – Society in Science; Universität Basel; European Bioinformatics Institute; Universitätsspital Zürich; Microsoft Research","keywords":"Computer science; Computational biology; Biology","score_opus":0.04298584498611423,"score_gpt":0.32671621241327764,"score_spread":0.2837303674271634,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2403880286","genre_codex":"other","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0030855497,0.116039045,0.20339201,0.076954864,0.25545096,0.0016763296,0.015329485,0.017062375,0.31100944],"genre_scores_gemma":[0.01429089,0.09912918,0.117507406,0.018036122,0.054154605,0.0027846566,0.061145708,0.016851218,0.6161002],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.98809785,0.0033020107,0.00092559634,0.0018321531,0.0048688403,0.00097345945],"domain_scores_gemma":[0.9826499,0.0047881347,0.0005694231,0.0033736029,0.0054956633,0.0031232983],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011946,0.003083994,0.0040788236,0.004288096,0.0022822493,0.014376272,0.0057923,0.00635271,0.24217132],"category_scores_gemma":[0.022703141,0.0013613371,0.002651358,0.004700114,0.0025127288,0.011008321,0.008643459,0.00935601,0.18814486],"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":[0.000059378646,0.000041874788,0.000084068495,0.00031644263,0.000029315279,0.000045786044,0.000047903028,0.00048612332,0.00036259886,0.009545935,0.92805713,0.060923427],"study_design_scores_gemma":[0.000016068307,0.000015975085,0.000183935,0.00024123651,0.000007992043,0.00007036463,0.000024649691,0.0007256647,0.00013810967,0.004648162,0.99390775,0.000020155767],"about_ca_topic_score_codex":0.0023710579,"about_ca_topic_score_gemma":0.0020628632,"teacher_disagreement_score":0.24217132,"about_ca_system_score_codex":0.0027181583,"about_ca_system_score_gemma":0.00678061,"threshold_uncertainty_score":0.81014377},"labels":[],"label_agreement":null},{"id":"W2509574344","doi":"10.1093/bioinformatics/btw481","title":"ECCB 2016: The 15th European Conference on Computational Biology","year":2016,"lang":"en","type":"editorial","venue":"Computer applications in the biosciences","topic":"Cancer Genomics and Diagnostics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Cancer Research UK Cambridge Institute, University of Cambridge; RIKEN; Swiss Institute of Bioinformatics; Università di Bologna; Rigshospitalet; University of Oxford; Vrije Universiteit Amsterdam; Wageningen University and Research; Aalto-Yliopisto; Universiteit Utrecht; Univerzita Karlova v Praze; Centre National de la Recherche Scientifique; Cancer Research UK; University College London; Institut National de la Recherche Agronomique; University of Toronto; Technische Universität Dresden; Eidgenössische Technische Hochschule Zürich; Universiteit Leiden; Okinawa Institute of Science and Technology Graduate University; Universität des Saarlandes; University of Notre Dame; European Bioinformatics Institute; University of California, Santa Cruz; Vrije Universiteit Brussel; Massachusetts General Hospital","keywords":"Computational biology; Computer science; Biology; Data science","score_opus":0.013841547144270381,"score_gpt":0.29380388828160175,"score_spread":0.27996234113733137,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2509574344","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000042337826,0.01593948,0.0010058373,0.041477166,0.9380383,0.000022322469,0.00011157318,0.00016582997,0.0031972064],"genre_scores_gemma":[0.00095057907,0.0164926,0.0010580766,0.018600553,0.92602146,0.00006973663,0.0002904914,0.00042533551,0.0360912],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99272716,0.0012625693,0.0006481367,0.0009341905,0.003979527,0.00044843304],"domain_scores_gemma":[0.95506173,0.01609253,0.0022370107,0.0018058021,0.018120527,0.006682523],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.013418012,0.0030770954,0.0049364204,0.0052472623,0.0027185124,0.014181489,0.0038892878,0.011193173,0.029260693],"category_scores_gemma":[0.0319596,0.0011261796,0.0021814012,0.002848761,0.0038093028,0.007912755,0.0038366402,0.021859348,0.027565388],"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":[0.000028712391,0.000007429299,0.000010898098,0.00013756,0.000013133523,0.000020125579,0.000007070047,0.000033290282,0.00004300837,0.00071499596,0.9911413,0.007842441],"study_design_scores_gemma":[0.000032489905,0.000011217893,0.000103038765,0.0003322508,0.000017780061,0.00006620764,0.000014958525,0.0001972887,0.00008253854,0.0028892495,0.99623764,0.00001534416],"about_ca_topic_score_codex":0.0019329515,"about_ca_topic_score_gemma":0.0051249177,"teacher_disagreement_score":0.029260693,"about_ca_system_score_codex":0.0038351314,"about_ca_system_score_gemma":0.0046769255,"threshold_uncertainty_score":0.0978868},"labels":[],"label_agreement":null},{"id":"W2725970836","doi":"10.1093/bioinformatics/btx460","title":"Analysis and prediction of protein folding energy changes upon mutation by element specific persistent homology","year":2017,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":108,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Mutation; Homology modeling; Pearson product-moment correlation coefficient; Stability (learning theory); Mutation rate; Computational biology; Correlation coefficient; Computer science; Correlation; Persistent homology; Mutagenesis; Protein folding; Folding (DSP implementation); Mathematics; Protein design; Protein structure; Algorithm; Biology; Genetics; Machine learning; Statistics; Gene; Biochemistry; Engineering; Geometry","score_opus":0.010582423357981768,"score_gpt":0.24057174214454452,"score_spread":0.22998931878656276,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2725970836","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":[0.9058331,0.00022112849,0.09145682,0.00006791159,0.000008999788,0.000036410856,0.0007278631,0.0009791937,0.00066848646],"genre_scores_gemma":[0.974974,0.000087178654,0.023162385,0.000010838931,0.0000049056257,0.00003237775,0.0014129907,0.000053141626,0.00026219265],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9997485,0.000057482157,0.000017249466,0.000066835935,0.000087979606,0.000021928781],"domain_scores_gemma":[0.9989812,0.00042197338,0.0002357368,0.00011772017,0.00018323604,0.000060064718],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00071089185,0.00043678758,0.00040858105,0.0013844708,0.0002191813,0.0003161821,0.00046076992,0.00039179163,0.00088546146],"category_scores_gemma":[0.0022305388,0.000116892814,0.00046778555,0.0007912529,0.00023764002,0.00040429799,0.0004893314,0.00035676447,0.00044987345],"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":[0.00068988814,0.00030151763,0.09292853,0.00034538354,0.00021566697,0.00046705172,0.0001507026,0.71429116,0.07938842,0.0029648924,0.0017675054,0.10648928],"study_design_scores_gemma":[0.0000076281585,0.00009188342,0.013068636,0.000005444595,0.00001549371,0.000098460485,0.00001770001,0.9748497,0.01082864,0.00077192474,0.00023220194,0.000012368317],"about_ca_topic_score_codex":0.0011821251,"about_ca_topic_score_gemma":0.0010013841,"teacher_disagreement_score":0.0013844708,"about_ca_system_score_codex":0.00031454707,"about_ca_system_score_gemma":0.0003893664,"threshold_uncertainty_score":0.003759563},"labels":[],"label_agreement":null},{"id":"W3108440423","doi":"10.1093/bioinformatics/btaa976","title":"Deep feature extraction of single-cell transcriptomes by generative adversarial network","year":2020,"lang":"en","type":"article","venue":"Computer applications in the biosciences","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":39,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Douglas Mental Health University Institute; McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Computer science; Generative model; Generative grammar; Feature (linguistics); Source code; Embedding; Generative adversarial network; Raw data; Code (set theory); Artificial intelligence; Data mining; Machine learning; Deep learning","score_opus":0.016535428181128544,"score_gpt":0.23602699513096556,"score_spread":0.219491566949837,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3108440423","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":[0.028336102,0.00042049726,0.96626943,0.00036180264,0.00006390104,0.00005195777,0.0012428547,0.0020289165,0.0012245731],"genre_scores_gemma":[0.70560855,0.00078576396,0.2760129,0.00076676894,0.00011731413,0.00046637846,0.007505556,0.00048032647,0.008256451],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99973506,0.000067682806,0.000009194689,0.000098358505,0.000049518883,0.00004015014],"domain_scores_gemma":[0.9993722,0.0004081867,0.000045808112,0.000081637954,0.000064334134,0.00002780576],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008297785,0.0009821946,0.0006552038,0.00037380273,0.00022893085,0.0006030512,0.0010906117,0.00085011485,0.0020242494],"category_scores_gemma":[0.0018297536,0.00040473836,0.0008230912,0.00042204597,0.0008088108,0.0005510873,0.0010996002,0.0017309124,0.0010442074],"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":[0.00015104952,0.00005714781,0.0029882106,0.00014459093,0.000090150475,0.00022946554,0.00011047668,0.87793404,0.028079294,0.008462374,0.006479615,0.075273514],"study_design_scores_gemma":[0.0000033787182,0.000012634701,0.00041966647,0.000007355285,0.0000064265078,0.000034579767,0.000007180078,0.989151,0.003982742,0.00558939,0.000776933,0.000008668379],"about_ca_topic_score_codex":0.0024160324,"about_ca_topic_score_gemma":0.0033490972,"teacher_disagreement_score":0.0024160324,"about_ca_system_score_codex":0.0006332793,"about_ca_system_score_gemma":0.0005921165,"threshold_uncertainty_score":0.0067718625},"labels":[],"label_agreement":null}]}