{"meta":{"query_hash":"92f54b72ab7b","filters":{"venue":"Molecular Informatics"},"cohort_total":26,"direct_labels_cover":0,"predictions_cover":26,"exported":26,"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/92f54b72ab7b","api":"https://metacan.xera.ac/api/v1/cohort?venue=Molecular+Informatics"},"results":[{"id":"W1899644141","doi":"10.1002/minf.201400118","title":"Towards Better BBB Passage Prediction Using an Extensive and Curated Data Set","year":2015,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":50,"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":"Linear discriminant analysis; Set (abstract data type); Test set; Computer science; Data set; Artificial intelligence; Machine learning; Regression; Linear regression; Data mining; Variance (accounting); Drug discovery; Linear model; Regression analysis; Statistics; Mathematics; Bioinformatics; Biology","score_opus":0.1242311728463903,"score_gpt":0.35289098644224953,"score_spread":0.22865981359585924,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1899644141","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":[0.50233,0.0041374,0.45728442,0.0010702689,0.0002092698,0.00059399195,0.026302002,0.004894157,0.0031784885],"genre_scores_gemma":[0.6890848,0.0023298333,0.2540661,0.00031261242,0.00008475394,0.00084520056,0.05112789,0.00032625496,0.001822436],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99915445,0.00025985928,0.00008372868,0.000244606,0.00020240604,0.000054938995],"domain_scores_gemma":[0.9965064,0.0019541213,0.00032940874,0.00055156,0.0005653698,0.00009317879],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019328136,0.0015887159,0.0015625836,0.002191075,0.0004559981,0.0016739842,0.0012516058,0.0013939616,0.0012659162],"category_scores_gemma":[0.0064214454,0.00035899284,0.0016572358,0.0018857375,0.00050689146,0.001274628,0.0010182792,0.0016640392,0.00077568326],"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":[0.0008530105,0.0014806782,0.023257138,0.0025652,0.0006024166,0.001225905,0.0002426858,0.7050898,0.0803821,0.0033141477,0.010919164,0.17006773],"study_design_scores_gemma":[0.000054147036,0.0006672733,0.010232842,0.00011453748,0.00020462666,0.0004492345,0.00013640906,0.9257069,0.045019966,0.0034787247,0.013811636,0.00012367753],"about_ca_topic_score_codex":0.0053244894,"about_ca_topic_score_gemma":0.0059741484,"teacher_disagreement_score":0.0053244894,"about_ca_system_score_codex":0.0007916277,"about_ca_system_score_gemma":0.0017987501,"threshold_uncertainty_score":0.010586977},"labels":[],"label_agreement":null},{"id":"W1980840780","doi":"10.1002/minf.201100111","title":"An Advanced Group Contribution Method for High‐Dimensional, Sparse Data Sets","year":2011,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Group (periodic table); Computer science; Data mining; Chemistry; Organic chemistry","score_opus":0.04836593607333022,"score_gpt":0.3458207124614303,"score_spread":0.2974547763881001,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1980840780","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":[0.0014938727,0.00015965819,0.9977635,0.00007900609,0.00003425356,0.00002796689,0.000040008035,0.00021906239,0.00018276487],"genre_scores_gemma":[0.06209595,0.00048097514,0.9337327,0.0001514316,0.00017968597,0.00036961562,0.0005261255,0.00018590823,0.0022776094],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99822515,0.0008080755,0.00008386744,0.00025752725,0.0005614022,0.00006386909],"domain_scores_gemma":[0.99689436,0.0020286643,0.00018394573,0.00031348466,0.00049212016,0.00008735859],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038592531,0.0011995342,0.0016810354,0.002230957,0.0007426165,0.0010777428,0.00226411,0.0017328655,0.0031295354],"category_scores_gemma":[0.009170289,0.0005684369,0.0018092506,0.0024154834,0.00090365723,0.0014755285,0.0019992439,0.0023234664,0.0015210117],"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":[0.0002451023,0.00014234034,0.0013319922,0.0003376058,0.00028919178,0.00024206826,0.00020037785,0.41630235,0.008020769,0.028875465,0.006691666,0.537321],"study_design_scores_gemma":[0.000011155485,0.000025113035,0.000109143286,0.000008882663,0.000009897304,0.000040576502,0.00000718037,0.98993874,0.0005573293,0.0073173875,0.0019651684,0.000009478088],"about_ca_topic_score_codex":0.003034087,"about_ca_topic_score_gemma":0.00262175,"teacher_disagreement_score":0.0038592531,"about_ca_system_score_codex":0.00049550715,"about_ca_system_score_gemma":0.001550864,"threshold_uncertainty_score":0.020409942},"labels":[],"label_agreement":null},{"id":"W1983545523","doi":"10.1002/minf.201100085","title":"Identification of Novel Inhibitors of UDP‐Galactopyranose Mutase by Structure‐Based Virtual Screening","year":2011,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Carbohydrate Chemistry and Synthesis","field":"Chemistry","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":"University of Alberta; University of Saskatchewan","funders":"University of Saskatchewan","keywords":"Mutase; In silico; Escherichia coli; Virtual screening; Enzyme; Docking (animal); Biochemistry; Klebsiella pneumoniae; Dehydratase; Chemistry; Mycobacterium tuberculosis; Biology; Klebsiella pneumonia; Stereochemistry; Drug discovery; Tuberculosis","score_opus":0.013410427002871003,"score_gpt":0.21194755187406325,"score_spread":0.19853712487119224,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1983545523","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":[0.94828665,0.01785818,0.016566034,0.0007752192,0.00014637568,0.0013195589,0.0036354742,0.001342035,0.010070511],"genre_scores_gemma":[0.9643947,0.010990695,0.014242731,0.0003445563,0.000050709346,0.0004867858,0.0055276807,0.000068151756,0.0038940061],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99977094,0.0000500426,0.000014163748,0.000042625856,0.00006616827,0.000056069435],"domain_scores_gemma":[0.9998373,0.00005108792,0.000035810852,0.000009734654,0.000027834145,0.000038184564],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00043377912,0.0013575411,0.0018480222,0.0011485163,0.00035443483,0.0010678123,0.0008847058,0.0005621575,0.0025931022],"category_scores_gemma":[0.0004942684,0.00047480222,0.00081895175,0.0008378976,0.00028093214,0.00031819122,0.00043802554,0.0008207443,0.00052907085],"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":[0.005563589,0.005021658,0.008451571,0.001975129,0.0012125588,0.002304821,0.00022748462,0.13865323,0.61796975,0.0024446452,0.0063015893,0.20987396],"study_design_scores_gemma":[0.011415453,0.032035243,0.018882196,0.00039297785,0.0034190766,0.005738964,0.00032467968,0.25706086,0.6152945,0.002247451,0.052828003,0.0003606742],"about_ca_topic_score_codex":0.001046978,"about_ca_topic_score_gemma":0.0027496298,"teacher_disagreement_score":0.0025931022,"about_ca_system_score_codex":0.00079780834,"about_ca_system_score_gemma":0.00066690345,"threshold_uncertainty_score":0.0086748},"labels":[],"label_agreement":null},{"id":"W1990193701","doi":"10.1002/minf.201400017","title":"Systematic Identification of Matching Molecular Series and Mapping of Screening Hits","year":2014,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Emergent BioSolutions (Canada)","funders":"","keywords":"Computer science; Series (stratigraphy); Matching (statistics); Identification (biology); Computational biology; Data mining; Information retrieval; Biology; Mathematics","score_opus":0.009911711928669237,"score_gpt":0.2481507745470005,"score_spread":0.23823906261833125,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1990193701","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":[0.6845781,0.008532493,0.19658485,0.0004751412,0.0001964356,0.0027704872,0.07668968,0.012382784,0.017789977],"genre_scores_gemma":[0.5657536,0.0028084728,0.31364337,0.0002485695,0.00008606274,0.0016184379,0.11255191,0.00063667,0.002652916],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9981668,0.0003661833,0.00029111936,0.00051287655,0.0005505641,0.00011239376],"domain_scores_gemma":[0.99719536,0.0012241555,0.0004912436,0.0005229366,0.0004627084,0.00010364285],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015387426,0.0010707505,0.0018595567,0.0074805417,0.00046728193,0.0011087846,0.0014149459,0.0005003473,0.004595159],"category_scores_gemma":[0.0076423693,0.0005034555,0.0013884922,0.007024539,0.0003296863,0.0013037571,0.0013336766,0.00057837425,0.0016320669],"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":[0.0023261684,0.0009409596,0.05027186,0.0057377946,0.0007763014,0.0018283178,0.00030475622,0.05588477,0.24903864,0.012356252,0.017899761,0.6026345],"study_design_scores_gemma":[0.0008899065,0.0047246045,0.06503473,0.00041866212,0.0025016908,0.0044235843,0.00063741475,0.42226446,0.37601694,0.023516936,0.09931763,0.00025335076],"about_ca_topic_score_codex":0.0007461925,"about_ca_topic_score_gemma":0.0013773147,"teacher_disagreement_score":0.0074805417,"about_ca_system_score_codex":0.00060622476,"about_ca_system_score_gemma":0.0012956021,"threshold_uncertainty_score":0.015372336},"labels":[],"label_agreement":null},{"id":"W2003032055","doi":"10.1002/minf.201100053","title":"Target‐Driven Subspace Mapping Methods and Their Applicability Domain Estimation","year":2011,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":22,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"Natural Sciences and Engineering Research Council of Canada; Consejo Nacional de Investigaciones Científicas y Técnicas; Deutsche Forschungsgemeinschaft","keywords":"Subspace topology; Estimation; Computer science; Domain (mathematical analysis); Data mining; Artificial intelligence; Computational biology; Pattern recognition (psychology); Mathematics; Biology; Engineering","score_opus":0.025776361414791102,"score_gpt":0.27292270232969645,"score_spread":0.24714634091490534,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2003032055","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":[0.009484752,0.00017877553,0.9896917,0.00007071529,0.000006348124,0.000019654819,0.000023316092,0.00015417633,0.00037051985],"genre_scores_gemma":[0.47482446,0.0005910646,0.52221084,0.00011811466,0.000066279696,0.0003197635,0.0003083902,0.00013852608,0.0014225875],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99878126,0.00062301575,0.00004527209,0.00013351739,0.00037286093,0.000044037384],"domain_scores_gemma":[0.99535817,0.003350549,0.00031859992,0.00045076915,0.00043117008,0.00009071951],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029278954,0.0005111256,0.00069322117,0.0011338309,0.00042380352,0.0007052291,0.0010050979,0.0006415411,0.0010048442],"category_scores_gemma":[0.009585235,0.0002960124,0.00077036733,0.0008054187,0.0009879251,0.0011483425,0.0015633738,0.0012910105,0.0003614417],"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":[0.00012617596,0.000093622155,0.0020251132,0.00015669118,0.000096994016,0.00007950255,0.00022463629,0.58683944,0.0070006736,0.03890313,0.0011252834,0.3633287],"study_design_scores_gemma":[0.000007278068,0.000035404184,0.00044030263,0.000009909327,0.0000058486276,0.000052219904,0.000020384301,0.97290015,0.0026118096,0.023190107,0.0007118754,0.000014645231],"about_ca_topic_score_codex":0.0010265495,"about_ca_topic_score_gemma":0.0006790681,"teacher_disagreement_score":0.0029278954,"about_ca_system_score_codex":0.0003923564,"about_ca_system_score_gemma":0.00068952056,"threshold_uncertainty_score":0.015484393},"labels":[],"label_agreement":null},{"id":"W2008947010","doi":"10.1002/minf.201200031","title":"Chemical Descriptors Are More Important Than Learning Algorithms for Modelling","year":2012,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":40,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"University of Waterloo; Mitacs","keywords":"Computer science; Cheminformatics; Machine learning; Artificial intelligence; Algorithm; Chemistry; Computational chemistry","score_opus":0.037541515635923485,"score_gpt":0.2965459023265248,"score_spread":0.2590043866906013,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2008947010","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":[0.064900324,0.04766675,0.8462719,0.008059924,0.0015354634,0.000264597,0.0011549874,0.0010545866,0.029091502],"genre_scores_gemma":[0.7212615,0.042723197,0.22024207,0.0018382898,0.0020180095,0.0002683568,0.001529654,0.0005371429,0.009581843],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9986563,0.00059201743,0.00007189772,0.00021540115,0.0004134856,0.00005093461],"domain_scores_gemma":[0.99033153,0.007879294,0.00066411245,0.000648398,0.00036580567,0.00011088435],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015966644,0.0010208776,0.0018238297,0.0012537469,0.0005087296,0.0028540592,0.0010378852,0.0016831141,0.0063182763],"category_scores_gemma":[0.012586691,0.00029324784,0.0004860552,0.0035864643,0.0016333557,0.0063291406,0.000833307,0.0022669004,0.0016490937],"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":[0.0005234667,0.00025489088,0.006482527,0.003403886,0.00033554368,0.00016921859,0.00009435776,0.17529283,0.008552531,0.26602983,0.007391356,0.53146964],"study_design_scores_gemma":[0.00015003856,0.0005177661,0.0026760276,0.00035952998,0.0002909586,0.00039498575,0.00015254556,0.32698876,0.012422521,0.6108592,0.045062087,0.00012554915],"about_ca_topic_score_codex":0.0011154945,"about_ca_topic_score_gemma":0.0008499589,"teacher_disagreement_score":0.0063182763,"about_ca_system_score_codex":0.0005370222,"about_ca_system_score_gemma":0.0005747029,"threshold_uncertainty_score":0.021136701},"labels":[],"label_agreement":null},{"id":"W2015042232","doi":"10.1002/minf.200900019","title":"How to Deal with Batch Effect in Sequential Microarray Experiments?","year":2010,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":13,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Lunenfeld-Tanenbaum Research Institute; Public Health Ontario; University of Toronto; Mount Sinai Hospital","funders":"National Cancer Institute; Canadian Prostate Cancer Research Initiative; Institute for Prostate Cancer Research","keywords":"Computer science; Computational biology; Microarray analysis techniques; Data mining; Chemistry; Biology; Gene; Gene expression; Biochemistry","score_opus":0.005950242800431988,"score_gpt":0.2494548033599986,"score_spread":0.2435045605595666,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2015042232","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005559368,0.000509897,0.98829496,0.0007277981,0.00045640883,0.00010656988,0.0001380617,0.0038927516,0.0003140833],"genre_scores_gemma":[0.05937986,0.00061770284,0.93232965,0.0012200058,0.0006648583,0.000706787,0.0005356469,0.0013648447,0.0031806882],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.98814,0.0048262645,0.0008021305,0.0026443512,0.0030265106,0.0005606985],"domain_scores_gemma":[0.95550185,0.027012372,0.002071249,0.011004546,0.003602793,0.0008071343],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.026451796,0.0029200437,0.0047167013,0.0011680174,0.0013944866,0.0022546914,0.004223462,0.0031688607,0.0037741512],"category_scores_gemma":[0.050400205,0.002669143,0.0026131596,0.0025069986,0.0021353725,0.0052762986,0.0022722376,0.0056124954,0.0025347315],"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":[0.0023929565,0.00072339765,0.012475255,0.0015615899,0.0016945015,0.00060207583,0.0006853009,0.07587544,0.16375002,0.02046683,0.025801778,0.6939709],"study_design_scores_gemma":[0.00039980342,0.0006685478,0.012705309,0.0001180916,0.0009851569,0.0011676092,0.00022128347,0.7078483,0.1521578,0.09560769,0.02771123,0.0004091313],"about_ca_topic_score_codex":0.0036321383,"about_ca_topic_score_gemma":0.00818015,"teacher_disagreement_score":0.9735482,"about_ca_system_score_codex":0.0017936664,"about_ca_system_score_gemma":0.0037989914,"threshold_uncertainty_score":0.13989216},"labels":[],"label_agreement":null},{"id":"W2032729355","doi":"10.1002/minf.201000018","title":"Finding Inspiration in the Protein Data Bank to Chemically Antagonize Readers of the Histone Code","year":2010,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Click Chemistry and Applications","field":"Chemistry","cited_by":22,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canada Research Chairs; University of New Brunswick; Structural Genomics Consortium; University of Toronto","funders":"Knut och Alice Wallenbergs Stiftelse; Natural Sciences and Engineering Research Council of Canada; Karolinska Institutet; Stiftelsen för Strategisk Forskning; Canadian Institutes of Health Research; Genome Canada; Ontario Genomics; Wellcome Trust; Ontario Genomics Institute; Ontario Innovation Trust","keywords":"Protein Data Bank (RCSB PDB); Protein Data Bank; Pharmacophore; Histone; Epigenetics; Chemistry; Cheminformatics; Computational biology; Affinities; Stereochemistry; Acetylation; Drug discovery; Biology; Biochemistry; Protein structure; Bioinformatics; DNA; Gene","score_opus":0.028061353100305983,"score_gpt":0.2972882026075814,"score_spread":0.2692268495072754,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2032729355","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":[0.2644494,0.010260398,0.061257046,0.0032780615,0.0004272068,0.0011673211,0.5830409,0.03963303,0.03648664],"genre_scores_gemma":[0.20624723,0.006679134,0.13441788,0.00074246293,0.000040279563,0.00066440814,0.6393319,0.0010228899,0.010853801],"study_design_codex":"not_applicable","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9997646,0.000033613982,0.000027497428,0.000053202773,0.000089742316,0.000031249492],"domain_scores_gemma":[0.99975413,0.000070318805,0.00004870451,0.000042648353,0.000047749,0.000036327252],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005727286,0.0009497667,0.0010957391,0.0011029762,0.0004084068,0.00079136295,0.00070570206,0.0006764902,0.012815982],"category_scores_gemma":[0.0012499737,0.0003369377,0.000513799,0.0028040186,0.00018793553,0.0005639733,0.00035904624,0.00069037767,0.008666652],"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":[0.0061632306,0.0012238463,0.03830825,0.006782017,0.00062813674,0.002362407,0.0003384684,0.011722908,0.33487687,0.013001254,0.3376536,0.24693894],"study_design_scores_gemma":[0.001855735,0.0012487494,0.03967021,0.00039407954,0.0007725164,0.001919489,0.00026242185,0.06488527,0.21778446,0.006281041,0.6647317,0.00019432037],"about_ca_topic_score_codex":0.0021066063,"about_ca_topic_score_gemma":0.0051994706,"teacher_disagreement_score":0.012815982,"about_ca_system_score_codex":0.0007557051,"about_ca_system_score_gemma":0.0017535149,"threshold_uncertainty_score":0.04287374},"labels":[],"label_agreement":null},{"id":"W2059197104","doi":"10.1002/minf.201200020","title":"Prospective Prediction of Antitarget Activity by Matched Molecular Pairs Analysis","year":2012,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"AstraZeneca (Canada)","funders":"","keywords":"Quantitative structure–activity relationship; Computer science; Aggregate (composite); Data mining; Machine learning; Nanotechnology; Materials science","score_opus":0.00911041723449239,"score_gpt":0.25838406969099165,"score_spread":0.24927365245649927,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2059197104","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":[0.60496813,0.0034758437,0.376753,0.00046042196,0.000055056127,0.00054815627,0.006159664,0.0030074394,0.004572258],"genre_scores_gemma":[0.9173644,0.0005000766,0.0762223,0.000103911334,0.000051508567,0.00019347618,0.004871112,0.00006930799,0.0006239195],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984567,0.0005310619,0.00011271863,0.00032236474,0.00050223625,0.000074863136],"domain_scores_gemma":[0.99670404,0.0021073204,0.0005164979,0.00027104927,0.00026557964,0.00013557886],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0026767561,0.0011910956,0.001831647,0.0035448724,0.00033906137,0.0011505126,0.0011385791,0.00088107,0.002850022],"category_scores_gemma":[0.0068001496,0.00045377208,0.00092572125,0.0020462917,0.0005292479,0.0013451484,0.0010422833,0.0011285981,0.0012204621],"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":[0.0040653213,0.0012086309,0.088742174,0.0006960774,0.00054763653,0.0005010071,0.000062943465,0.46293494,0.06395177,0.007228596,0.0053088954,0.364752],"study_design_scores_gemma":[0.00013085589,0.0010404926,0.006373829,0.000011902691,0.00011094544,0.0002085485,0.000012563298,0.96321803,0.023377191,0.0039797244,0.0015072846,0.000028654686],"about_ca_topic_score_codex":0.00049863424,"about_ca_topic_score_gemma":0.00060068036,"teacher_disagreement_score":0.0035448724,"about_ca_system_score_codex":0.000557466,"about_ca_system_score_gemma":0.00063958997,"threshold_uncertainty_score":0.014156222},"labels":[],"label_agreement":null},{"id":"W2061937956","doi":"10.1002/minf.201200141","title":"Chemically Advanced Template Search (CATS) for Scaffold‐Hopping and Prospective Target Prediction for ‘Orphan’ Molecules","year":2013,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":164,"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":"Ministry of Earth Sciences; Ministry of Education, India; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung","keywords":"Drug discovery; Scaffold; Computational biology; Nanotechnology; Chemistry; Computer science; Combinatorial chemistry; Materials science; Medicine; Biology; Biomedical engineering; Biochemistry","score_opus":0.010784107179727472,"score_gpt":0.26918250934824844,"score_spread":0.25839840216852095,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2061937956","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":[0.4515797,0.0028875247,0.5334153,0.000172785,0.00006626675,0.00048793986,0.0013165006,0.0041305786,0.0059434064],"genre_scores_gemma":[0.6798569,0.00083324726,0.3157246,0.000082387036,0.000047052436,0.0003798444,0.0018321042,0.00019204452,0.001051845],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991948,0.00027877538,0.000053230237,0.00011168443,0.00031240098,0.00004902158],"domain_scores_gemma":[0.99918646,0.00047786304,0.0001040828,0.00010052363,0.00009151305,0.000039560495],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014066771,0.0006770313,0.00085664965,0.0010624197,0.00026566104,0.00075303775,0.0009829033,0.0005401526,0.0020162459],"category_scores_gemma":[0.0025429912,0.00019146397,0.00058735907,0.0010401168,0.00039432893,0.0007936799,0.00084095984,0.0005057274,0.0003005814],"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":[0.0011144264,0.00032171444,0.0061616707,0.00033268705,0.00019639473,0.00014975501,0.000095095435,0.63210183,0.042135905,0.015001356,0.0017010134,0.30068824],"study_design_scores_gemma":[0.0000531854,0.00064102345,0.00072400464,0.000009749596,0.000032202723,0.00007100207,0.00001699079,0.9806224,0.014201424,0.0021939396,0.0014162187,0.000017846301],"about_ca_topic_score_codex":0.0012895832,"about_ca_topic_score_gemma":0.0014544824,"teacher_disagreement_score":0.0020162459,"about_ca_system_score_codex":0.00050675235,"about_ca_system_score_gemma":0.0008714446,"threshold_uncertainty_score":0.007439375},"labels":[],"label_agreement":null},{"id":"W2066256098","doi":"10.1002/minf.200900040","title":"Forecasting CYP2D6 and CYP3A4 Risk with a Global/Local Fusion Model of CYP450 Inhibition","year":2010,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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 Health Network; Ontario Institute for Cancer Research","funders":"Neurocrine Biosciences","keywords":"Quantitative structure–activity relationship; Computer science; Data mining; Pharmacophore; Similarity (geometry); Set (abstract data type); Machine learning; Artificial intelligence; Chemistry","score_opus":0.011373145632268525,"score_gpt":0.23896628912515772,"score_spread":0.2275931434928892,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2066256098","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":[0.30030164,0.00041625087,0.69274867,0.00041284776,0.000045754965,0.00012424141,0.0011916067,0.0014174094,0.003341593],"genre_scores_gemma":[0.8868462,0.00015499427,0.110213794,0.00010945208,0.00004812097,0.00009305809,0.0012609031,0.00005995735,0.0012135552],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992599,0.0001891692,0.000044887845,0.00022739217,0.00022298486,0.000055642326],"domain_scores_gemma":[0.9984206,0.00084407267,0.00026981402,0.00021924115,0.0001845268,0.000061755876],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0022360033,0.000908352,0.0010841887,0.0010444238,0.00030409824,0.0008001751,0.0008540408,0.00071701675,0.0012595418],"category_scores_gemma":[0.003451499,0.00033107688,0.001185949,0.0008373018,0.00042363373,0.0010517852,0.000885488,0.0011646854,0.00032854572],"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":[0.00010319497,0.000081770144,0.008277342,0.000021337648,0.00008761802,0.00003941804,0.000022515802,0.948422,0.0020668542,0.00083915365,0.00054846465,0.039490286],"study_design_scores_gemma":[0.0000041586636,0.000038776303,0.00094523834,0.000001405929,0.000014955072,0.000015091913,0.000005409905,0.9973564,0.0006310173,0.00081634644,0.00016614751,0.0000051074007],"about_ca_topic_score_codex":0.006002321,"about_ca_topic_score_gemma":0.0069815326,"teacher_disagreement_score":0.006002321,"about_ca_system_score_codex":0.00077882485,"about_ca_system_score_gemma":0.0007525743,"threshold_uncertainty_score":0.011934757},"labels":[],"label_agreement":null},{"id":"W2155678816","doi":"10.1002/minf.201200171","title":"Rendering Molecular Sketches for Publication Quality Output","year":2013,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec","funders":"","keywords":"Rendering (computer graphics); Computer science; Cheminformatics; Computer graphics (images); File format; Graphics; Software; Computer graphics; Computational science; Engineering drawing; Database; Programming language; Bioinformatics; Engineering","score_opus":0.03700972570440528,"score_gpt":0.3195687943401153,"score_spread":0.28255906863571,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2155678816","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":[0.004147851,0.00066000025,0.94897276,0.00084334135,0.0010027321,0.00025221606,0.0012191959,0.027090713,0.015811121],"genre_scores_gemma":[0.06774283,0.0020883223,0.9066217,0.00017511455,0.0002596589,0.00033023278,0.0018853045,0.006420733,0.014476102],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9978696,0.00041321406,0.00022688314,0.00019873583,0.0011583808,0.00013316343],"domain_scores_gemma":[0.98803115,0.0057639806,0.00044689054,0.002611346,0.0028011221,0.000345476],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003977637,0.0014952915,0.0012285126,0.0033510143,0.0008977973,0.008430454,0.0018822756,0.0015482605,0.08668551],"category_scores_gemma":[0.041305207,0.0010026433,0.0015311154,0.0028415231,0.0007992309,0.004385833,0.003005253,0.0018836969,0.019697387],"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":[0.00071666576,0.00007873668,0.001033172,0.001742459,0.00010988964,0.0010850214,0.0010933764,0.084885836,0.03433359,0.22922334,0.08483292,0.560865],"study_design_scores_gemma":[0.00032352295,0.00017082653,0.0007109764,0.00063182565,0.000115428884,0.0008307155,0.00040699707,0.38233525,0.04075777,0.10790672,0.46554887,0.00026115455],"about_ca_topic_score_codex":0.0013535382,"about_ca_topic_score_gemma":0.0014744764,"teacher_disagreement_score":0.08668551,"about_ca_system_score_codex":0.0011982482,"about_ca_system_score_gemma":0.0014278317,"threshold_uncertainty_score":0.28999192},"labels":[],"label_agreement":null},{"id":"W2621981149","doi":"10.1002/minf.201700022","title":"Non a Priori Automatic Discovery of 3D Chemical Patterns: Application to Mutagenicity","year":2017,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Artificial Intelligence in Medicine (Canada)","funders":"European Regional Development Fund; Région Normandie","keywords":"Cheminformatics; A priori and a posteriori; Computer science; Relation (database); Set (abstract data type); Feature (linguistics); Property (philosophy); Function (biology); Fragment (logic); Data mining; Biological system; Theoretical computer science; Algorithm; Chemistry; Computational chemistry; Biology","score_opus":0.010406116238549094,"score_gpt":0.291363309884864,"score_spread":0.2809571936463149,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2621981149","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":[0.17936568,0.0010006952,0.80880314,0.00044982528,0.00004272633,0.00020893775,0.0031630755,0.0047733807,0.0021924877],"genre_scores_gemma":[0.4605474,0.00036472175,0.53335434,0.00010166404,0.000032517088,0.00016458827,0.004623385,0.00024142975,0.0005700376],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991524,0.00024175069,0.00006633888,0.00024761728,0.00024988517,0.000041978215],"domain_scores_gemma":[0.9961468,0.0022956743,0.00047383024,0.00059677835,0.00038166033,0.00010524476],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008802959,0.00063739624,0.00085481134,0.0024178121,0.00045104657,0.0008898205,0.0010786626,0.000840953,0.0015102064],"category_scores_gemma":[0.005346052,0.0003905264,0.0010243728,0.0018052644,0.00067927846,0.0008925577,0.0011224588,0.00082682905,0.00058635254],"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":[0.0009316873,0.00043892805,0.034434337,0.0011665684,0.00029629905,0.00065942574,0.00030434408,0.23029217,0.0749038,0.019436933,0.004815809,0.6323197],"study_design_scores_gemma":[0.000051253188,0.00014721448,0.0046499344,0.00003320434,0.00004421378,0.00029596454,0.000054651948,0.9545416,0.013353787,0.02324208,0.0035599791,0.000026241754],"about_ca_topic_score_codex":0.0018989433,"about_ca_topic_score_gemma":0.0032849591,"teacher_disagreement_score":0.0024178121,"about_ca_system_score_codex":0.00045182504,"about_ca_system_score_gemma":0.0008439067,"threshold_uncertainty_score":0.0050521493},"labels":[],"label_agreement":null},{"id":"W2799656935","doi":"10.1002/minf.201800043","title":"Cheminformatics Driven Development of Novel Therapies for Drug Resistant Prostate Cancer","year":2018,"lang":"en","type":"review","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia; Prevention of Organ Failure","funders":"Canadian Institutes of Health Research; Genome British Columbia; Prostate Cancer Canada","keywords":"Cheminformatics; Androgen receptor; Prostate cancer; Antiandrogens; Drug resistance; Computational biology; Drug development; Regulator; Drug; Chemistry; Pharmacology; Cancer; Cancer research; Bioinformatics; Medicine; Biology; Internal medicine; Biochemistry; Genetics","score_opus":0.051230954180605616,"score_gpt":0.35990716918224513,"score_spread":0.3086762150016395,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2799656935","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":[0.0004656515,0.9913368,0.0023793543,0.00068029587,0.000257557,0.000016300566,0.00006031627,0.000044446995,0.0047592307],"genre_scores_gemma":[0.0031682078,0.9923779,0.0019334406,0.00043751154,0.00016839862,0.000020072968,0.000111222034,0.000013043357,0.0017701323],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9997626,0.00005566076,0.000017086375,0.000043526565,0.00009689909,0.000024259172],"domain_scores_gemma":[0.9996418,0.0001911037,0.00003479278,0.00001797044,0.00008696018,0.000027337594],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00064905547,0.0008074878,0.0010293135,0.001885978,0.00025450048,0.0011786288,0.0010824864,0.0010656705,0.004263359],"category_scores_gemma":[0.00090977474,0.0003187988,0.00057514815,0.0020385126,0.00046318138,0.0013818627,0.0008854821,0.0021905464,0.0025390417],"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":[0.000045280427,0.00009882876,0.00014797066,0.0117241,0.0001494503,0.00015370193,0.000072863,0.0026158635,0.0025111209,0.032234456,0.028452748,0.9217936],"study_design_scores_gemma":[0.000016647235,0.00006897508,0.00028044992,0.00209737,0.00007559424,0.00057846855,0.000033560416,0.0009954037,0.0014592594,0.008940152,0.9854301,0.000024050583],"about_ca_topic_score_codex":0.00079480873,"about_ca_topic_score_gemma":0.0011635066,"teacher_disagreement_score":0.004263359,"about_ca_system_score_codex":0.00069260376,"about_ca_system_score_gemma":0.0010976907,"threshold_uncertainty_score":0.014262378},"labels":[],"label_agreement":null},{"id":"W2919214297","doi":"10.1002/minf.201800160","title":"Application of an Iterative Fragment Selection (IFS) Method to Estimate Entropies of Fusion and Melting Points of Organic Chemicals","year":2019,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":17,"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; ASL Environmental Sciences (Canada); ARC Resources (Canada); The Scarborough Hospital; Greenfield Research (Canada)","funders":"European Chemical Industry Council","keywords":"Selection (genetic algorithm); Mean squared error; Computer science; Biological system; Mathematics; Artificial intelligence; Biology; Statistics","score_opus":0.005338842299840955,"score_gpt":0.30577566503761855,"score_spread":0.3004368227377776,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2919214297","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":[0.38465023,0.00057206815,0.6104683,0.0001067843,0.000024296367,0.00014648896,0.00054966455,0.0016215129,0.0018606167],"genre_scores_gemma":[0.88108236,0.0001252275,0.117694,0.000026991971,0.000011578873,0.000102384096,0.00044704607,0.00006685415,0.00044356758],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994734,0.00017839707,0.000020944462,0.000075127566,0.00022597036,0.000026133457],"domain_scores_gemma":[0.9981116,0.0012451526,0.00022391426,0.000086909495,0.0003006339,0.00003169323],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016790192,0.00057000783,0.00064590125,0.0010284578,0.00024010178,0.00029103746,0.0006137333,0.00035368465,0.0007722821],"category_scores_gemma":[0.0034117915,0.00025388828,0.0007422532,0.00050823647,0.00031882524,0.00036587942,0.00035608216,0.0004221709,0.00017806998],"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":[0.00049778266,0.00017387093,0.016962312,0.00024991386,0.00023744293,0.00012456499,0.00010160641,0.73840845,0.07696311,0.0015335615,0.0008860654,0.16386133],"study_design_scores_gemma":[0.000012780025,0.00013918312,0.0026562398,0.0000042183133,0.000017455515,0.000034389937,0.000005882493,0.9799129,0.016834592,0.00018525503,0.00018319946,0.000013841781],"about_ca_topic_score_codex":0.0037328089,"about_ca_topic_score_gemma":0.0024130612,"teacher_disagreement_score":0.0037328089,"about_ca_system_score_codex":0.0005036797,"about_ca_system_score_gemma":0.0005404918,"threshold_uncertainty_score":0.008879602},"labels":[],"label_agreement":null},{"id":"W2944666543","doi":"10.1002/minf.201800110","title":"QSPRs for Molecular Diffusion Coefficients in Polymeric Passive Samplers: A Comparison of Simple Molecular and Quantum‐mechanical Sigma‐moment Descriptors","year":2019,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Trent University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Quantitative structure–activity relationship; Molecular descriptor; Low-density polyethylene; Mean squared error; Correlation coefficient; Chemistry; Moment (physics); Diffusion; Computational chemistry; Mathematics; Polyethylene; Thermodynamics; Organic chemistry; Stereochemistry; Statistics; Physics","score_opus":0.016193422162381347,"score_gpt":0.3002332653078655,"score_spread":0.28403984314548414,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2944666543","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":[0.6542438,0.0038593644,0.33522224,0.0003075632,0.000040316816,0.0002855968,0.0023424742,0.0011001291,0.0025985267],"genre_scores_gemma":[0.95780766,0.00090389524,0.039012697,0.000076335105,0.000021581487,0.0001443438,0.0012644933,0.00003696038,0.0007319921],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990901,0.00036010807,0.000047275513,0.00010556216,0.0003591904,0.000037739992],"domain_scores_gemma":[0.99549073,0.003550203,0.00049251626,0.0001665887,0.00027236086,0.000027602204],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0031064732,0.00068686693,0.00065368065,0.0010480189,0.00009666351,0.00054865103,0.0005665519,0.00047896383,0.0007978905],"category_scores_gemma":[0.004943199,0.00025619738,0.0011266345,0.0007748592,0.0003798503,0.00081051094,0.00038189703,0.0006041088,0.00030227692],"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":[0.0012250524,0.00053906214,0.027067939,0.001250038,0.0006709943,0.00016875843,0.00012050572,0.58911616,0.23638307,0.004196065,0.0010346613,0.13822773],"study_design_scores_gemma":[0.00005700838,0.0013343196,0.015453775,0.000042853873,0.00013292683,0.000108087195,0.000044846016,0.8932794,0.08675705,0.0013035947,0.0014080944,0.000078081604],"about_ca_topic_score_codex":0.0018627758,"about_ca_topic_score_gemma":0.0017762656,"teacher_disagreement_score":0.0031064732,"about_ca_system_score_codex":0.0005485216,"about_ca_system_score_gemma":0.0004691994,"threshold_uncertainty_score":0.016428828},"labels":[],"label_agreement":null},{"id":"W2947303722","doi":"10.1002/minf.201900024","title":"Molecular Modelling of Potential Candidates for the Treatment of Depression","year":2019,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Conselho Nacional de Desenvolvimento Científico e Tecnológico; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior","keywords":"Pharmacophore; Quantitative structure–activity relationship; Computational biology; Serotonin transporter; Docking (animal); Mode of action; Isoquinoline; Chemistry; Pharmacology; Serotonin; Stereochemistry; Biology; Medicine; Receptor; Biochemistry","score_opus":0.01543643617665239,"score_gpt":0.2676626665438475,"score_spread":0.2522262303671951,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2947303722","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":[0.70755905,0.0048272084,0.18208784,0.0021406547,0.00032551712,0.0005186312,0.005629124,0.0011214623,0.09579059],"genre_scores_gemma":[0.97003436,0.0010538994,0.021314038,0.00011528958,0.000020971918,0.0002820166,0.0010677222,0.000051772364,0.0060599884],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9999013,0.000036199515,0.000003989478,0.000011327312,0.000028007624,0.000019205974],"domain_scores_gemma":[0.99986124,0.00007740694,0.000017308263,0.00000764506,0.000025108075,0.000011257987],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00022918875,0.0005782471,0.0008291403,0.00057457964,0.00032536482,0.0009115952,0.00069769163,0.0010555283,0.009381132],"category_scores_gemma":[0.00054297806,0.00030556094,0.00062109204,0.00060249504,0.0002926652,0.00032597364,0.00033247905,0.00046407434,0.0010338761],"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":[0.000081107486,0.000038903814,0.00035406754,0.00007596474,0.000017147424,0.000078460704,0.000011053715,0.9919708,0.0017661398,0.0025201316,0.00040355348,0.0026826523],"study_design_scores_gemma":[0.00004571305,0.00007654912,0.00018199836,0.000010197841,0.000013438959,0.0000147483,0.00001449881,0.99664235,0.00072623073,0.0008570029,0.0014116646,0.0000056625877],"about_ca_topic_score_codex":0.005089136,"about_ca_topic_score_gemma":0.003546982,"teacher_disagreement_score":0.009381132,"about_ca_system_score_codex":0.0007133142,"about_ca_system_score_gemma":0.00089689484,"threshold_uncertainty_score":0.031383038},"labels":[],"label_agreement":null},{"id":"W2981596486","doi":"10.1002/minf.201900096","title":"An <i>in silico</i> Approach for Integrating Phenotypic and Target‐based Approaches in Drug Discovery","year":2019,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Innovation Cluster (Canada)","funders":"Core Research for Evolutional Science and Technology; Japan Society for the Promotion of Science; New Energy and Industrial Technology Development Organization","keywords":"In silico; Drug discovery; Computational biology; Phenotypic screening; Computer science; Deconvolution; Phenotype; Cheminformatics; Drug target; Systems biology; Bioinformatics; Biology; Genetics; Algorithm; Pharmacology; Gene","score_opus":0.018396756260697183,"score_gpt":0.2593566582448479,"score_spread":0.2409599019841507,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2981596486","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":[0.0021360463,0.00005931156,0.9954991,0.00013846991,0.000029154093,0.000048201196,0.000049059057,0.00038287102,0.0016578505],"genre_scores_gemma":[0.1208247,0.0004972503,0.87555796,0.0003514304,0.000075463686,0.00032078155,0.0002335297,0.00023532315,0.0019035806],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990901,0.0004955383,0.000041197603,0.00011064899,0.00022505842,0.000037388985],"domain_scores_gemma":[0.9986254,0.00078666664,0.0001617368,0.0002528666,0.000112905305,0.000060393526],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017694843,0.00092682714,0.0009417327,0.0010900145,0.0006164124,0.0012225687,0.0017647393,0.0009785406,0.0024892078],"category_scores_gemma":[0.0029370792,0.0008647919,0.0015668329,0.00071879657,0.0010918461,0.0014287246,0.0015332928,0.0015501892,0.00075865304],"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":[0.00016445018,0.0004882033,0.003932805,0.0005142841,0.0004855187,0.00039971783,0.00015012322,0.6178663,0.09677148,0.17614569,0.0032406135,0.099840805],"study_design_scores_gemma":[0.000020490757,0.00011407006,0.00033593015,0.000016258478,0.00005496853,0.00020221107,0.00001313262,0.96005356,0.012344701,0.021961093,0.0048566125,0.000026988799],"about_ca_topic_score_codex":0.0013372615,"about_ca_topic_score_gemma":0.0016372583,"teacher_disagreement_score":0.0024892078,"about_ca_system_score_codex":0.00080331165,"about_ca_system_score_gemma":0.001119551,"threshold_uncertainty_score":0.009358048},"labels":[],"label_agreement":null},{"id":"W3009970748","doi":"10.1002/minf.202000028","title":"Rapid Identification of Potential Inhibitors of SARS‐CoV‐2 Main Protease by Deep Docking of 1.3 Billion Compounds","year":2020,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":518,"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":"National Institute on Deafness and Other Communication Disorders; Canadian Institutes of Health Research","keywords":"Docking (animal); Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Virtual screening; Coronavirus disease 2019 (COVID-19); Drug discovery; Protease; Coronavirus; Computational biology; Drug development; 2019-20 coronavirus outbreak; Drug; Computer science; Virology; Infectious disease (medical specialty); Medicine; Chemistry; Outbreak; Biology; Bioinformatics; Pharmacology; Disease; Biochemistry; Enzyme","score_opus":0.01770774515941695,"score_gpt":0.2675397446277249,"score_spread":0.24983199946830795,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3009970748","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":[0.90912706,0.009852291,0.05168629,0.0010502742,0.00017356013,0.00036658574,0.004223952,0.006084344,0.017435651],"genre_scores_gemma":[0.9474475,0.0027915505,0.03826631,0.0005011501,0.000030014788,0.00017910595,0.005822221,0.00011817751,0.004843968],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9998055,0.000035216362,0.000011644021,0.000034992092,0.00006424765,0.000048363334],"domain_scores_gemma":[0.9999027,0.000029879673,0.000017147868,0.0000131681545,0.000020315916,0.000016760257],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00033560456,0.0011957188,0.0011475891,0.0008281029,0.00025118297,0.00070786965,0.0006093061,0.000575681,0.0026257506],"category_scores_gemma":[0.00064151204,0.00028481512,0.0010048762,0.00071849074,0.00018759591,0.00048089738,0.00060049514,0.0006331801,0.00063426205],"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":[0.001254312,0.0013746223,0.009200648,0.0010066289,0.0006762725,0.0003565851,0.000046864516,0.6707702,0.069534525,0.0029646682,0.014330725,0.22848392],"study_design_scores_gemma":[0.00023397159,0.00093845505,0.0016062356,0.000024445495,0.00012977068,0.00010256165,0.000030151747,0.9652132,0.026224917,0.0013226161,0.0041474635,0.00002629792],"about_ca_topic_score_codex":0.0033528048,"about_ca_topic_score_gemma":0.005770765,"teacher_disagreement_score":0.0033528048,"about_ca_system_score_codex":0.000504485,"about_ca_system_score_gemma":0.00071182434,"threshold_uncertainty_score":0.008784056},"labels":[],"label_agreement":null},{"id":"W3037787710","doi":"10.1002/minf.202000033","title":"Using Language Representation Learning Approach to Efficiently Identify Protein Complex Categories in Electron Transport Chain","year":2020,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Machine Learning in Bioinformatics","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":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Computer science; Artificial intelligence; Feature learning; Representation (politics); Word embedding; Natural language processing; Machine learning; Feature (linguistics); Pattern recognition (psychology); Support vector machine; Categorization; Multi-task learning; Embedding; Task (project management)","score_opus":0.02120448430788708,"score_gpt":0.3007711008899546,"score_spread":0.27956661658206755,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3037787710","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":[0.17910717,0.00065970636,0.8133099,0.0005319896,0.00006894759,0.00014140789,0.00049207686,0.0024938986,0.003194954],"genre_scores_gemma":[0.71490395,0.0002933015,0.28008756,0.0001963601,0.00005546996,0.00019176204,0.0013010734,0.000095031544,0.002875464],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999539,0.00012604508,0.000043412005,0.00011924141,0.000091392605,0.000080981714],"domain_scores_gemma":[0.99945,0.00020239834,0.00010019578,0.00006418097,0.00015114331,0.000032037937],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00064810016,0.00063240837,0.00049787224,0.0018641647,0.0003961405,0.00089030433,0.00068419153,0.00057526823,0.0013385357],"category_scores_gemma":[0.0015547207,0.00016130044,0.0008328981,0.0011009747,0.00038427065,0.001984435,0.00079832546,0.0007498265,0.00064438436],"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":[0.00037447695,0.00043519036,0.012717697,0.00035723572,0.00014807086,0.0005274544,0.00063090655,0.07081539,0.083183974,0.019999636,0.006354911,0.80445504],"study_design_scores_gemma":[0.000026072248,0.00016385212,0.002664255,0.000023700211,0.00005200586,0.00025492453,0.00019106298,0.95772105,0.01640219,0.019549098,0.0029047304,0.000047074336],"about_ca_topic_score_codex":0.0021445039,"about_ca_topic_score_gemma":0.0016797963,"teacher_disagreement_score":0.0021445039,"about_ca_system_score_codex":0.00058398413,"about_ca_system_score_gemma":0.0007496632,"threshold_uncertainty_score":0.0044778585},"labels":[],"label_agreement":null},{"id":"W4206247747","doi":"10.1002/minf.202100264","title":"An AI‐based Prediction Model for Drug‐drug Interactions in Osteoporosis and Paget's Diseases from SMILES","year":2022,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Bone health and treatments","field":"Medicine","cited_by":89,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Ministry of Science and Technology, Taiwan","keywords":"Drug; Cheminformatics; Osteoporosis; Computer science; Medicine; Pharmacology; Chemistry; Internal medicine; Computational chemistry","score_opus":0.013212614403985358,"score_gpt":0.2909295781902771,"score_spread":0.27771696378629174,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4206247747","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":[0.63250864,0.0052868,0.34397897,0.0029057665,0.00029748777,0.0004275155,0.0061985548,0.0027492018,0.0056470837],"genre_scores_gemma":[0.9598682,0.0007614751,0.03320107,0.0003629434,0.00010445843,0.00024500757,0.003548343,0.000025492982,0.0018829802],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9996809,0.00007571099,0.000035637237,0.00008372471,0.0000723408,0.000051745857],"domain_scores_gemma":[0.99866414,0.0008814191,0.00013706418,0.000030198064,0.00023857728,0.000048627797],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000994593,0.0009827446,0.0008998313,0.001540796,0.00031485554,0.0006318496,0.0007229997,0.0008094705,0.0013046365],"category_scores_gemma":[0.0023797047,0.00024485978,0.0009574363,0.0008872578,0.00021107093,0.0005080991,0.00032151103,0.0008477321,0.0003622499],"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":[0.0004032273,0.00047127978,0.029670337,0.00015563027,0.00023129366,0.00031287197,0.000050939518,0.85393286,0.0013426581,0.0011430702,0.004306119,0.107979685],"study_design_scores_gemma":[0.000009053476,0.000043818367,0.001305425,0.000005927683,0.00002433249,0.00002364951,0.0000046561718,0.9977254,0.00017659646,0.00042837212,0.0002486273,0.0000040529862],"about_ca_topic_score_codex":0.017537026,"about_ca_topic_score_gemma":0.014481684,"teacher_disagreement_score":0.017537026,"about_ca_system_score_codex":0.00070189807,"about_ca_system_score_gemma":0.0010615936,"threshold_uncertainty_score":0.03486991},"labels":[],"label_agreement":null},{"id":"W4220753843","doi":"10.1002/minf.202100240","title":"A Descriptor Set for Quantitative Structure‐property Relationship Prediction in Biologics","year":2022,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Monoclonal and Polyclonal Antibodies Research","field":"Medicine","cited_by":78,"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; University of Waterloo","funders":"","keywords":"Quantitative structure–activity relationship; In silico; Computer science; Stability (learning theory); Drug development; Artificial intelligence; Machine learning; Drug discovery; Set (abstract data type); Process (computing); Biochemical engineering; Data mining; Biological system; Computational biology; Drug; Chemistry; Bioinformatics; Biology","score_opus":0.08329791264474877,"score_gpt":0.3406507462846687,"score_spread":0.2573528336399199,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4220753843","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":[0.18907765,0.0015380654,0.79503745,0.00067087804,0.00008787169,0.00043714297,0.008209409,0.0020621375,0.0028794026],"genre_scores_gemma":[0.74089825,0.0005555332,0.24739072,0.00019143973,0.000055431046,0.0007411288,0.009232412,0.00009402684,0.0008410016],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999225,0.0002789487,0.00007059105,0.0001257884,0.00024302905,0.000056550933],"domain_scores_gemma":[0.9968021,0.0025899236,0.00015461254,0.00012209787,0.00027864063,0.000052576735],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019931667,0.0010641691,0.0012646767,0.0021783889,0.0004800583,0.0010403785,0.0010156623,0.0010513214,0.0019491777],"category_scores_gemma":[0.0056517827,0.0003143145,0.0018622732,0.0017829438,0.00061239494,0.0008221493,0.0008400781,0.0013117971,0.0002812784],"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":[0.000071624745,0.00015165201,0.0021074242,0.00017832985,0.00006711244,0.00007272722,0.000023292969,0.9608302,0.0014469156,0.0032680933,0.0010570496,0.030725535],"study_design_scores_gemma":[0.000015312193,0.00005855523,0.00033061992,0.0000092758955,0.000011834197,0.000013151847,0.000010052319,0.9964259,0.0004929056,0.0022052838,0.00042038478,0.00000677196],"about_ca_topic_score_codex":0.006789865,"about_ca_topic_score_gemma":0.0049504186,"teacher_disagreement_score":0.006789865,"about_ca_system_score_codex":0.001107282,"about_ca_system_score_gemma":0.0012401019,"threshold_uncertainty_score":0.01350069},"labels":[],"label_agreement":null},{"id":"W4313816625","doi":"10.1002/minf.202200186","title":"A comparison between 2D and 3D descriptors in QSAR modeling based on bio‐active conformations","year":2023,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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 British Columbia Hospital","funders":"","keywords":"Quantitative structure–activity relationship; Cheminformatics; Random forest; Protein Data Bank (RCSB PDB); Molecular descriptor; Lasso (programming language); Computer science; Set (abstract data type); Artificial intelligence; Test set; Data mining; Pattern recognition (psychology); Machine learning; Chemistry; Computational chemistry; Stereochemistry","score_opus":0.05288652475164202,"score_gpt":0.337118736541096,"score_spread":0.28423221178945396,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4313816625","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":[0.434488,0.009096988,0.54060674,0.0009829961,0.00016523941,0.00030680216,0.00510168,0.0015038198,0.0077477256],"genre_scores_gemma":[0.897436,0.002666413,0.094733335,0.00016665082,0.0000562229,0.00027195414,0.0039285608,0.00017845567,0.0005622532],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9981712,0.00094869884,0.00015334254,0.00019955834,0.00043669224,0.00009044595],"domain_scores_gemma":[0.9964843,0.0025974174,0.00024779045,0.00037576872,0.0002440917,0.000050620525],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003480259,0.0009034429,0.001285922,0.0024513158,0.00035209322,0.0013841848,0.0008375471,0.0007327533,0.0011796547],"category_scores_gemma":[0.0061945515,0.0003256503,0.0015973251,0.0027737182,0.00052123744,0.0013521943,0.0008763398,0.0008517206,0.00043888486],"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":[0.00043933527,0.00013508218,0.0065749227,0.00027585187,0.00019522959,0.00007244364,0.000053317548,0.9249422,0.002910662,0.005101214,0.0007731951,0.058526512],"study_design_scores_gemma":[0.000039319882,0.00022072394,0.002339981,0.000035187175,0.000040841118,0.000050102095,0.000027023465,0.99031645,0.0015971102,0.0037595679,0.0015397,0.00003401973],"about_ca_topic_score_codex":0.004274156,"about_ca_topic_score_gemma":0.0028903258,"teacher_disagreement_score":0.004274156,"about_ca_system_score_codex":0.0006818989,"about_ca_system_score_gemma":0.0008140834,"threshold_uncertainty_score":0.018405557},"labels":[],"label_agreement":null},{"id":"W4376636211","doi":"10.1002/minf.202300026","title":"Novel Inhibitors of androgen receptor's DNA binding domain identified using an ultra‐large virtual screening","year":2023,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Prostate Cancer Treatment and Research","field":"Medicine","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":"University of British Columbia","funders":"DOD Prostate Cancer Research Program; Canadian Cancer Society Research Institute; U.S. Department of Defense","keywords":"Androgen receptor; Virtual screening; Computational biology; Chemistry; Domain (mathematical analysis); DNA; Computer science; Biology; Genetics; Drug discovery; Biochemistry; Prostate cancer; Mathematics","score_opus":0.04133740417371557,"score_gpt":0.3232444522595325,"score_spread":0.28190704808581696,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4376636211","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":[0.92582023,0.011618157,0.03710226,0.0008320815,0.00018995815,0.00081912224,0.0037647013,0.0015652243,0.018288312],"genre_scores_gemma":[0.96108365,0.0056706546,0.025373677,0.0004541987,0.000029914372,0.00039339534,0.0031901253,0.000058801867,0.003745673],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9998635,0.00003443763,0.000006293703,0.000030332843,0.00003863583,0.00002677155],"domain_scores_gemma":[0.9999298,0.000026724294,0.000014073907,0.0000057994216,0.00001049652,0.00001314402],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00030367353,0.0009350562,0.0009399358,0.0007310824,0.00026812524,0.00059043686,0.00059745234,0.00027518597,0.0021659816],"category_scores_gemma":[0.00033194444,0.00019273476,0.00077282387,0.00060333346,0.00023043422,0.0002867808,0.0004970197,0.0004943275,0.00031256516],"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":[0.0030345202,0.002563093,0.0074524675,0.002634009,0.0011692587,0.0021864313,0.00013010477,0.44797423,0.28783822,0.010698073,0.011573813,0.22274579],"study_design_scores_gemma":[0.0028762326,0.011906866,0.008742691,0.00023655826,0.0023849583,0.0019274915,0.00024322349,0.73837245,0.16793133,0.006625677,0.058502324,0.0002502039],"about_ca_topic_score_codex":0.0009762532,"about_ca_topic_score_gemma":0.0026161985,"teacher_disagreement_score":0.0021659816,"about_ca_system_score_codex":0.0004303917,"about_ca_system_score_gemma":0.00068426493,"threshold_uncertainty_score":0.0072458982},"labels":[],"label_agreement":null},{"id":"W4408701827","doi":"10.1002/minf.70029","title":"The Use of DeepQSAR Models for the Discovery of Peptides With Enhanced Antimicrobial and Antibiofilm Potential","year":2025,"lang":"en","type":"preprint","venue":"Molecular Informatics","topic":"Machine Learning in Bioinformatics","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 British Columbia","funders":"British Columbia Knowledge Development Fund; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; Canada Foundation for Innovation","keywords":"Antimicrobial; Antimicrobial peptides; Chemistry; Biochemical engineering; Nanotechnology; Computational biology; Microbiology; Biology; Materials science; Engineering","score_opus":0.010574617349845829,"score_gpt":0.2379863335103369,"score_spread":0.22741171616049108,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408701827","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":[0.2063974,0.0030141775,0.78493065,0.000636221,0.00008146526,0.000079818754,0.00040817403,0.0012993212,0.0031527423],"genre_scores_gemma":[0.86162925,0.0018898909,0.13292058,0.00045617248,0.000041940337,0.00018948565,0.0007130031,0.00007687043,0.0020828797],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99976796,0.000066152934,0.000018970186,0.000044827484,0.00007052676,0.000031582516],"domain_scores_gemma":[0.9994887,0.00027849092,0.000084363055,0.000031723186,0.000092667244,0.000023950166],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010771962,0.0010941244,0.0009833786,0.00046005016,0.00016220921,0.0006062606,0.00082375377,0.00092024956,0.0008659901],"category_scores_gemma":[0.0012250055,0.00046199124,0.0009550274,0.00040669288,0.00046693286,0.0008915136,0.00057709694,0.0015577563,0.00026626582],"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":[0.000055107117,0.000067581204,0.00063801225,0.00011188897,0.000048590922,0.00005156598,0.000016262762,0.958976,0.019466981,0.002242579,0.00023557943,0.018089835],"study_design_scores_gemma":[0.000003820528,0.000040662628,0.00003953116,0.0000025097204,0.0000064789438,0.0000052340592,0.0000015193913,0.9966286,0.002512139,0.00058903504,0.00016774639,0.0000027390413],"about_ca_topic_score_codex":0.0028520366,"about_ca_topic_score_gemma":0.0027436805,"teacher_disagreement_score":0.0028520366,"about_ca_system_score_codex":0.00075438985,"about_ca_system_score_gemma":0.0006820529,"threshold_uncertainty_score":0.0056967735},"labels":[],"label_agreement":null},{"id":"W4409979195","doi":"10.1002/minf.202500018","title":"Deep Modeling of Gain‐of‐Function Mutations on Androgen Receptor","year":2025,"lang":"en","type":"article","venue":"Molecular Informatics","topic":"Prostate Cancer Treatment and Research","field":"Medicine","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 British Columbia","funders":"","keywords":"Androgen receptor; Prostate cancer; Mutant; Computational biology; Mutation; Antiandrogens; Function (biology); Computer science; Biology; Bioinformatics; Genetics; Gene; Cancer","score_opus":0.01792586232352345,"score_gpt":0.3033978119549671,"score_spread":0.2854719496314437,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4409979195","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":[0.53563136,0.0018395091,0.45278332,0.0009050259,0.00012958798,0.000053882057,0.0010652518,0.0012291917,0.006362908],"genre_scores_gemma":[0.9745213,0.0005134476,0.021465437,0.00018021715,0.000023162649,0.00006750362,0.0005550015,0.00005528543,0.002618764],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99992585,0.000021074025,0.0000034354862,0.000016582826,0.000014011685,0.000018923525],"domain_scores_gemma":[0.9998542,0.00009222848,0.00001782622,0.000008453153,0.000014924793,0.000012279375],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00023882004,0.0006666823,0.00057511305,0.00024671695,0.00012695741,0.0004765043,0.0005804675,0.00067406846,0.0010795863],"category_scores_gemma":[0.00061068946,0.00029363166,0.000735481,0.00023944041,0.0003089772,0.00033382955,0.0003416251,0.00073665136,0.00016201708],"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":[0.00004260582,0.000025482015,0.0005862356,0.000023128117,0.000024829584,0.00003711945,0.000005583452,0.99127114,0.0019783292,0.00091143185,0.00034279042,0.004751308],"study_design_scores_gemma":[0.0000016818376,0.000009116501,0.000046903402,7.28343e-7,0.0000025886604,0.000003270994,0.0000010231278,0.999171,0.0003308056,0.00035481344,0.00007709213,8.8240193e-7],"about_ca_topic_score_codex":0.005476729,"about_ca_topic_score_gemma":0.005392646,"teacher_disagreement_score":0.005476729,"about_ca_system_score_codex":0.0004302024,"about_ca_system_score_gemma":0.0006734077,"threshold_uncertainty_score":0.010889709},"labels":[],"label_agreement":null}]}