{"meta":{"query_hash":"09521ba2f298","filters":{"venue":"Journal of Cheminformatics"},"cohort_total":43,"direct_labels_cover":0,"predictions_cover":43,"exported":43,"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/09521ba2f298","api":"https://metacan.xera.ac/api/v1/cohort?venue=Journal+of+Cheminformatics"},"results":[{"id":"W1495258543","doi":"10.1186/s13321-015-0085-3","title":"ChemDoodle Web Components: HTML5 toolkit for chemical graphics, interfaces, and informatics","year":2015,"lang":"en","type":"review","venue":"Journal of Cheminformatics","topic":"Various Chemistry Research Topics","field":"Chemistry","cited_by":75,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Research Canada; University of Toronto","funders":"Canadian Institutes of Health Research","keywords":"HTML5; JavaScript; Computer science; World Wide Web; Graphics; Multimedia; Web application; Computer graphics; Computer graphics (images)","score_opus":0.0822703604728123,"score_gpt":0.35068284433762875,"score_spread":0.2684124838648164,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1495258543","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":[0.002970554,0.12759624,0.36433184,0.003011239,0.0022135573,0.0017120654,0.048860665,0.16226089,0.28704295],"genre_scores_gemma":[0.034026306,0.160252,0.4220993,0.0057469085,0.001110394,0.0031333845,0.11675207,0.026569685,0.23030998],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99947447,0.00008101264,0.000046971014,0.000065326705,0.00028182127,0.000050458522],"domain_scores_gemma":[0.99948776,0.00014451762,0.00004175155,0.000051058178,0.00019121409,0.00008366942],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010126358,0.0015951433,0.0011852814,0.0023476197,0.00048671776,0.0022477563,0.002468616,0.0011643043,0.09533019],"category_scores_gemma":[0.0015064763,0.0008872126,0.0010026767,0.0022904747,0.00037185466,0.0021982454,0.0014776897,0.0018668949,0.09353561],"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":[0.00009618731,0.0000768197,0.00014745494,0.0058736447,0.00006898016,0.00014948874,0.00010697405,0.0008196308,0.011646695,0.014669402,0.5293268,0.43701792],"study_design_scores_gemma":[0.000017802902,0.000009191834,0.00009667936,0.00014455758,0.000013660911,0.00013316696,0.000008079955,0.00041510415,0.00427223,0.0015521778,0.9933177,0.000019661808],"about_ca_topic_score_codex":0.0010868747,"about_ca_topic_score_gemma":0.001687615,"teacher_disagreement_score":0.09533019,"about_ca_system_score_codex":0.0005807379,"about_ca_system_score_gemma":0.001689906,"threshold_uncertainty_score":0.31891125},"labels":[],"label_agreement":null},{"id":"W1722424871","doi":"10.1186/1758-2946-5-s1-p40","title":"A 3D-QSAR model for cannabinoid receptor (CB2) ligands derived from aligned pharmacophors","year":2013,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Pharmacological Receptor Mechanisms and Effects","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"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":"","keywords":"Pharmacophore; Cannabinoid receptor; Cannabinoid receptor type 2; Positron emission tomography; Radioligand; Neuroscience; Drug discovery; Cannabinoid; Computational biology; Quantitative structure–activity relationship; Receptor; Chemistry; Computer science; Biology; Biochemistry; Stereochemistry; Agonist","score_opus":0.012406226783392487,"score_gpt":0.26236953832106275,"score_spread":0.24996331153767026,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1722424871","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":[0.46775693,0.006291424,0.48394138,0.0013594159,0.0002843494,0.0008345643,0.0159235,0.0023991012,0.021209292],"genre_scores_gemma":[0.89148545,0.0017278885,0.09522317,0.00026172315,0.00006044639,0.0011908857,0.0063552465,0.00016208304,0.003533067],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9997631,0.00008298497,0.000013518454,0.00004078484,0.00006531492,0.000034380388],"domain_scores_gemma":[0.9995683,0.0003034441,0.000034867517,0.0000131153565,0.00006284328,0.00001747603],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00058719027,0.0012082562,0.001207869,0.0009412747,0.0004173617,0.0010676454,0.0010755336,0.0017155396,0.0065057827],"category_scores_gemma":[0.00088750216,0.00039934326,0.0016991777,0.0008954881,0.00040539616,0.00037414944,0.00044059302,0.0012331917,0.0009063191],"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":[0.00004341769,0.000024200963,0.0002904103,0.00007095748,0.000023696552,0.000038983704,0.000008315449,0.99532074,0.00075007096,0.00044087882,0.00028861786,0.0026997612],"study_design_scores_gemma":[0.000019369607,0.0000758473,0.00023001769,0.000010545998,0.000020060634,0.000015866544,0.0000075740213,0.9982243,0.0004242523,0.00036597296,0.00059931737,0.0000068784116],"about_ca_topic_score_codex":0.015628617,"about_ca_topic_score_gemma":0.0111259045,"teacher_disagreement_score":0.015628617,"about_ca_system_score_codex":0.001122207,"about_ca_system_score_gemma":0.001384527,"threshold_uncertainty_score":0.031075299},"labels":[],"label_agreement":null},{"id":"W174610764","doi":"10.1186/1758-2946-4-s1-p3","title":"Targeting protein-protein interactions using methods of cheminformatics","year":2012,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Microbial Natural Products and Biosynthesis","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":"Cheminformatics; Computer science; Computational biology; Data science; Data mining; Bioinformatics; Biology","score_opus":0.036522105356600616,"score_gpt":0.34954319290856406,"score_spread":0.31302108755196345,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W174610764","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":[0.012678415,0.002298745,0.97293144,0.0005025529,0.000063110194,0.00016871776,0.00064782833,0.002956147,0.0077529904],"genre_scores_gemma":[0.13673313,0.004573419,0.85467017,0.00029363515,0.000080178776,0.0005443569,0.0011274804,0.00033345088,0.0016442223],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994099,0.0002033353,0.000026975404,0.00010728197,0.0002298113,0.000022780456],"domain_scores_gemma":[0.9995084,0.00027377188,0.000068988455,0.00008624833,0.000043259577,0.000019377125],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009535122,0.0011868983,0.0010467108,0.0016341497,0.00051700673,0.001656075,0.0011596847,0.00070331164,0.0035065715],"category_scores_gemma":[0.0012254274,0.00044788505,0.00092139887,0.0014645848,0.0009058444,0.001338024,0.0011693331,0.0016277916,0.0016369397],"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":[0.0004426491,0.00055091904,0.0017145954,0.002107777,0.0006035951,0.0004093405,0.00025636624,0.20464095,0.19462733,0.19410275,0.007069978,0.39347377],"study_design_scores_gemma":[0.00019656641,0.0002601354,0.0008296118,0.000092239854,0.00014983576,0.000313366,0.000056344346,0.6456944,0.15900314,0.12622155,0.067074895,0.00010785803],"about_ca_topic_score_codex":0.0008042384,"about_ca_topic_score_gemma":0.0010645235,"teacher_disagreement_score":0.0035065715,"about_ca_system_score_codex":0.0011020084,"about_ca_system_score_gemma":0.0011719352,"threshold_uncertainty_score":0.011730671},"labels":[],"label_agreement":null},{"id":"W1850375728","doi":"10.1186/s13321-015-0092-4","title":"Visual analytics in cheminformatics: user-supervised descriptor selection for QSAR methods","year":2015,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":32,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"Natural Sciences and Engineering Research Council of Canada; Consejo Nacional de Investigaciones Científicas y Técnicas","keywords":"Computer science; Cheminformatics; Interpretability; Quantitative structure–activity relationship; Feature selection; Applicability domain; Visual analytics; Data mining; Machine learning; Artificial intelligence; Set (abstract data type); Selection (genetic algorithm); Property (philosophy); Software; Feature (linguistics); Process (computing); Generality; Domain (mathematical analysis); Visualization; Bioinformatics; Mathematics","score_opus":0.0632942398036562,"score_gpt":0.3859116379978224,"score_spread":0.32261739819416624,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1850375728","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":[0.007495169,0.00030351008,0.9754024,0.0002554897,0.000022354014,0.00015284981,0.0005550537,0.014501276,0.0013118521],"genre_scores_gemma":[0.121985324,0.0004656353,0.87234145,0.00016040007,0.00005543363,0.0006866719,0.0016827589,0.0017542024,0.0008681373],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9968527,0.0016155927,0.00019675566,0.0003914954,0.00084645516,0.00009698512],"domain_scores_gemma":[0.99014366,0.0073060356,0.00041197435,0.0009969794,0.0008985373,0.00024273177],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005951828,0.0018295342,0.0013009674,0.0036907697,0.00047734077,0.0023284447,0.002147174,0.0009946655,0.00711384],"category_scores_gemma":[0.012926275,0.0005814986,0.0016333737,0.002285473,0.0009994511,0.0017575268,0.002430069,0.0016162343,0.002157814],"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":[0.0011181997,0.00047488324,0.0045155874,0.0016771357,0.000329949,0.00032811082,0.0016888143,0.08842261,0.042359848,0.023033218,0.024452824,0.8115987],"study_design_scores_gemma":[0.00018670548,0.00021663563,0.0019902368,0.00022729866,0.00007663927,0.0002693319,0.00025966717,0.89042616,0.03761988,0.04496979,0.023625238,0.0001323594],"about_ca_topic_score_codex":0.0014383949,"about_ca_topic_score_gemma":0.0012294544,"teacher_disagreement_score":0.00711384,"about_ca_system_score_codex":0.00075063296,"about_ca_system_score_gemma":0.0011167858,"threshold_uncertainty_score":0.031476617},"labels":[],"label_agreement":null},{"id":"W2017222169","doi":"10.1186/1758-2946-6-1","title":"Prediction of novel drug indications using network driven biological data prioritization and integration","year":2014,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":51,"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 Regina; University of Calgary","funders":"Natural Sciences and Engineering Research Council of Canada; Alberta Innovates","keywords":"Computer science; Prioritization; Data integration; Data science; Data mining; Drug; Risk analysis (engineering); Management science; Pharmacology; Medicine; Engineering","score_opus":0.03665510091314504,"score_gpt":0.25880139960248577,"score_spread":0.22214629868934072,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2017222169","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":[0.37259784,0.002309682,0.5968478,0.0020440184,0.00009099694,0.0007734666,0.015484964,0.0060331095,0.0038181678],"genre_scores_gemma":[0.7894008,0.0007519096,0.19644183,0.00024625956,0.00007640159,0.00037261398,0.011579804,0.00011706139,0.0010133443],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99961543,0.00010840212,0.000040726616,0.00009852276,0.00010298957,0.000033900287],"domain_scores_gemma":[0.9976042,0.0016004831,0.00032532503,0.00006338611,0.00029970234,0.00010695565],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009391261,0.0010670425,0.0007807833,0.0052112048,0.00032310456,0.0010443844,0.0005971267,0.0006493411,0.002831665],"category_scores_gemma":[0.004033083,0.00026994897,0.0013357907,0.0021884847,0.00023341543,0.00060043024,0.000522902,0.0006656002,0.00048594052],"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":[0.0010552364,0.00049428846,0.049925532,0.00052530394,0.00043403447,0.00058604294,0.000058506987,0.7503597,0.008991825,0.0029892933,0.004195988,0.18038423],"study_design_scores_gemma":[0.000025647001,0.0000654011,0.0024141322,0.00001488356,0.00006974218,0.000080315105,0.000010816233,0.9926923,0.0015056512,0.00244805,0.0006643,0.000008787567],"about_ca_topic_score_codex":0.009137061,"about_ca_topic_score_gemma":0.013162253,"teacher_disagreement_score":0.009137061,"about_ca_system_score_codex":0.0012774275,"about_ca_system_score_gemma":0.0017217802,"threshold_uncertainty_score":0.018167734},"labels":[],"label_agreement":null},{"id":"W2019688884","doi":"10.1186/1758-2946-4-24","title":"mol2chemfig, a tool for rendering chemical structures from molfile or SMILES format to LATE X code","year":2012,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":8,"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; Python (programming language); Rendering (computer graphics); Documentation; Programming language; Source code; Software; Graphics; Computer graphics (images); Coding (social sciences)","score_opus":0.041670292449581564,"score_gpt":0.3269340572626618,"score_spread":0.28526376481308025,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2019688884","genre_codex":"software","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":"software","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0038387184,0.0004812279,0.35250285,0.00048659503,0.00044635823,0.0003013322,0.04854259,0.57146394,0.021936474],"genre_scores_gemma":[0.042764004,0.0024458729,0.54799676,0.0013211003,0.00022806111,0.0021432114,0.09652899,0.26463798,0.04193398],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9995701,0.00006888569,0.000036996717,0.000070118636,0.00019703986,0.00005688961],"domain_scores_gemma":[0.9991233,0.00038859618,0.00009378653,0.0001353316,0.00016822499,0.000090894704],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015514806,0.0017938125,0.0010586557,0.002159776,0.0009549,0.0019005466,0.0041872375,0.0011958976,0.17113134],"category_scores_gemma":[0.003787233,0.0012131897,0.0012016447,0.0016530133,0.00054111885,0.0021162007,0.002297349,0.003043574,0.054071326],"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":[0.0003688754,0.00010444633,0.001007034,0.0011653481,0.00009890704,0.00048479871,0.000267639,0.004264663,0.010175491,0.012664466,0.8656281,0.10377014],"study_design_scores_gemma":[0.0003683319,0.0001187074,0.0014543788,0.0002429841,0.00006180182,0.0008953458,0.000097614946,0.045656707,0.042462654,0.01533309,0.8931142,0.0001942244],"about_ca_topic_score_codex":0.002642118,"about_ca_topic_score_gemma":0.0025538977,"teacher_disagreement_score":0.17113134,"about_ca_system_score_codex":0.00076854887,"about_ca_system_score_gemma":0.0019810041,"threshold_uncertainty_score":0.5724914},"labels":[],"label_agreement":null},{"id":"W2083474203","doi":"10.1186/1758-2946-6-s1-p43","title":"Exploiting solvent effects in drug design and optimization","year":2014,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Vertex Pharmaceuticals (Canada)","funders":"","keywords":"Ligand (biochemistry); Solvent; Computer science; Molecule; Computational chemistry; Binding energy; Chemistry; Computational science; Physics; Atomic physics; Organic chemistry","score_opus":0.012668857480028543,"score_gpt":0.2551218885689458,"score_spread":0.24245303108891725,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2083474203","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":[0.06285735,0.014263402,0.8973054,0.000999217,0.00022611323,0.00011472812,0.00019247604,0.0015507822,0.022490596],"genre_scores_gemma":[0.62545687,0.01558615,0.34853247,0.00043197488,0.00014416041,0.00032145728,0.00020570305,0.00043390717,0.008887354],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99973303,0.000111498266,0.000011215279,0.000028210592,0.00008927137,0.000026789176],"domain_scores_gemma":[0.99980265,0.00012045724,0.000024642093,0.000020680116,0.000021665073,0.000009820886],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007102888,0.0007225647,0.0008411322,0.0005475166,0.0003527613,0.0006702965,0.00047923764,0.0006267491,0.0017855248],"category_scores_gemma":[0.00096005696,0.00046786983,0.0006918573,0.0004634011,0.00067279505,0.00081334723,0.00087376434,0.00063486956,0.0005846457],"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":[0.00016389642,0.000047610232,0.0003819571,0.00033676476,0.00007137483,0.00010825137,0.000043552696,0.82623374,0.022588031,0.048144214,0.0013036095,0.10057706],"study_design_scores_gemma":[0.00007463637,0.00015621858,0.0001596503,0.000044064906,0.000036001857,0.00006196489,0.00002029631,0.9299905,0.013684908,0.042516332,0.013224561,0.000030930765],"about_ca_topic_score_codex":0.0011200688,"about_ca_topic_score_gemma":0.0012626312,"teacher_disagreement_score":0.0017855248,"about_ca_system_score_codex":0.000594742,"about_ca_system_score_gemma":0.0006793257,"threshold_uncertainty_score":0.0059731603},"labels":[],"label_agreement":null},{"id":"W2103126992","doi":"10.1186/s13321-015-0064-8","title":"Creation of a free, Internet-accessible database: the Multiple Target Ligand Database","year":2015,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Jewish General Hospital","funders":"Fundamental Research Funds for the Central Universities","keywords":"Computer science; The Internet; Database; World Wide Web","score_opus":0.04316942399164634,"score_gpt":0.3173300065991756,"score_spread":0.27416058260752924,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2103126992","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":[0.042954966,0.00941465,0.42972445,0.006488515,0.002536841,0.0032683213,0.37952307,0.09404105,0.03204819],"genre_scores_gemma":[0.045425612,0.0036252288,0.4655902,0.0011804078,0.00036688504,0.0016536197,0.47255015,0.0027220987,0.0068858066],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9983524,0.00030592957,0.00039437084,0.00029197827,0.0005527711,0.00010252333],"domain_scores_gemma":[0.995419,0.0012321486,0.00041505598,0.00079008174,0.0014142452,0.0007294912],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0035132975,0.0009921583,0.0019926243,0.0049751545,0.0008894565,0.0032414324,0.0031630218,0.0014573969,0.025701204],"category_scores_gemma":[0.008056128,0.0006389758,0.0010484533,0.004713792,0.0003970132,0.0031948998,0.0026849883,0.0023158102,0.020388382],"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":[0.0017912986,0.00070728286,0.007895094,0.0040411083,0.00032082375,0.0013230856,0.0002861308,0.0042622774,0.03266475,0.013993878,0.3706347,0.5620796],"study_design_scores_gemma":[0.0011735805,0.00047623523,0.004576563,0.0005279038,0.00025039894,0.0025989192,0.00018988058,0.018546203,0.03301496,0.0080214245,0.93042177,0.0002021754],"about_ca_topic_score_codex":0.0012437524,"about_ca_topic_score_gemma":0.0010070403,"teacher_disagreement_score":0.025701204,"about_ca_system_score_codex":0.00079808815,"about_ca_system_score_gemma":0.004407539,"threshold_uncertainty_score":0.085979044},"labels":[],"label_agreement":null},{"id":"W2110732896","doi":"10.1186/1758-2946-3-16","title":"Semantic Web integration of Cheminformatics resources with the SADI framework","year":2011,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":34,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University","funders":"Natural Sciences and Engineering Research Council of Canada; Canarie; Pfizer","keywords":"Computer science; Cheminformatics; Interoperability; Workflow; Semantic Web; Ontology; SPARQL; Web service; RDF; World Wide Web; Social Semantic Web; Data science; Database; Chemistry","score_opus":0.10075100365193504,"score_gpt":0.3231345185300068,"score_spread":0.22238351487807176,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2110732896","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":[0.007937906,0.00054717466,0.9520249,0.001088957,0.00012085717,0.00050686573,0.0017617246,0.024692383,0.011319399],"genre_scores_gemma":[0.082814395,0.0007472861,0.9008402,0.0005625351,0.000083165316,0.00043488765,0.010251894,0.0016779577,0.002587629],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9960496,0.00075432024,0.0005941642,0.0005163069,0.0018201479,0.00026554882],"domain_scores_gemma":[0.9970728,0.00078011124,0.00021217093,0.0010438559,0.0006235164,0.0002675554],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006080665,0.0011157171,0.0011535507,0.0037630713,0.0013626382,0.0052761585,0.0032570069,0.0013093669,0.0026104804],"category_scores_gemma":[0.0055329734,0.0007832686,0.002870484,0.00244065,0.002066167,0.005365782,0.0069023604,0.0024920828,0.001629561],"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":[0.0007757447,0.0009373568,0.0072779185,0.0022142632,0.0007714098,0.0024121667,0.002276524,0.08149595,0.029542131,0.5628203,0.039162844,0.27031335],"study_design_scores_gemma":[0.00013449004,0.00008887519,0.0011216864,0.00039635535,0.00024155212,0.00064947765,0.00036177848,0.39612433,0.03134412,0.23391168,0.3354614,0.00016419849],"about_ca_topic_score_codex":0.006918493,"about_ca_topic_score_gemma":0.0075040828,"teacher_disagreement_score":0.006918493,"about_ca_system_score_codex":0.0024579193,"about_ca_system_score_gemma":0.004170199,"threshold_uncertainty_score":0.032158077},"labels":[],"label_agreement":null},{"id":"W2110899240","doi":"10.1186/1758-2946-1-4","title":"A constructive approach for discovering new drug leads: Using a kernel methodology for the inverse-QSAR problem","year":2009,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":53,"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":"Quantitative structure–activity relationship; Kernel (algebra); Feature vector; Molecular descriptor; Computer science; Support vector machine; Kernel method; Feature (linguistics); Artificial intelligence; Inverse; Algorithm; Pattern recognition (psychology); Mathematics; Machine learning","score_opus":0.09032694935977614,"score_gpt":0.348785861803831,"score_spread":0.25845891244405483,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2110899240","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":[0.0070402343,0.00021802806,0.9911243,0.00019105601,0.000012398051,0.00007834891,0.00003072409,0.00021539941,0.001089498],"genre_scores_gemma":[0.16762924,0.00040249227,0.83004826,0.00020459188,0.000042531385,0.00026282322,0.00016943674,0.0001012897,0.001139401],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99768674,0.001017945,0.00013414155,0.00024573941,0.00082182745,0.00009357125],"domain_scores_gemma":[0.99123704,0.0064387163,0.0006093339,0.0007011498,0.0008873589,0.00012629504],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0038201748,0.0010284407,0.0012880414,0.0025436794,0.00049072126,0.0013156829,0.002150629,0.0015064556,0.0040567494],"category_scores_gemma":[0.0103313895,0.0006509459,0.001954382,0.0014728141,0.002235332,0.0026652797,0.0018230529,0.0017594938,0.0009269812],"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":[0.00014477668,0.00044103328,0.0020180645,0.00070765644,0.0001919286,0.000230138,0.00021271015,0.5895173,0.010827903,0.13341746,0.0022556183,0.2600354],"study_design_scores_gemma":[0.000025838224,0.00014449953,0.00016830812,0.000027548673,0.0000379117,0.00012429156,0.000032545886,0.9728981,0.003358706,0.021267978,0.001896161,0.000018188168],"about_ca_topic_score_codex":0.0010705818,"about_ca_topic_score_gemma":0.0011187722,"teacher_disagreement_score":0.0040567494,"about_ca_system_score_codex":0.0012084696,"about_ca_system_score_gemma":0.0021077844,"threshold_uncertainty_score":0.020203292},"labels":[],"label_agreement":null},{"id":"W2118054110","doi":"10.1186/1758-2946-2-8","title":"Basic primitives for molecular diagram sketching","year":2010,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Chemical Synthesis and Analysis","field":"Biochemistry, Genetics and Molecular Biology","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":"Quebec Labrador Foundation","funders":"","keywords":"Computer science; Software; Construct (python library); Set (abstract data type); Process (computing); Interface (matter); Theoretical computer science; Programming language; Parallel computing","score_opus":0.003713436710431373,"score_gpt":0.2402804689753334,"score_spread":0.23656703226490203,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2118054110","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":[0.00093292596,0.00023823221,0.98860663,0.00007170007,0.000055117296,0.00021018446,0.00060010154,0.005702816,0.003582396],"genre_scores_gemma":[0.019288616,0.0006785639,0.9740029,0.00004571463,0.0000404969,0.00046094292,0.001281895,0.0010224058,0.0031784915],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99855906,0.0001924422,0.00018063071,0.0001929643,0.00078053493,0.00009443064],"domain_scores_gemma":[0.9980623,0.0006984971,0.000103778875,0.00062017,0.00036115196,0.00015413889],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014182439,0.0017452659,0.0013644042,0.0026392706,0.0011600317,0.0041281986,0.0032217186,0.0014019789,0.02264195],"category_scores_gemma":[0.005546268,0.0012617445,0.0019614561,0.0024495486,0.0018135252,0.0038136204,0.00283428,0.003274677,0.013553127],"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":[0.0002840289,0.00010975386,0.0005120881,0.0014801074,0.000050209703,0.00036807582,0.00065695285,0.035119515,0.043641135,0.45109665,0.019851563,0.44682977],"study_design_scores_gemma":[0.00012591813,0.00017879218,0.00041996856,0.00034922332,0.000054553428,0.0010223555,0.00015171204,0.2106348,0.059958905,0.3430889,0.38382953,0.00018535371],"about_ca_topic_score_codex":0.00078776316,"about_ca_topic_score_gemma":0.0007547933,"teacher_disagreement_score":0.02264195,"about_ca_system_score_codex":0.00078804884,"about_ca_system_score_gemma":0.0012367212,"threshold_uncertainty_score":0.07574487},"labels":[],"label_agreement":null},{"id":"W2128119041","doi":"10.1186/1758-2946-4-9","title":"A chemical specialty semantic network for the Unified Medical Language System","year":2012,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"New York Institute of Technology","funders":"","keywords":"Unified Medical Language System; Computer science; Ontology; Semantics (computer science); Semantic network; Natural language processing; Information retrieval; Artificial intelligence; Data mining; Programming language","score_opus":0.012390122717380214,"score_gpt":0.2734242568030976,"score_spread":0.2610341340857174,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2128119041","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":[0.036695845,0.00046838325,0.92996615,0.0016274519,0.00008808133,0.00078158744,0.0061626886,0.0029531366,0.021256618],"genre_scores_gemma":[0.1411197,0.00031357925,0.8491815,0.00021951976,0.00003888755,0.0006213902,0.0068354285,0.0001600506,0.0015098507],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99756145,0.00087672117,0.00023163401,0.00043292268,0.00080447196,0.000092790615],"domain_scores_gemma":[0.9959741,0.002451974,0.00032678002,0.00034701728,0.0007852844,0.000114886636],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003177079,0.00039765125,0.00037574957,0.0045696977,0.0012657655,0.0021652777,0.000878936,0.0010391185,0.006542001],"category_scores_gemma":[0.013613797,0.00027134572,0.00097539544,0.0044894903,0.0010088529,0.0033806746,0.0016616953,0.0007101985,0.0011330697],"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":[0.00035413142,0.00013534905,0.00976141,0.0012080373,0.000119920165,0.00062703184,0.0019454266,0.119667366,0.006307434,0.5733807,0.021405853,0.2650873],"study_design_scores_gemma":[0.000064180895,0.00007517538,0.0047515435,0.00043034583,0.00016255608,0.00047879765,0.00068901543,0.5312783,0.005527392,0.33878276,0.117696814,0.00006316044],"about_ca_topic_score_codex":0.0090535525,"about_ca_topic_score_gemma":0.0137745105,"teacher_disagreement_score":0.0090535525,"about_ca_system_score_codex":0.003766058,"about_ca_system_score_gemma":0.003957933,"threshold_uncertainty_score":0.027324796},"labels":[],"label_agreement":null},{"id":"W2144342206","doi":"10.1186/s13321-015-0056-8","title":"Atom-Atom-Path similarity and Sphere Exclusion clustering: tools for prioritizing fragment hits","year":2015,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Discovery Centre","funders":"","keywords":"Fragment (logic); Cluster analysis; Similarity (geometry); Atom (system on chip); Path (computing); Computer science; Data mining; Information retrieval; Combinatorics; Algorithm; Artificial intelligence; Mathematics; Parallel computing","score_opus":0.06401385121419086,"score_gpt":0.3221815610132692,"score_spread":0.2581677097990784,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2144342206","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":[0.034158424,0.0005566043,0.93402654,0.0002356011,0.00005145159,0.00038064906,0.0023543702,0.024808379,0.0034279085],"genre_scores_gemma":[0.1137969,0.00033856436,0.87810606,0.000070211965,0.000026912881,0.0007149228,0.003403869,0.002087088,0.0014554523],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99835855,0.00036506035,0.00014894262,0.00024220634,0.000778961,0.00010632579],"domain_scores_gemma":[0.997156,0.0014565754,0.0004054415,0.00026294214,0.0005397026,0.00017929781],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024840331,0.0016623377,0.0013485723,0.0063146516,0.0010626151,0.0023228475,0.0021041518,0.001290032,0.0090970695],"category_scores_gemma":[0.0076794964,0.0006232439,0.0012705547,0.004553852,0.0007425765,0.0019722963,0.0024629284,0.0012242444,0.0027022902],"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":[0.0016511094,0.0004816511,0.016758826,0.0019281264,0.0005706917,0.000402013,0.0010548423,0.111509204,0.045986954,0.046699364,0.03431276,0.73864454],"study_design_scores_gemma":[0.00019343161,0.00020853002,0.0053003267,0.000074141884,0.000087450266,0.00037003355,0.00024192675,0.89121014,0.04297964,0.033508122,0.025670176,0.00015604272],"about_ca_topic_score_codex":0.0035342686,"about_ca_topic_score_gemma":0.0040911753,"teacher_disagreement_score":0.0090970695,"about_ca_system_score_codex":0.0013086705,"about_ca_system_score_gemma":0.0023079244,"threshold_uncertainty_score":0.030432701},"labels":[],"label_agreement":null},{"id":"W2145210171","doi":"10.1186/1758-2946-6-s1-o9","title":"Development of cannabinoid receptor (CB 2 R) ligands for application in PET studies - where to attach the radiolabel?","year":2014,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Pharmacological Receptor Mechanisms and Effects","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"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":"","keywords":"Cannabinoid receptor; Positron emission tomography; Receptor; Chemistry; Cannabinoid; High resolution; Ligand (biochemistry); Docking (animal); Carbazole; Stereochemistry; Biochemistry; Nuclear medicine; Medicine; Organic chemistry; Antagonist","score_opus":0.016062371692487018,"score_gpt":0.3144542415294359,"score_spread":0.2983918698369489,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2145210171","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":[0.38028783,0.13425699,0.44586676,0.0049995836,0.00072693126,0.0014581865,0.0076698153,0.0038377144,0.020896185],"genre_scores_gemma":[0.6655373,0.07166971,0.24474593,0.0013639738,0.0001464769,0.0008187293,0.0048628305,0.00025378464,0.010601235],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9997831,0.000060285794,0.000009819551,0.00004804281,0.000063918895,0.000034878765],"domain_scores_gemma":[0.99992037,0.00001832446,0.000018404127,0.000010632503,0.000020258542,0.000012085812],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00062846794,0.0009136465,0.0009990922,0.00047804316,0.00020269703,0.0009330835,0.0008203208,0.00089857785,0.0029729316],"category_scores_gemma":[0.00046782766,0.0004687805,0.00071480183,0.000713404,0.00027170999,0.0007724189,0.00041084795,0.0013122552,0.002244153],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00041062312,0.0003510605,0.0007745175,0.0015663016,0.00008914008,0.00052870566,0.00007111951,0.024320649,0.8723559,0.0041147284,0.003857144,0.091560155],"study_design_scores_gemma":[0.00027287714,0.0035956344,0.0023133564,0.0003191124,0.000349154,0.0015221112,0.00009958995,0.069820836,0.82012486,0.0029839089,0.0983909,0.00020775797],"about_ca_topic_score_codex":0.0016101795,"about_ca_topic_score_gemma":0.0025115053,"teacher_disagreement_score":0.0029729316,"about_ca_system_score_codex":0.0006949118,"about_ca_system_score_gemma":0.0005013238,"threshold_uncertainty_score":0.009945452},"labels":[],"label_agreement":null},{"id":"W2146172880","doi":"10.1186/1758-2946-4-5","title":"Blind trials of computer-assisted structure elucidation software","year":2012,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Molecular spectroscopy and chirality","field":"Chemistry","cited_by":29,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Systems, Applications & Products in Data Processing (Canada); Canadian Institute for Advanced Research","funders":"","keywords":"Software; Computer science; Set (abstract data type); Process (computing); Data mining; Data science; Information retrieval","score_opus":0.03362422348583312,"score_gpt":0.31109870900897996,"score_spread":0.27747448552314685,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2146172880","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":[0.6020009,0.033233233,0.1631042,0.0025709323,0.0053795082,0.17517905,0.0038703254,0.0029007362,0.011761138],"genre_scores_gemma":[0.7781281,0.0047017476,0.12024429,0.0015227126,0.00079741265,0.09006545,0.0012997425,0.0005073497,0.0027331535],"study_design_codex":"randomized_trial","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.7974472,0.15492268,0.024283577,0.0072835027,0.01423463,0.0018283318],"domain_scores_gemma":[0.47957346,0.38403636,0.04533188,0.040205892,0.047839258,0.0030131326],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.20900562,0.0024206145,0.0038633305,0.001997246,0.0015335184,0.0021128252,0.0026932368,0.0043109446,0.0061496776],"category_scores_gemma":[0.31353882,0.0012015427,0.004853277,0.0022642035,0.0034173378,0.0026431887,0.0021707832,0.002873614,0.0018403743],"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":[0.62851954,0.014902934,0.0068866024,0.019999698,0.00774786,0.00023870995,0.0022793007,0.006370594,0.010864705,0.0042761825,0.007587692,0.29032618],"study_design_scores_gemma":[0.28987628,0.56630456,0.0152716115,0.0063951258,0.009609679,0.00074743887,0.0009276761,0.016630381,0.040231004,0.010628194,0.04292115,0.0004569841],"about_ca_topic_score_codex":0.0006246,"about_ca_topic_score_gemma":0.00049290794,"teacher_disagreement_score":0.20900562,"about_ca_system_score_codex":0.0022377614,"about_ca_system_score_gemma":0.004501725,"threshold_uncertainty_score":0.97543687},"labels":[],"label_agreement":null},{"id":"W2146538435","doi":"10.1186/1758-2946-3-19","title":"Linked open drug data for pharmaceutical research and development","year":2011,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":187,"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":"Pfizer","keywords":"Data science; Computer science; Task (project management); Open data; Data sharing; External Data Representation; Web application; World Wide Web; Data mining; Medicine; Engineering; Alternative medicine; Artificial intelligence","score_opus":0.4477610663788747,"score_gpt":0.4690782952067872,"score_spread":0.021317228827912482,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2146538435","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":[0.009411902,0.009239428,0.48444423,0.02004854,0.0019681675,0.002484183,0.33370206,0.02403599,0.11466548],"genre_scores_gemma":[0.09087215,0.012795625,0.37291595,0.005327028,0.0008652868,0.0022846863,0.49768093,0.002509231,0.014749144],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98904175,0.003734181,0.0015141806,0.0010151854,0.004270122,0.00042452817],"domain_scores_gemma":[0.95214534,0.013097966,0.0050553414,0.021611976,0.0059074247,0.0021819267],"candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.012590253,0.0010700001,0.0013553817,0.013268317,0.001879292,0.009439941,0.0031559872,0.003020263,0.024467932],"category_scores_gemma":[0.058719076,0.0008190629,0.0020186326,0.021591099,0.0012624328,0.0098049315,0.009478928,0.0038760905,0.012049742],"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":[0.0007903516,0.00051335414,0.008604794,0.0039690943,0.00067814736,0.0005706637,0.00062872283,0.010027361,0.0028580218,0.28922445,0.29400378,0.3881313],"study_design_scores_gemma":[0.00012955547,0.000079524434,0.0026375027,0.00086578587,0.00015564087,0.00023950556,0.0002430727,0.007559355,0.00358627,0.13512146,0.84928924,0.00009315277],"about_ca_topic_score_codex":0.0041689593,"about_ca_topic_score_gemma":0.003732621,"teacher_disagreement_score":0.996844,"about_ca_system_score_codex":0.003116334,"about_ca_system_score_gemma":0.008772796,"threshold_uncertainty_score":0.08185345},"labels":[],"label_agreement":null},{"id":"W2155459813","doi":"10.1186/1758-2946-3-20","title":"Chemical Entity Semantic Specification: Knowledge representation for efficient semantic cheminformatics and facile data integration","year":2011,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":31,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Carleton University","funders":"","keywords":"Computer science; Cheminformatics; Representation (politics); Semantic Web; Semantics (computer science); Knowledge representation and reasoning; Chemical database; Domain (mathematical analysis); Data science; Information retrieval; Artificial intelligence; Chemistry; Programming language","score_opus":0.08371741914909708,"score_gpt":0.32162875154499426,"score_spread":0.23791133239589718,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2155459813","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":[0.004114557,0.000175286,0.98266083,0.000841108,0.000064910644,0.00024332498,0.0021934102,0.006443933,0.0032627236],"genre_scores_gemma":[0.06879234,0.00052514154,0.9180071,0.00044303964,0.000050596303,0.00037265825,0.009146486,0.0008438031,0.0018189223],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9967121,0.00088667445,0.0006799523,0.0004228123,0.0011351982,0.0001633036],"domain_scores_gemma":[0.9936539,0.0021333713,0.0004737099,0.0023123166,0.0012621371,0.00016456803],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0062705567,0.00093702134,0.00074895774,0.0028567335,0.000981076,0.005075602,0.0030984299,0.0016515023,0.00524621],"category_scores_gemma":[0.011672893,0.0007053723,0.0022710748,0.003955973,0.0020310122,0.0071408483,0.0040053576,0.002431142,0.0025371655],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00030055325,0.00024053457,0.004130525,0.0014343538,0.00018573338,0.000846136,0.0015600462,0.06975249,0.016720673,0.63564575,0.038727425,0.23045577],"study_design_scores_gemma":[0.00010570032,0.00009327885,0.0007472254,0.00062352454,0.00018026482,0.0006509949,0.0007341059,0.3690949,0.054142393,0.2795235,0.29397672,0.00012743448],"about_ca_topic_score_codex":0.006608024,"about_ca_topic_score_gemma":0.00836796,"teacher_disagreement_score":0.006608024,"about_ca_system_score_codex":0.0017235095,"about_ca_system_score_gemma":0.0047040014,"threshold_uncertainty_score":0.033162296},"labels":[],"label_agreement":null},{"id":"W2169151607","doi":"10.1186/1758-2946-2-s1-p35","title":"Prediction of highly-connected 'hub'-proteins in protein interaction networks using QSAR","year":2010,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Quantitative structure–activity relationship; Computer science; Interpretability; Cheminformatics; Computational biology; Machine learning; Proteome; Artificial intelligence; Drug discovery; Virtual screening; Data mining; Chemistry; Biology; Biochemistry; Computational chemistry","score_opus":0.027586025133478874,"score_gpt":0.2815810124379254,"score_spread":0.25399498730444653,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2169151607","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":[0.63219243,0.0009232852,0.36125913,0.00021447237,0.000027950044,0.00024521232,0.0024395045,0.0017219918,0.0009760786],"genre_scores_gemma":[0.9396476,0.0003382763,0.057637658,0.000037835896,0.00001425337,0.000117082054,0.0019374769,0.00003815131,0.000231689],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9997526,0.000069637645,0.000016664939,0.00006390737,0.000054429303,0.000042711334],"domain_scores_gemma":[0.99873227,0.0008753643,0.00017404895,0.000035044515,0.00013862981,0.000044695844],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009151302,0.0011990503,0.0010490123,0.0020144726,0.00036108197,0.000569391,0.00058606983,0.00082734996,0.0007623371],"category_scores_gemma":[0.0014678122,0.0003259236,0.0012113716,0.0010164934,0.00031459896,0.0005257369,0.00031529047,0.00079598965,0.00035969],"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":[0.000069122165,0.00012241412,0.011670254,0.00009240911,0.00006824186,0.00010723549,0.000025369307,0.9639942,0.0087103145,0.00039445728,0.00041616452,0.014329825],"study_design_scores_gemma":[0.0000018327048,0.000014737051,0.00084613403,0.0000018313351,0.0000057882157,0.000006363098,0.0000026333075,0.998109,0.0008180991,0.00015826245,0.00003353071,0.0000018142179],"about_ca_topic_score_codex":0.005845999,"about_ca_topic_score_gemma":0.0047319853,"teacher_disagreement_score":0.005845999,"about_ca_system_score_codex":0.0009768858,"about_ca_system_score_gemma":0.00057587435,"threshold_uncertainty_score":0.011623919},"labels":[],"label_agreement":null},{"id":"W2518754389","doi":"10.1186/s13321-016-0156-0","title":"Post-acquisition filtering of salt cluster artefacts for LC-MS based human metabolomic studies","year":2016,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":38,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University; Agriculture and Agri-Food Canada; Lawson Health Research Institute","funders":"Agriculture and Agri-Food Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Metabolomics; Metabolome; Mass spectrometry; Metabolite; Electrospray ionization; Chromatography; Chemistry; Liquid chromatography–mass spectrometry; Computer science; Analytical Chemistry (journal)","score_opus":0.018617142550887603,"score_gpt":0.29418090206411834,"score_spread":0.27556375951323075,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2518754389","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":[0.21645361,0.0015224147,0.769254,0.00037664035,0.0001942251,0.00044100537,0.0023997747,0.008229417,0.0011289382],"genre_scores_gemma":[0.3129727,0.0010788376,0.6751938,0.00045545006,0.00012271268,0.00074802124,0.005767961,0.0017845097,0.0018760881],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9982919,0.00032895769,0.00018172937,0.00044177548,0.00059588783,0.00015974455],"domain_scores_gemma":[0.99573463,0.001725879,0.0006316535,0.00076304376,0.0010083962,0.00013649579],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0051750983,0.001273631,0.0013078736,0.0019674983,0.0006912388,0.001485283,0.00095251336,0.001119447,0.0028723464],"category_scores_gemma":[0.010786277,0.000371988,0.0012539034,0.0016986212,0.0005639905,0.0009081171,0.0013645848,0.0011368337,0.0024404074],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0016148775,0.00027761995,0.018808754,0.00084144826,0.00036240264,0.001190568,0.00044578628,0.004844866,0.7277259,0.0013390179,0.0037425873,0.23880623],"study_design_scores_gemma":[0.000079135956,0.00069968007,0.059938684,0.00009682975,0.00039487687,0.002305007,0.0002261264,0.08191246,0.82372487,0.0036031716,0.026817933,0.00020127391],"about_ca_topic_score_codex":0.0011889278,"about_ca_topic_score_gemma":0.0017335502,"teacher_disagreement_score":0.0051750983,"about_ca_system_score_codex":0.0003225964,"about_ca_system_score_gemma":0.0013358009,"threshold_uncertainty_score":0.027368844},"labels":[],"label_agreement":null},{"id":"W2548357532","doi":"10.1186/s13321-016-0174-y","title":"ClassyFire: automated chemical classification with a comprehensive, computable taxonomy","year":2016,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":1587,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"The Metabolomics Innovation Centre; Alberta Innovates; Athabasca University; Ottawa Hospital; University of Ottawa; University of Alberta","funders":"Alberta Innovates; National Institutes of Health; Genome Alberta; U.S. National Library of Medicine; University of Alberta; Canadian Institutes of Health Research; Genome Canada","keywords":"Computer science; Taxonomy (biology); Class (philosophy); Druggability; Information retrieval; Chemical database; Artificial intelligence; Data science; Chemistry","score_opus":0.04246514513245019,"score_gpt":0.27671566847275636,"score_spread":0.23425052334030616,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2548357532","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":[0.009610136,0.0007409761,0.8567485,0.00068126153,0.00016847264,0.00077805517,0.016577395,0.10910814,0.005586981],"genre_scores_gemma":[0.021377059,0.00032866304,0.96137,0.00020124337,0.00003445581,0.0005909476,0.013574541,0.0012602403,0.0012627728],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99662846,0.00040538123,0.00033353732,0.0009256734,0.0015357941,0.00017119749],"domain_scores_gemma":[0.9928369,0.0036323022,0.00066512334,0.0012117694,0.0013273408,0.00032656195],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0034923623,0.0023972199,0.0018987436,0.007243982,0.0017595256,0.0041178004,0.004293725,0.0025315075,0.011044949],"category_scores_gemma":[0.014344235,0.0012122593,0.003697022,0.00519488,0.0014497001,0.0076576197,0.0037891727,0.0023885644,0.005406154],"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":[0.00074115914,0.00052500545,0.011306733,0.003418722,0.00034376464,0.0006459636,0.00072079565,0.08408399,0.013409278,0.06260638,0.14089492,0.68130326],"study_design_scores_gemma":[0.00030631677,0.00021287319,0.0023586466,0.00054112996,0.00011566245,0.00063204655,0.00029248066,0.6874703,0.02124349,0.11094115,0.17573157,0.00015424858],"about_ca_topic_score_codex":0.009956568,"about_ca_topic_score_gemma":0.013207433,"teacher_disagreement_score":0.011044949,"about_ca_system_score_codex":0.0033535555,"about_ca_system_score_gemma":0.007202392,"threshold_uncertainty_score":0.03694904},"labels":[],"label_agreement":null},{"id":"W2603853601","doi":"10.1186/s13321-017-0207-1","title":"Critical Assessment of Small Molecule Identification 2016: automated methods","year":2017,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":192,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Academy of Finland; Leibniz-Gemeinschaft; Deutsche Forschungsgemeinschaft; Metabolomics Innovation Centre; European Commission","keywords":"Computer science; Identification (biology); CONTEST; Machine learning; Annotation; Ranking (information retrieval); Artificial intelligence; Metadata; Data mining; Information retrieval; Data science; World Wide Web","score_opus":0.04345306440196599,"score_gpt":0.4468853515237422,"score_spread":0.4034322871217762,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2603853601","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":[0.034790795,0.0379897,0.6983398,0.06881592,0.013821026,0.0064201984,0.010603522,0.032243863,0.09697519],"genre_scores_gemma":[0.23445472,0.011135084,0.6538991,0.01152018,0.007273489,0.0057285437,0.023752363,0.009006516,0.043230083],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.89641666,0.037201844,0.005990778,0.0075709554,0.04959431,0.0032254641],"domain_scores_gemma":[0.77676153,0.06430675,0.0118509475,0.018266486,0.12075036,0.008063914],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.10408809,0.003026157,0.0021127688,0.00813144,0.0037711277,0.010208159,0.0058202134,0.0030832991,0.033612996],"category_scores_gemma":[0.16569352,0.0010488224,0.0019469081,0.0031417066,0.0036382608,0.006511166,0.01209165,0.004815624,0.023123244],"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":[0.0015146175,0.0003137675,0.0039033303,0.0055060275,0.0002652675,0.00029006822,0.0012899552,0.005927013,0.015913183,0.034001518,0.2890972,0.64197814],"study_design_scores_gemma":[0.00026568255,0.0008598958,0.0043172142,0.0023143308,0.0001643693,0.0006634179,0.0012035703,0.0410502,0.039789934,0.07090485,0.83813566,0.00033090624],"about_ca_topic_score_codex":0.0025947443,"about_ca_topic_score_gemma":0.0026643658,"teacher_disagreement_score":0.10408809,"about_ca_system_score_codex":0.007172266,"about_ca_system_score_gemma":0.014628305,"threshold_uncertainty_score":0.550477},"labels":[],"label_agreement":null},{"id":"W2605952223","doi":"10.1186/s13321-017-0209-z","title":"SimBoost: a read-across approach for predicting drug–target binding affinities using gradient boosting machines","year":2017,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":509,"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; Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Cheminformatics; Benchmark (surveying); Computer science; Binding affinities; Binary number; Boosting (machine learning); Drug discovery; Drug target; Affinities; Gradient boosting; Machine learning; Binary classification; Artificial intelligence; Data mining; Bioinformatics; Support vector machine; Random forest; Mathematics; Chemistry; Biology","score_opus":0.061090638483814026,"score_gpt":0.3506717947180562,"score_spread":0.2895811562342422,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2605952223","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":[0.04763115,0.0010915835,0.938666,0.0003050372,0.000143457,0.00024563007,0.0006147927,0.009577798,0.0017245956],"genre_scores_gemma":[0.4621108,0.0003753816,0.5295679,0.00057266344,0.00013408897,0.00043555224,0.0024933843,0.00071741425,0.0035928984],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9983747,0.0006462129,0.000079809884,0.00030011495,0.00044202685,0.00015710815],"domain_scores_gemma":[0.99828845,0.00081770564,0.00017035015,0.00025228396,0.0003526147,0.0001186182],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0043065874,0.0016534555,0.002681984,0.0021119958,0.0006306457,0.0010327086,0.0031468647,0.0017307979,0.0023576096],"category_scores_gemma":[0.004925207,0.00072395767,0.0012891012,0.0015623159,0.0006932671,0.0011091908,0.0013670914,0.0019393547,0.0011525205],"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":[0.0009180756,0.00067822146,0.0051360764,0.00033751395,0.0006233371,0.00013006678,0.00006554355,0.6391361,0.0064612436,0.0035597573,0.012681505,0.33027253],"study_design_scores_gemma":[0.000028108363,0.00007732849,0.00034817093,0.0000061833857,0.000018387369,0.000025129846,0.000006362218,0.99512744,0.0018900349,0.0017508978,0.00071311183,0.00000892099],"about_ca_topic_score_codex":0.0032721176,"about_ca_topic_score_gemma":0.0040340046,"teacher_disagreement_score":0.0043065874,"about_ca_system_score_codex":0.0010847414,"about_ca_system_score_gemma":0.0015807782,"threshold_uncertainty_score":0.02277565},"labels":[],"label_agreement":null},{"id":"W2727000434","doi":"10.1186/s13321-017-0242-y","title":"A posteriori metadata from automated provenance tracking: integration of AiiDA and TCOD","year":2017,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":30,"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":"Horizon 2020 Framework Programme; European Commission; SCIEX; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Metadata; Protocol (science); Interoperability; Interface (matter); Field (mathematics); Process (computing); Proof of concept","score_opus":0.022860132749665727,"score_gpt":0.305870177292973,"score_spread":0.28301004454330725,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2727000434","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":[0.0073676547,0.00054314645,0.92604893,0.001060076,0.00050170353,0.0010130255,0.00860171,0.04262595,0.012237737],"genre_scores_gemma":[0.08220992,0.0010575979,0.85803837,0.0006537475,0.00022113937,0.001405435,0.038988136,0.009086289,0.008339424],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9894203,0.0021471318,0.0016401062,0.0012849963,0.005128259,0.00037925923],"domain_scores_gemma":[0.93816954,0.011769895,0.0022319579,0.034690443,0.011788644,0.001349494],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.017573113,0.0012932732,0.0016582586,0.0073821954,0.0024249926,0.011309579,0.0035485632,0.0017687819,0.0067942617],"category_scores_gemma":[0.053601045,0.0013950398,0.0017625114,0.004698542,0.0024645398,0.012671445,0.011874855,0.004162663,0.005022801],"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":[0.0014858,0.00070784014,0.013706172,0.0026720501,0.00046205684,0.001247318,0.0041054315,0.02204206,0.025857078,0.320712,0.08436294,0.5226392],"study_design_scores_gemma":[0.00021949483,0.00015334341,0.0032958013,0.00090401893,0.00022922251,0.00072196405,0.00092367735,0.1720532,0.072342575,0.23603855,0.51264864,0.0004694833],"about_ca_topic_score_codex":0.009310615,"about_ca_topic_score_gemma":0.01077572,"teacher_disagreement_score":0.017573113,"about_ca_system_score_codex":0.0026217902,"about_ca_system_score_gemma":0.010464242,"threshold_uncertainty_score":0.092936635},"labels":[],"label_agreement":null},{"id":"W2745961286","doi":"10.1186/s13321-017-0234-y","title":"Comparative analysis of chemical similarity methods for modular natural products with a hypothetical structure enumeration algorithm","year":2017,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Microbial Natural Products and Biosynthesis","field":"Medicine","cited_by":63,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"MacEwan University; McMaster University","funders":"Canadian Institutes of Health Research; Canada Research Chairs","keywords":"Cheminformatics; Chemical space; Chemical similarity; Computer science; Natural product; Similarity (geometry); Enumeration; Modular design; Data mining; Drug discovery; Biochemical engineering; Algorithm; Structural similarity; Artificial intelligence; Bioinformatics; Chemistry; Mathematics; Biology; Engineering","score_opus":0.0263425672799847,"score_gpt":0.3490112789775224,"score_spread":0.3226687116975377,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2745961286","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":[0.38697255,0.0023658727,0.5973652,0.0004242213,0.00010895569,0.00034273593,0.00095304486,0.0027809918,0.008686428],"genre_scores_gemma":[0.39749166,0.0004244516,0.5978351,0.00017029369,0.000046441564,0.00025592739,0.0021252627,0.00025135837,0.0013994639],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99835134,0.0006771374,0.00014653345,0.00024971005,0.0004520559,0.00012332601],"domain_scores_gemma":[0.9870783,0.010038135,0.0004999448,0.0008317114,0.0013227505,0.00022910979],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003909474,0.00076300505,0.0009047696,0.0026363027,0.0006262022,0.001090964,0.001561961,0.0012628615,0.0027935847],"category_scores_gemma":[0.014910677,0.00022128997,0.0009424965,0.0020433734,0.0005622877,0.0018139284,0.001177131,0.00073648297,0.0004528137],"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":[0.0011241765,0.0006261516,0.010073941,0.00041992872,0.00019468936,0.000108990884,0.00012706261,0.5921496,0.00677487,0.025460107,0.0032445104,0.35969606],"study_design_scores_gemma":[0.00004006838,0.00017791745,0.00073001644,0.0000119666865,0.000016899761,0.00004558343,0.000029336408,0.98965,0.0023587614,0.006052182,0.00087767676,0.000009530044],"about_ca_topic_score_codex":0.0020934688,"about_ca_topic_score_gemma":0.002692166,"teacher_disagreement_score":0.003909474,"about_ca_system_score_codex":0.0013172351,"about_ca_system_score_gemma":0.0020435182,"threshold_uncertainty_score":0.02067548},"labels":[],"label_agreement":null},{"id":"W2905553692","doi":"10.1186/s13321-018-0319-2","title":"SIA: a scalable interoperable annotation server for biomedical named entities","year":2018,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"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":"Bundesministerium für Wirtschaft und Energie; Banting and Best Diabetes Centre, University of Toronto; Bundesministerium für Bildung und Forschung","keywords":"Annotation; Interoperability; Workflow; Scalability; Server; Information extraction; Named-entity recognition","score_opus":0.013462996530450582,"score_gpt":0.27915431022398707,"score_spread":0.2656913136935365,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2905553692","genre_codex":"software","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":"software","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.003671203,0.00077404076,0.3540823,0.0015075714,0.00047560327,0.0007816495,0.07350397,0.55823874,0.0069649257],"genre_scores_gemma":[0.03189932,0.0010794302,0.35832286,0.0013463984,0.00030271176,0.001666068,0.55846334,0.035667393,0.011252467],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9953324,0.00078091864,0.00079041324,0.0010293645,0.0017321702,0.00033461436],"domain_scores_gemma":[0.99025327,0.0022768828,0.001003957,0.003070474,0.0020371107,0.001358263],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0084144175,0.0032732126,0.0023334324,0.008837342,0.0022746224,0.005433147,0.005276086,0.0027500181,0.019135471],"category_scores_gemma":[0.017397791,0.0018836188,0.0026274784,0.008321518,0.0012548723,0.008022275,0.008962981,0.004550034,0.04316473],"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":[0.001973238,0.00054495584,0.005692155,0.0028678677,0.00074857683,0.0012118061,0.0011190607,0.004570455,0.047524355,0.02714977,0.7294292,0.17716864],"study_design_scores_gemma":[0.0006183413,0.00023789502,0.0076083546,0.00064868235,0.0003367625,0.001474851,0.00061223086,0.11746537,0.046129193,0.07848434,0.7457469,0.00063707307],"about_ca_topic_score_codex":0.0038503474,"about_ca_topic_score_gemma":0.0030804232,"teacher_disagreement_score":0.019135471,"about_ca_system_score_codex":0.0016750618,"about_ca_system_score_gemma":0.0058957757,"threshold_uncertainty_score":0.064014554},"labels":[],"label_agreement":null},{"id":"W2907470641","doi":"10.1186/s13321-018-0324-5","title":"BioTransformer: a comprehensive computational tool for small molecule metabolism prediction and metabolite identification","year":2019,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":513,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Eidgenössische Anstalt für Wasserversorgung Abwasserreinigung und Gewässerschutz; Genome Alberta; Agence Nationale de la Recherche; Alberta Innovates; Alberta Innovates - Health Solutions; Joint Programming Initiative A healthy diet for a healthy life; Canadian Institutes of Health Research; Genome Canada","keywords":"Computer science; Identification (biology); In silico; Computational biology; Metabolomics; Data mining; Machine learning; Bioinformatics; Chemistry; Biology; Biochemistry","score_opus":0.00830872921048072,"score_gpt":0.23225723172089133,"score_spread":0.2239485025104106,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2907470641","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":[0.06152808,0.0035328725,0.73360634,0.0011244874,0.00028048904,0.0006170503,0.03560005,0.15552232,0.008188278],"genre_scores_gemma":[0.21531704,0.003451796,0.7176613,0.00071793573,0.00013759562,0.002151752,0.048431046,0.008599698,0.0035318544],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9994295,0.00015757425,0.00006475734,0.00012178438,0.00018476347,0.000041479252],"domain_scores_gemma":[0.99854773,0.0009679861,0.00011306472,0.000121857396,0.00016229649,0.00008706131],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016064222,0.0025007627,0.0019537702,0.001639838,0.0005746701,0.0012261841,0.0024564047,0.001313485,0.008723734],"category_scores_gemma":[0.0035822901,0.0009489363,0.0023678637,0.0015235535,0.0004669631,0.0012231399,0.00163785,0.001604976,0.0026441517],"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":[0.0025107346,0.0007582209,0.01253211,0.005518966,0.002647784,0.0014187403,0.00032472715,0.5893566,0.028579276,0.01763078,0.089726955,0.24899508],"study_design_scores_gemma":[0.00033532962,0.00018523588,0.00090996403,0.00007010334,0.00017583367,0.00018972288,0.000019826388,0.96417034,0.0077484595,0.005085112,0.02104705,0.00006309424],"about_ca_topic_score_codex":0.0041377232,"about_ca_topic_score_gemma":0.006006012,"teacher_disagreement_score":0.008723734,"about_ca_system_score_codex":0.00064876484,"about_ca_system_score_gemma":0.0028860618,"threshold_uncertainty_score":0.029183805},"labels":[],"label_agreement":null},{"id":"W2963037859","doi":"10.1186/s13321-019-0349-4","title":"DNAmod: the DNA modification database","year":2019,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"RNA modifications and cancer","field":"Biochemistry, Genetics and Molecular Biology","cited_by":89,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Vector Institute; Canada Research Chairs; Princess Margaret Cancer Centre; Fleming College; University of Toronto","funders":"Canadian Institutes of Health Research; China Scholarship Council; Ontario Ministry of Research, Innovation and Science; Government of Ontario; University of Toronto; Canadian Cancer Society; Ontario Institute for Cancer Research; Natural Sciences and Engineering Research Council of Canada; Princess Margaret Cancer Foundation","keywords":"Computer science; Database; Computational biology; DNA; Biology; Genetics","score_opus":0.010571237936122725,"score_gpt":0.2545547639121071,"score_spread":0.24398352597598436,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2963037859","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":[0.0031366595,0.00721051,0.050280403,0.0006987322,0.0003408559,0.000488336,0.81729406,0.10409616,0.016454274],"genre_scores_gemma":[0.009301069,0.004061339,0.06710658,0.0009081854,0.000099756915,0.00076554116,0.9034499,0.00971306,0.0045946008],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9986486,0.00017882773,0.00031209327,0.00033522063,0.00043186994,0.00009330441],"domain_scores_gemma":[0.99829227,0.0005074718,0.00032796667,0.0003789178,0.0003098033,0.00018355309],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00212043,0.0025034419,0.002458966,0.0056360345,0.00088912045,0.0029485812,0.0045507103,0.0029351136,0.04368128],"category_scores_gemma":[0.006067653,0.0013958067,0.0015005084,0.0053358385,0.00050359016,0.003814907,0.0028661513,0.002264849,0.057874534],"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":[0.0017180375,0.00021819308,0.003239562,0.017092789,0.0004981731,0.0009354354,0.00031274694,0.0043552136,0.04069387,0.01477647,0.7451452,0.17101431],"study_design_scores_gemma":[0.00025908675,0.00010289909,0.0018163946,0.00045914724,0.00015071277,0.00058840326,0.00005825259,0.0026490348,0.015948035,0.009004751,0.9688525,0.00011076117],"about_ca_topic_score_codex":0.00167176,"about_ca_topic_score_gemma":0.002095785,"teacher_disagreement_score":0.04368128,"about_ca_system_score_codex":0.0009556236,"about_ca_system_score_gemma":0.002698407,"threshold_uncertainty_score":0.14612848},"labels":[],"label_agreement":null},{"id":"W2994270865","doi":"10.1186/s13321-019-0398-8","title":"The chemfp project","year":2019,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":55,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Discovery Air (Canada)","funders":"","keywords":"Computer science; x86; Benchmark (surveying); Cheminformatics; Data mining; Implementation; Software; Information retrieval; Parallel computing; Operating system; Programming language","score_opus":0.009351925075971965,"score_gpt":0.27854023117147897,"score_spread":0.269188306095507,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2994270865","genre_codex":"other","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":[0.010616523,0.0131958425,0.14839353,0.02564835,0.008609617,0.0012695115,0.1393464,0.15301853,0.49990168],"genre_scores_gemma":[0.05058012,0.009471916,0.19546212,0.010618929,0.0019871073,0.0023644497,0.37376833,0.052673828,0.30307323],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99276274,0.0015287483,0.00030296625,0.0012478203,0.0034066744,0.000750935],"domain_scores_gemma":[0.99120235,0.0016924731,0.00035159473,0.0017477231,0.00309279,0.0019130253],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009467661,0.002423127,0.001476949,0.0031338183,0.002103415,0.0069476175,0.0055659115,0.0032361015,0.22436665],"category_scores_gemma":[0.015327023,0.0008406778,0.0016101863,0.003457808,0.0012125757,0.007883887,0.0059478907,0.003531378,0.19960563],"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":[0.0004777384,0.00016904458,0.0006073038,0.00054878223,0.000034530553,0.0001280604,0.00009262198,0.0008299347,0.0022788753,0.03392616,0.8115398,0.1493672],"study_design_scores_gemma":[0.00011368532,0.00008694687,0.0002649691,0.00009383097,0.000011672013,0.00013921088,0.00004551988,0.0013522474,0.0019549723,0.0071903365,0.98872095,0.000025672565],"about_ca_topic_score_codex":0.0033767954,"about_ca_topic_score_gemma":0.0021268579,"teacher_disagreement_score":0.22436665,"about_ca_system_score_codex":0.0031985235,"about_ca_system_score_gemma":0.006870845,"threshold_uncertainty_score":0.7505812},"labels":[],"label_agreement":null},{"id":"W3044334750","doi":"10.1186/s13321-020-00450-7","title":"A review of computational drug repositioning: strategies, approaches, opportunities, challenges, and directions","year":2020,"lang":"en","type":"review","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":391,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"","keywords":"Drug repositioning; Computer science; Drug discovery; Process (computing); Data science; Multidisciplinary approach; Drug development; Computational model; Drug; Field (mathematics); Risk analysis (engineering); Artificial intelligence; Bioinformatics; Medicine; Pharmacology","score_opus":0.2181327886717617,"score_gpt":0.3515936227001733,"score_spread":0.1334608340284116,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3044334750","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":[0.00019717142,0.9943786,0.00263495,0.000629784,0.00023782573,0.000014871013,0.00005235715,0.000056004672,0.0017984604],"genre_scores_gemma":[0.0010328918,0.994995,0.003001634,0.00024277138,0.0002182151,0.00001909838,0.00007473817,0.000013921704,0.00040170783],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99949694,0.00011295019,0.000068714915,0.00008213579,0.00020163815,0.000037671016],"domain_scores_gemma":[0.99804294,0.0013083434,0.0001250689,0.00006948192,0.000362088,0.000092142036],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013036837,0.0013659206,0.0014324964,0.0037962582,0.0004129284,0.0015673566,0.0016766374,0.0013351681,0.004367395],"category_scores_gemma":[0.002712394,0.0006883417,0.0009905773,0.006268958,0.00071543123,0.0026037276,0.0010117322,0.002442312,0.0025424915],"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":[0.00005462435,0.00011215524,0.0002660459,0.018438604,0.00013620315,0.00007865921,0.00006871891,0.003386151,0.00084659463,0.013663417,0.03671928,0.92622966],"study_design_scores_gemma":[0.00002556703,0.00015561907,0.00064039056,0.0075362744,0.00019609588,0.00072129467,0.00007746255,0.0032423441,0.0010153805,0.009843831,0.9764806,0.00006526361],"about_ca_topic_score_codex":0.0024438517,"about_ca_topic_score_gemma":0.0029261934,"teacher_disagreement_score":0.004367395,"about_ca_system_score_codex":0.0009917741,"about_ca_system_score_gemma":0.0023571788,"threshold_uncertainty_score":0.01461035},"labels":[],"label_agreement":null},{"id":"W3045213199","doi":"10.1186/s13321-020-00449-0","title":"rMSIcleanup: an open-source tool for matrix-related peak annotation in mass spectrometry imaging and its application to silver-assisted laser desorption/ionization","year":2020,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"Agència de Gestió d'Ajuts Universitaris i de Recerca; Natural Sciences and Engineering Research Council of Canada; Universitat Rovira i Virgili; Horizon 2020; Ministerio de Economía y Competitividad","keywords":"Mass spectrometry imaging; Mass spectrometry; Computer science; Annotation; Workflow; Matrix-assisted laser desorption/ionization; Matrix (chemical analysis); Principal component analysis; Desorption; Analyte; MALDI imaging; Chemistry; Analytical Chemistry (journal); Data mining; Artificial intelligence; Chromatography; Database","score_opus":0.009794563412108884,"score_gpt":0.26528218273723764,"score_spread":0.25548761932512876,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3045213199","genre_codex":"software","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":"software","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005874652,0.0008763529,0.30872148,0.00028309273,0.00021084242,0.00034484628,0.025865115,0.65524644,0.0025771498],"genre_scores_gemma":[0.042715773,0.001383797,0.65260786,0.00089039124,0.00013052422,0.0022878994,0.07466067,0.21763222,0.007690897],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9981317,0.00038021,0.0002075382,0.0005582455,0.0005400579,0.00018227832],"domain_scores_gemma":[0.9964461,0.0018885332,0.000505559,0.00047325893,0.00050494436,0.0001815872],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003894007,0.004443344,0.0020612648,0.0028641734,0.0010737573,0.0032225947,0.004045506,0.0015819006,0.028833503],"category_scores_gemma":[0.010634013,0.0020183711,0.0031318197,0.0018658335,0.0009170513,0.0024103227,0.004074712,0.0032029627,0.02852589],"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":[0.0034839702,0.00059560145,0.012266592,0.009082064,0.002996105,0.0024394752,0.0017884666,0.017489064,0.09699379,0.011113863,0.53514016,0.30661088],"study_design_scores_gemma":[0.0007808326,0.0005033896,0.014127106,0.0009567444,0.00061086466,0.002435983,0.00037775704,0.18778662,0.17021854,0.021491172,0.59980446,0.00090649066],"about_ca_topic_score_codex":0.0019139051,"about_ca_topic_score_gemma":0.0030821506,"teacher_disagreement_score":0.028833503,"about_ca_system_score_codex":0.0007378893,"about_ca_system_score_gemma":0.002618063,"threshold_uncertainty_score":0.09645772},"labels":[],"label_agreement":null},{"id":"W3155267227","doi":"10.1186/s13321-021-00527-x","title":"Twitter integration of chemistry software tools","year":2021,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Scientific Computing and Data Management","field":"Decision Sciences","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":"University of Toronto","funders":"Japan Society for the Promotion of Science","keywords":"Cheminformatics; Computer science; Software; Data science; World Wide Web; Social media; Social software; Measure (data warehouse); Chemistry; Data mining","score_opus":0.13531568106620018,"score_gpt":0.36827355986071786,"score_spread":0.23295787879451768,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3155267227","genre_codex":"empirical","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":[0.3136229,0.002365385,0.1452808,0.01151763,0.0022499075,0.0023491785,0.0588554,0.20242402,0.2613348],"genre_scores_gemma":[0.68607765,0.0015827687,0.14856277,0.0028007957,0.0012727836,0.0023257886,0.07251008,0.015633304,0.06923404],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9970304,0.0006542548,0.0003665325,0.00045637708,0.0012118614,0.00028057766],"domain_scores_gemma":[0.9852916,0.007237022,0.0015898962,0.0028457695,0.0020943363,0.00094137515],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0028236182,0.0011893326,0.00073547335,0.007168892,0.0016386947,0.0027512398,0.0011723952,0.001113975,0.0145038655],"category_scores_gemma":[0.017982496,0.0005653606,0.00083631783,0.0045429342,0.00053178484,0.005265488,0.0043488853,0.0011484469,0.0075333887],"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":[0.002422163,0.0006873701,0.06148248,0.002452342,0.00041643,0.0030023595,0.008098369,0.0063221254,0.037046727,0.024645302,0.25280347,0.60062087],"study_design_scores_gemma":[0.0002430643,0.0005006146,0.030970708,0.0004769514,0.00028103872,0.0010675256,0.0023485585,0.09204575,0.033097852,0.023009028,0.81564337,0.0003156391],"about_ca_topic_score_codex":0.0031301351,"about_ca_topic_score_gemma":0.005125696,"teacher_disagreement_score":0.0145038655,"about_ca_system_score_codex":0.000967425,"about_ca_system_score_gemma":0.0012846342,"threshold_uncertainty_score":0.048520267},"labels":[],"label_agreement":null},{"id":"W4220847490","doi":"10.1186/s13321-022-00596-6","title":"Deep learning-driven prediction of drug mechanism of action from large-scale chemical-genetic interaction profiles","year":2022,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":23,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Manitoba","funders":"Canadian Institutes of Health Research; Cystic Fibrosis Canada","keywords":"Computer science; Mechanism (biology); Scale (ratio); Action (physics); Drug discovery; Artificial intelligence; Machine learning; Data science; Bioinformatics; Biology","score_opus":0.015299762500538215,"score_gpt":0.26690328197015045,"score_spread":0.25160351946961224,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4220847490","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":[0.7496743,0.0031406372,0.23484907,0.0015201605,0.00010712207,0.00014344923,0.0043093995,0.0036914004,0.002564492],"genre_scores_gemma":[0.96755946,0.0003511301,0.027305283,0.00026182763,0.000031194246,0.000080951926,0.0031320413,0.000039760467,0.0012384517],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99981624,0.000038119684,0.000011005399,0.00005970436,0.00003346148,0.000041510106],"domain_scores_gemma":[0.99914813,0.00046185285,0.00012450901,0.000048030757,0.00014466484,0.00007284077],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006726496,0.0010342785,0.000952945,0.00080994965,0.0001977197,0.0005237512,0.0012183291,0.0013998996,0.0009972559],"category_scores_gemma":[0.0015967382,0.00035593877,0.00084423,0.00053046254,0.00041915596,0.00068678835,0.00044077556,0.0013985743,0.00027857276],"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":[0.00016556721,0.00018960914,0.0027957885,0.0000871445,0.000051549832,0.000045050812,0.00000592169,0.97391564,0.00301565,0.0005318389,0.0008409206,0.01835534],"study_design_scores_gemma":[0.0000044943,0.000016772643,0.00013841194,0.0000014483205,0.0000040617524,0.000002458923,8.674667e-7,0.9986607,0.0007849449,0.00033908876,0.000045483986,0.0000013407581],"about_ca_topic_score_codex":0.0130666895,"about_ca_topic_score_gemma":0.0128138475,"teacher_disagreement_score":0.0130666895,"about_ca_system_score_codex":0.0017739428,"about_ca_system_score_gemma":0.0014231537,"threshold_uncertainty_score":0.025981247},"labels":[],"label_agreement":null},{"id":"W4322102662","doi":"10.1186/s13321-023-00698-9","title":"ABT-MPNN: an atom-bond transformer-based message-passing neural network for molecular property prediction","year":2023,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":51,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University; University of Manitoba","funders":"Canadian Institutes of Health Research; Canada Research Chairs; Cystic Fibrosis Canada; Manitoba Medical Service Foundation","keywords":"Computer science; Interpretability; Message passing; Molecular graph; Theoretical computer science; Embedding; Artificial neural network; Graph embedding; Artificial intelligence; Graph; Distributed computing","score_opus":0.02954166971114737,"score_gpt":0.3017053752043101,"score_spread":0.27216370549316277,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4322102662","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":[0.04949101,0.001250818,0.93877375,0.0005453817,0.00020820451,0.00012676169,0.0006163827,0.0064332583,0.0025543603],"genre_scores_gemma":[0.6499392,0.0011855249,0.33552742,0.00064181536,0.00013800185,0.00032810261,0.0031753655,0.00036901212,0.008695533],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99976,0.000039343016,0.000014071972,0.00007906458,0.00007867038,0.000028860048],"domain_scores_gemma":[0.99961495,0.00013924876,0.000054499666,0.000060330636,0.00010440037,0.000026587373],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00048013814,0.0011666779,0.00076232513,0.0007213256,0.00033142994,0.00057335483,0.002552818,0.0012224568,0.001952693],"category_scores_gemma":[0.0019114364,0.00042678235,0.00090735126,0.0008013174,0.00048925,0.0016424053,0.00084979896,0.001776169,0.00064761605],"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":[0.00017574173,0.00018154619,0.0016025145,0.00023556314,0.0001747015,0.00017170164,0.00006910155,0.6438874,0.013883827,0.008581399,0.0067058904,0.32433066],"study_design_scores_gemma":[0.0000034584168,0.000015261754,0.00007001111,0.000002721988,0.000008627819,0.000010769737,0.0000017339372,0.9963236,0.0014577458,0.0017229469,0.00038003863,0.0000031197985],"about_ca_topic_score_codex":0.009427505,"about_ca_topic_score_gemma":0.010682403,"teacher_disagreement_score":0.009427505,"about_ca_system_score_codex":0.0010309033,"about_ca_system_score_gemma":0.0010898,"threshold_uncertainty_score":0.018745303},"labels":[],"label_agreement":null},{"id":"W4399175963","doi":"10.1186/s13321-024-00853-w","title":"Identifying uncertainty in physical–chemical property estimation with IFSQSAR","year":2024,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; The Scarborough Hospital; ARC Resources (Canada)","funders":"","keywords":"Property (philosophy); Computer science; Estimation; Data mining; Data science; Systems engineering; Engineering","score_opus":0.009905848516422483,"score_gpt":0.2654185823120778,"score_spread":0.2555127337956553,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399175963","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":[0.21966341,0.00095137995,0.7670004,0.00033287337,0.000040505318,0.0001803196,0.0023285227,0.006965063,0.0025375856],"genre_scores_gemma":[0.8749078,0.00025425898,0.11916627,0.00019551034,0.000034386067,0.00016874453,0.004521817,0.00027380435,0.00047733067],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9963117,0.0013600981,0.00021376634,0.0006559122,0.0012800136,0.00017850866],"domain_scores_gemma":[0.9812362,0.014103399,0.0014211762,0.0016426814,0.0014696028,0.00012690136],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.013433164,0.0014680933,0.001329788,0.0022750413,0.00037561168,0.0016806497,0.001714321,0.0007689668,0.001306455],"category_scores_gemma":[0.019284181,0.000542163,0.0022852458,0.0012057159,0.00091587595,0.001768669,0.0017957123,0.0014566475,0.0005595834],"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":[0.0003339593,0.00016261292,0.019532219,0.0002424,0.00037943458,0.000106515065,0.000101194426,0.9207675,0.0046199365,0.0023869576,0.001087075,0.050280225],"study_design_scores_gemma":[0.00001003876,0.0000785308,0.0016917952,0.000011952788,0.000024552839,0.000034212197,0.000016018106,0.99308807,0.002974283,0.0015861624,0.00046953856,0.000014813167],"about_ca_topic_score_codex":0.006172399,"about_ca_topic_score_gemma":0.0031134572,"teacher_disagreement_score":0.013433164,"about_ca_system_score_codex":0.0010753163,"about_ca_system_score_gemma":0.0013769652,"threshold_uncertainty_score":0.07104218},"labels":[],"label_agreement":null},{"id":"W4401589835","doi":"10.1186/s13321-024-00892-3","title":"Metis: a python-based user interface to collect expert feedback for generative chemistry models","year":2024,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":7,"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":"Horizon 2020; Engineering and Physical Sciences Research Council; Vetenskapsrådet; UK Research and Innovation; Knut och Alice Wallenbergs Stiftelse; Finnish Center for Artificial Intelligence; European Commission; Chalmers Tekniska Högskola","keywords":"Computer science; Python (programming language); Graphical user interface; Human–computer interaction; Interface (matter); User interface; Generative grammar; Generative model; Human-in-the-loop; Software engineering; Data science; Artificial intelligence; Programming language","score_opus":0.030714670423141927,"score_gpt":0.33359502077316355,"score_spread":0.30288035035002164,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4401589835","genre_codex":"software","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":[0.0032057832,0.00018124968,0.3330146,0.00045863038,0.00015820959,0.00069916144,0.025886193,0.62988365,0.0065125246],"genre_scores_gemma":[0.106763735,0.0006864042,0.5781303,0.0027252506,0.00022384933,0.007787487,0.07195832,0.20801239,0.023712296],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9986173,0.00037592577,0.00013790591,0.0002914697,0.00042315637,0.00015423975],"domain_scores_gemma":[0.9949297,0.003119068,0.00023184174,0.00055469695,0.0007590591,0.00040570094],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0031621906,0.0031544229,0.0013852972,0.001353907,0.00055371254,0.0020139473,0.004170636,0.0015874526,0.11348994],"category_scores_gemma":[0.012463156,0.0011825626,0.0019091709,0.00068704237,0.0009201347,0.002972328,0.0051488867,0.002907859,0.035811774],"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":[0.003407503,0.0006405517,0.0055377213,0.005023065,0.00038980835,0.0013923856,0.0018551601,0.02233345,0.020660387,0.021370826,0.69745815,0.21993102],"study_design_scores_gemma":[0.0018740847,0.00038075537,0.0063837697,0.00076486863,0.000113800605,0.00087976287,0.0002520177,0.34369949,0.04144622,0.06292133,0.54073673,0.0005472184],"about_ca_topic_score_codex":0.0018879746,"about_ca_topic_score_gemma":0.003037669,"teacher_disagreement_score":0.11348994,"about_ca_system_score_codex":0.0010153769,"about_ca_system_score_gemma":0.0020927468,"threshold_uncertainty_score":0.37966168},"labels":[],"label_agreement":null},{"id":"W4403833137","doi":"10.1186/s13321-024-00911-3","title":"Towards the prediction of drug solubility in binary solvent mixtures at various temperatures using machine learning","year":2024,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium; Canadian Institute for Advanced Research; Vector Institute; University of Toronto","funders":"Leslie Dan Faculty of Pharmacy, University of Toronto; Natural Sciences and Engineering Research Council of Canada; Advanced Research Projects Agency; Defense Advanced Research Projects Agency; Canada First Research Excellence Fund; University of Toronto","keywords":"Solubility; Machine learning; Computer science; Solvent; Binary number; Gradient boosting; Hildebrand solubility parameter; Artificial intelligence; Biochemical engineering; Materials science; Chemistry; Mathematics; Organic chemistry; Random forest","score_opus":0.023630155489136693,"score_gpt":0.28875265640825043,"score_spread":0.2651225009191137,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403833137","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":[0.5188651,0.0069395374,0.41443217,0.002714504,0.00064645574,0.00040938603,0.029373644,0.0210055,0.005613769],"genre_scores_gemma":[0.6893915,0.0018859948,0.26663977,0.00081705995,0.00019809618,0.00047959446,0.036782395,0.0008540926,0.0029515019],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9992692,0.00018297818,0.000049030416,0.00025347449,0.00017268883,0.00007254461],"domain_scores_gemma":[0.9982168,0.00086687657,0.00019559142,0.00020106307,0.0004177339,0.00010186242],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0018432593,0.0017172571,0.0014309325,0.0013943287,0.0005812811,0.0015206821,0.0013988707,0.0018750927,0.0016597363],"category_scores_gemma":[0.0055661467,0.00055029185,0.0022196167,0.0011997313,0.00063780765,0.0020262857,0.00092755945,0.0026008945,0.0019642801],"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":[0.00075720047,0.00090450526,0.019464219,0.0010656143,0.00028823104,0.00026953072,0.00011379631,0.8213681,0.031268183,0.0030835422,0.021916267,0.09950076],"study_design_scores_gemma":[0.000034136512,0.00007636083,0.0011290584,0.000027029866,0.000023661822,0.00003199473,0.00001601998,0.9806395,0.013479343,0.0020131508,0.0024998733,0.000029895033],"about_ca_topic_score_codex":0.009693057,"about_ca_topic_score_gemma":0.010089525,"teacher_disagreement_score":0.009693057,"about_ca_system_score_codex":0.0015822991,"about_ca_system_score_gemma":0.0020324162,"threshold_uncertainty_score":0.019273281},"labels":[],"label_agreement":null},{"id":"W4407140906","doi":"10.1186/s13321-025-00962-0","title":"Improving drug repositioning with negative data labeling using large language models","year":2025,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":8,"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":"Computer science; Drug repositioning; Drug; Data mining; Natural language processing; Data science; Medicine; Pharmacology","score_opus":0.024940952440858137,"score_gpt":0.32035736894867095,"score_spread":0.2954164165078128,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4407140906","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":[0.25529286,0.0016801372,0.7263515,0.0016772235,0.00028003566,0.00040217637,0.0026894002,0.007242258,0.0043844674],"genre_scores_gemma":[0.7375065,0.0004246054,0.2527961,0.00077427627,0.00015493031,0.00037893528,0.00544268,0.00023593003,0.0022859958],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99797195,0.0009039337,0.0001654996,0.00046634296,0.00037449022,0.00011776724],"domain_scores_gemma":[0.9945885,0.0031456586,0.00061529916,0.0006178606,0.0009001631,0.00013251511],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003608568,0.0015208009,0.0010388859,0.0014030554,0.0005039868,0.0012057504,0.001174429,0.00095970347,0.0010973618],"category_scores_gemma":[0.008206487,0.0003816141,0.001967724,0.0007350382,0.0005263589,0.0016061179,0.0009985799,0.0015084156,0.0007356829],"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":[0.0011379417,0.0014012823,0.02477352,0.0005875696,0.0004709648,0.00061060063,0.00026014601,0.4324687,0.031526253,0.0038700902,0.012081505,0.4908114],"study_design_scores_gemma":[0.000029998891,0.00016167693,0.00093707396,0.000020480209,0.00007186148,0.00006709799,0.000028945504,0.9861611,0.008125213,0.0028053566,0.0015625684,0.000028605522],"about_ca_topic_score_codex":0.0053075105,"about_ca_topic_score_gemma":0.0079906285,"teacher_disagreement_score":0.0053075105,"about_ca_system_score_codex":0.0010170136,"about_ca_system_score_gemma":0.0020271442,"threshold_uncertainty_score":0.019084156},"labels":[],"label_agreement":null},{"id":"W4408693799","doi":"10.1186/s13321-025-00979-5","title":"An interpretable deep geometric learning model to predict the effects of mutations on protein–protein interactions using large-scale protein language model","year":2025,"lang":"en","type":"article","venue":"Journal of Cheminformatics","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":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Children’s Health Research Institute; Lawson Health Research Institute; University of Manitoba; Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Computer science; Scale (ratio); Artificial intelligence; Machine learning","score_opus":0.004592795866963382,"score_gpt":0.28068493732233607,"score_spread":0.2760921414553727,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408693799","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":[0.07520451,0.0007625433,0.9179806,0.0007376602,0.000087908964,0.00006092679,0.00071654463,0.002233264,0.0022159414],"genre_scores_gemma":[0.8665784,0.0005712962,0.123071,0.00052123034,0.00007003035,0.0002558508,0.001958931,0.00017979684,0.0067935335],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99983585,0.000029006525,0.000008902866,0.000059039663,0.000037880236,0.000029234528],"domain_scores_gemma":[0.9996455,0.00017276967,0.000051038238,0.000024588644,0.00008354205,0.000022612125],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004254997,0.00096811337,0.0007388463,0.0007005973,0.00022912868,0.00060575403,0.001589156,0.0011805951,0.0017089808],"category_scores_gemma":[0.0012162408,0.00044713018,0.0010231722,0.00054093293,0.0005389343,0.0009442245,0.0007101565,0.0014509865,0.0005559651],"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":[0.00004434878,0.000039067516,0.0007655178,0.000022557539,0.00002850333,0.00007057441,0.000014270358,0.96965003,0.0019058227,0.0026436837,0.00095363264,0.023862055],"study_design_scores_gemma":[0.0000014506963,0.000003404622,0.000027517555,6.1835425e-7,0.00000144664,0.0000037842178,7.0160723e-7,0.99917954,0.00011438879,0.00061735435,0.000048717342,0.0000010025659],"about_ca_topic_score_codex":0.011485747,"about_ca_topic_score_gemma":0.012544046,"teacher_disagreement_score":0.011485747,"about_ca_system_score_codex":0.0011253464,"about_ca_system_score_gemma":0.0009944319,"threshold_uncertainty_score":0.022837818},"labels":[],"label_agreement":null},{"id":"W4411371888","doi":"10.1186/s13321-025-00958-w","title":"Crossover operators for molecular graphs with an application to virtual drug screening","year":2025,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Artificial Intelligence in Medicine (Canada)","funders":"National Institutes of Health; German Network for Bioinformatics Infrastructure; Novo Nordisk Fonden; Universität Leipzig; Bundesministerium für Bildung und Forschung; Deutsche Forschungsgemeinschaft; Max Kade Foundation; Alexander von Humboldt-Stiftung","keywords":"Crossover; Computer science; Theoretical computer science; Context (archaeology); Algorithm; Artificial intelligence","score_opus":0.007400817636468401,"score_gpt":0.2996973154753661,"score_spread":0.2922964978388977,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4411371888","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":[0.01997151,0.0015628845,0.971958,0.00037062194,0.00011345644,0.00013786233,0.0001683984,0.0007718625,0.004945388],"genre_scores_gemma":[0.31599957,0.002045279,0.67536366,0.00034229463,0.00014789143,0.00040082447,0.00045889872,0.00048375523,0.00475785],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990332,0.00032308808,0.000053659514,0.00013943382,0.00038456198,0.00006602254],"domain_scores_gemma":[0.99853635,0.0008872157,0.00017351187,0.00019507032,0.00015363887,0.00005419687],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013534874,0.00074116536,0.0006225179,0.0015121044,0.0006118241,0.0011945405,0.001071545,0.0009883676,0.0023915898],"category_scores_gemma":[0.0033617348,0.00024231923,0.0011415009,0.0015445035,0.0017380152,0.0010311401,0.0012983346,0.0018560005,0.00035498856],"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":[0.00014246383,0.00012025568,0.00092009816,0.00028758045,0.00007153891,0.0003794424,0.00025064286,0.324794,0.028111337,0.48606887,0.0039084465,0.15494533],"study_design_scores_gemma":[0.00007845647,0.00020583734,0.00081123883,0.000095305884,0.00006288158,0.00049809675,0.000064303444,0.649646,0.015145162,0.29961407,0.033711053,0.000067655375],"about_ca_topic_score_codex":0.0010747885,"about_ca_topic_score_gemma":0.0009103245,"teacher_disagreement_score":0.0023915898,"about_ca_system_score_codex":0.0014235488,"about_ca_system_score_gemma":0.00059875974,"threshold_uncertainty_score":0.0103286505},"labels":[],"label_agreement":null},{"id":"W4412435986","doi":"10.1186/s13321-025-01038-9","title":"SuperMetal: a generative AI framework for rapid and precise metal ion location prediction in proteins","year":2025,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Vanderbilt University","keywords":"Metal ions in aqueous solution; Generative grammar; Computer science; Ion; Generative model; Metal; Zinc; Chemistry; Artificial intelligence","score_opus":0.009415692670339681,"score_gpt":0.2635056148016386,"score_spread":0.2540899221312989,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4412435986","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":[0.010051559,0.00023965572,0.98485374,0.00025447411,0.000024480412,0.000039350303,0.00027773733,0.0029010738,0.0013579637],"genre_scores_gemma":[0.43880585,0.00036874355,0.5540903,0.0006405313,0.000106070605,0.00031072384,0.0014012535,0.00060799654,0.0036685315],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9995383,0.00014065467,0.00001839441,0.000115615425,0.00014164606,0.000045413384],"domain_scores_gemma":[0.9989146,0.00076453434,0.0000617756,0.00007954824,0.00011915646,0.00006028371],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010399087,0.00086435204,0.0009856978,0.0009374272,0.00043631133,0.0010989989,0.0034248335,0.0012696069,0.00407987],"category_scores_gemma":[0.0027750772,0.00061158417,0.0013190166,0.0007569264,0.001052119,0.0010327395,0.0016132179,0.0016207263,0.0008460322],"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":[0.00010368891,0.000048464073,0.0010091359,0.000114928705,0.000073840274,0.00014117529,0.00008059267,0.90718627,0.0040649083,0.028609967,0.0034322597,0.055134863],"study_design_scores_gemma":[0.0000029874318,0.000004025699,0.000017080534,0.000002055569,0.0000019192214,0.000007203038,0.0000018833626,0.9946126,0.00024328522,0.0048653884,0.00023965484,0.000001997539],"about_ca_topic_score_codex":0.008338928,"about_ca_topic_score_gemma":0.011646033,"teacher_disagreement_score":0.008338928,"about_ca_system_score_codex":0.001280229,"about_ca_system_score_gemma":0.0011706799,"threshold_uncertainty_score":0.01658076},"labels":[],"label_agreement":null},{"id":"W4413807406","doi":"10.1186/s13321-025-01073-6","title":"FusionCLM: enhanced molecular property prediction via knowledge fusion of chemical language models","year":2025,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","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":"Children’s Health Research Institute; Western University; Lawson Health Research Institute; Public Health Ontario; University of Toronto","funders":"Canadian Institutes of Health Research; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Computer science; Property (philosophy); Artificial intelligence; Ensemble learning; Inference; Machine learning; Representation (politics); Feature (linguistics); Feature learning; Deep learning; Data mining","score_opus":0.009688278155474715,"score_gpt":0.28108803723704645,"score_spread":0.2713997590815717,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413807406","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":[0.07475719,0.0018608766,0.89808357,0.0009088543,0.00012059876,0.00014492776,0.0024960232,0.017951036,0.003676964],"genre_scores_gemma":[0.6658366,0.00087787583,0.32128277,0.0009253039,0.000113180206,0.00034644094,0.0065669436,0.00052669086,0.0035242252],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9993343,0.00013331654,0.000034791665,0.00015767346,0.0002630618,0.00007692941],"domain_scores_gemma":[0.9991078,0.0004111848,0.00012296149,0.0001488575,0.00016018692,0.000049034654],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012080634,0.001443017,0.0014012407,0.0018660146,0.00039518884,0.0010551824,0.002247659,0.0013484566,0.0019634464],"category_scores_gemma":[0.0034896003,0.00053576543,0.0018422017,0.001398396,0.0004922886,0.0028599491,0.002016668,0.0021140168,0.0010096946],"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":[0.00024369062,0.00034674318,0.0045400388,0.00022580782,0.00031241204,0.00023984174,0.0000807715,0.6214088,0.00766311,0.007981999,0.011757486,0.34519935],"study_design_scores_gemma":[0.0000076078495,0.000028806238,0.00010821462,0.000004925723,0.000013248652,0.000017142462,0.0000043723676,0.99432594,0.00150377,0.003364728,0.0006135568,0.0000075951193],"about_ca_topic_score_codex":0.008492187,"about_ca_topic_score_gemma":0.010757177,"teacher_disagreement_score":0.008492187,"about_ca_system_score_codex":0.0012519384,"about_ca_system_score_gemma":0.0018826205,"threshold_uncertainty_score":0.01688546},"labels":[],"label_agreement":null},{"id":"W4414507809","doi":"10.1186/s13321-025-01084-3","title":"Cache: Utilizing ultra-large library screening in Rosetta to identify novel binders of the WD-repeat domain of Leucine-Rich Repeat Kinase 2","year":2025,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Institutes of Health; Alexander von Humboldt-Stiftung; Enamine; Max Kade Foundation; University of Toronto; Bundesministerium für Bildung und Forschung; German Network for Bioinformatics Infrastructure; Deutscher Akademischer Austauschdienst; Structural Genomics Consortium; Vanderbilt University; Deutsche Forschungsgemeinschaft; Universität Leipzig; Michael J. Fox Foundation for Parkinson's Research","keywords":"Enamine; Virtual screening; Benchmark (surveying); Drug discovery; Domain (mathematical analysis); Docking (animal); Chemical space","score_opus":0.019745294908351658,"score_gpt":0.3157505918754207,"score_spread":0.29600529696706906,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4414507809","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":[0.73216903,0.000900794,0.2521254,0.00026414206,0.000053740183,0.000239771,0.0007488727,0.0081381705,0.0053599863],"genre_scores_gemma":[0.6778936,0.00041696668,0.3175329,0.00019829693,0.000010615106,0.0003681752,0.0013871545,0.0002829757,0.0019093785],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9997955,0.000077137374,0.000010077132,0.00004052665,0.000053199594,0.00002354135],"domain_scores_gemma":[0.9998012,0.00009184512,0.000019968826,0.000030599,0.000035389734,0.000020913183],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00058175693,0.0004715826,0.00074167125,0.0003427302,0.00022064759,0.0005643793,0.0011242476,0.00040989713,0.0013401475],"category_scores_gemma":[0.0007102827,0.0002256488,0.00059462246,0.00028859076,0.00018797157,0.00035652428,0.00057631586,0.00047246064,0.00033711764],"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":[0.0011437485,0.0007835709,0.010307481,0.0005117741,0.00049285364,0.00052883296,0.00013490547,0.4930035,0.30265903,0.00891263,0.0041860333,0.1773356],"study_design_scores_gemma":[0.00007627531,0.0004146831,0.0006860393,0.000004303733,0.000049017774,0.000095373245,0.00002045375,0.9467069,0.049064618,0.0005132964,0.002346204,0.0000227765],"about_ca_topic_score_codex":0.0013916085,"about_ca_topic_score_gemma":0.0024982614,"teacher_disagreement_score":0.0013916085,"about_ca_system_score_codex":0.0003032861,"about_ca_system_score_gemma":0.0005512089,"threshold_uncertainty_score":0.004483223},"labels":[],"label_agreement":null},{"id":"W4415517900","doi":"10.1186/s13321-025-01104-2","title":"VNFlow: integration of variational autoencoders and normalizing flows for novel molecular design","year":2025,"lang":"en","type":"article","venue":"Journal of Cheminformatics","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Defence Research and Development Canada; National Research Council Canada","funders":"National Research Council Canada; Defence Research and Development Canada","keywords":"Generative grammar; Key (lock); Inverse; Range (aeronautics); Sampling (signal processing); Generative model; Inverse problem; Chemical space","score_opus":0.0154423853595157,"score_gpt":0.28014446830229384,"score_spread":0.26470208294277814,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4415517900","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":[0.007455415,0.00021112827,0.9898207,0.0001084776,0.000047190064,0.000042172735,0.000051900657,0.0006541004,0.0016088706],"genre_scores_gemma":[0.34221372,0.0004258377,0.651347,0.00031511756,0.000082772785,0.00034605424,0.00038618548,0.0005170138,0.004366225],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99973255,0.00009133659,0.0000140457105,0.00004847381,0.00008647228,0.000026998834],"domain_scores_gemma":[0.99934274,0.00042780087,0.000048300535,0.0000556124,0.00009330509,0.000032267646],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001556036,0.0010007517,0.0008038614,0.0005641997,0.00031590942,0.0007302128,0.0010336378,0.0011324438,0.0018827306],"category_scores_gemma":[0.0025663697,0.00054451654,0.00083452393,0.00039720305,0.0007457916,0.00080229,0.0010592252,0.0013966829,0.00041457175],"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":[0.00002423845,0.000028587996,0.00032103,0.000040291434,0.000035577083,0.000030701198,0.000023576564,0.9441273,0.0024427252,0.013659476,0.0005990193,0.03866749],"study_design_scores_gemma":[0.0000020030545,0.000008130246,0.000012875363,0.0000027546885,0.000002536637,0.000004541796,0.0000011730817,0.9968093,0.0004687153,0.002311165,0.0003748147,0.0000020287096],"about_ca_topic_score_codex":0.00525533,"about_ca_topic_score_gemma":0.007446356,"teacher_disagreement_score":0.00525533,"about_ca_system_score_codex":0.0010171254,"about_ca_system_score_gemma":0.0013960812,"threshold_uncertainty_score":0.010449529},"labels":[],"label_agreement":null}]}