{"meta":{"query_hash":"95db5a6f7e4e","filters":{"venue":"Journal of Structural and Functional Genomics"},"cohort_total":41,"direct_labels_cover":0,"predictions_cover":41,"exported":41,"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/95db5a6f7e4e","api":"https://metacan.xera.ac/api/v1/cohort?venue=Journal+of+Structural+and+Functional+Genomics"},"results":[{"id":"W1513446941","doi":"10.1023/a:1021261026670","title":"Role for NMR in structural genomics","year":2002,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":31,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Cancer Care Ontario; University of Toronto; Ontario Institute for Cancer Research","funders":"National Institute of General Medical Sciences","keywords":"Structural genomics; Genomics; National laboratory; Functional genomics; Computational biology; Biology; Engineering; Genetics; Genome; Engineering physics; Biochemistry; Protein structure; Gene","score_opus":0.01320816260975611,"score_gpt":0.20827084317933703,"score_spread":0.19506268056958093,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1513446941","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":[0.083720826,0.05926274,0.61493325,0.085112974,0.0073429393,0.00015138142,0.0007309943,0.0055204113,0.14322442],"genre_scores_gemma":[0.6719703,0.026664803,0.25787696,0.0075422605,0.004753202,0.00034334447,0.0006318391,0.00069725036,0.029519977],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9994821,0.00025102135,0.000015104836,0.00010784306,0.00009527578,0.00004856252],"domain_scores_gemma":[0.99754447,0.0014001937,0.000092288115,0.00048866065,0.00026118328,0.00021324235],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0035329135,0.0007245178,0.0010485167,0.00076433894,0.0012156587,0.0015783252,0.0016043717,0.0019404319,0.007322221],"category_scores_gemma":[0.0041656825,0.00046585142,0.0004002149,0.00048978685,0.0037122264,0.007130764,0.0012954072,0.0028429003,0.0018415616],"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":[0.0003766542,0.00018683434,0.00039330908,0.000489254,0.000040372946,0.00017931736,0.00018997557,0.00458427,0.086320676,0.7999939,0.013356677,0.093888834],"study_design_scores_gemma":[0.00018064272,0.00024483976,0.0006015426,0.0001382146,0.000050883667,0.00034478123,0.00015219826,0.046197515,0.050958857,0.7483618,0.15264022,0.00012844228],"about_ca_topic_score_codex":0.0009943827,"about_ca_topic_score_gemma":0.00071919366,"teacher_disagreement_score":0.007322221,"about_ca_system_score_codex":0.001640087,"about_ca_system_score_gemma":0.0007940302,"threshold_uncertainty_score":0.024495244},"labels":[],"label_agreement":null},{"id":"W1589570762","doi":"10.1023/a:1026156025612","title":"Coverage of protein sequence space by current structural genomics targets","year":2003,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":13,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Biotechnology Research Institute; McGill University","funders":"Canadian Institutes of Health Research","keywords":"Structural genomics; Protein Data Bank (RCSB PDB); Proteome; Computational biology; Protein superfamily; Homology (biology); Biology; Protein Data Bank; Homologous chromosome; Genome; Cluster analysis; Proteomics; Genetics; Protein structure; Computer science; Gene; Biochemistry; Artificial intelligence","score_opus":0.015035922793876211,"score_gpt":0.22811302812100925,"score_spread":0.21307710532713303,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1589570762","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":[0.9645452,0.0035295172,0.026506873,0.00059712166,0.000027103533,0.000010231932,0.0012976353,0.00041520997,0.0030710234],"genre_scores_gemma":[0.9911567,0.0011069964,0.0050465697,0.0001187973,0.000034536362,0.000008712022,0.0022068878,0.000070684946,0.00025031136],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.9994049,0.000105205116,0.000039169176,0.00015337268,0.00022523329,0.00007198064],"domain_scores_gemma":[0.9977812,0.0010751543,0.00036444774,0.00029389522,0.00036231562,0.0001230207],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00092007685,0.00035965093,0.00071040256,0.0011444095,0.0004137237,0.0012496897,0.00040457613,0.00053501426,0.001672161],"category_scores_gemma":[0.0030439903,0.00026678573,0.0004888866,0.0014781366,0.0004873405,0.0015481728,0.00048613534,0.0007113009,0.000684763],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0025524406,0.00022767279,0.073880516,0.0005329754,0.0002603447,0.00043288138,0.0007157982,0.022206103,0.77638984,0.011638479,0.001660498,0.10950245],"study_design_scores_gemma":[0.00019011257,0.0020381054,0.25474864,0.00020457777,0.0010088502,0.005079166,0.001452197,0.19674145,0.4618415,0.041892834,0.034662973,0.00013965716],"about_ca_topic_score_codex":0.0009990995,"about_ca_topic_score_gemma":0.0010593241,"teacher_disagreement_score":0.001672161,"about_ca_system_score_codex":0.000446922,"about_ca_system_score_gemma":0.00032202454,"threshold_uncertainty_score":0.005593896},"labels":[],"label_agreement":null},{"id":"W171658374","doi":"10.1023/a:1021352716284","title":"The cadherin superfamily database","year":2002,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Wnt/β-catenin signaling in development and cancer","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Ontario Institute for Cancer Research","funders":"National Cancer Institute; Canadian Institutes of Health Research","keywords":"SUPERFAMILY; Biology; Computational biology; Cadherin; Annotation; Protein family; Bioinformatics; Database; Genetics; Computer science; Cell; Gene","score_opus":0.01727630489704733,"score_gpt":0.20372120345385347,"score_spread":0.18644489855680613,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W171658374","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":[0.0090936655,0.0058879512,0.009994637,0.0002836722,0.00009030903,0.00010779394,0.9372878,0.022173997,0.015080193],"genre_scores_gemma":[0.015714115,0.002852241,0.007326468,0.00012536964,0.00002568681,0.00013280044,0.9686174,0.0011269188,0.004079089],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9998412,0.000013036873,0.00002697531,0.000041156844,0.00004915946,0.000028530707],"domain_scores_gemma":[0.9996971,0.00008771053,0.00003473307,0.000071470466,0.00005691811,0.000052093892],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00024998529,0.0012949309,0.0011880384,0.0058796415,0.00051250844,0.0013823017,0.0011863619,0.0009760723,0.054781873],"category_scores_gemma":[0.0016223352,0.00046493043,0.0006354416,0.0060549728,0.00017898428,0.0012156651,0.0010398523,0.0007871583,0.044237312],"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":[0.0008989748,0.00013187982,0.003189043,0.0044923117,0.00023609315,0.0006144511,0.0001910091,0.0018996899,0.025651881,0.009140726,0.77473617,0.17881775],"study_design_scores_gemma":[0.0001767984,0.000048031827,0.0053876936,0.00029446307,0.00016870175,0.0007276897,0.000071513925,0.0035741455,0.005302461,0.0054882066,0.9787147,0.00004551736],"about_ca_topic_score_codex":0.002749776,"about_ca_topic_score_gemma":0.004376351,"teacher_disagreement_score":0.054781873,"about_ca_system_score_codex":0.0004120413,"about_ca_system_score_gemma":0.0010737077,"threshold_uncertainty_score":0.1832636},"labels":[],"label_agreement":null},{"id":"W1767701883","doi":"10.1023/a:1011320027914","title":"Calmodulin Target Database","year":2000,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Protein Kinase Regulation and GTPase Signaling","field":"Biochemistry, Genetics and Molecular Biology","cited_by":548,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Ontario Institute for Cancer Research","funders":"Medical Research Council; Howard Hughes Medical Institute","keywords":"Calmodulin; Calcium-binding protein; Computational biology; Biology; Subclass; Protein structure; Calcium; Biochemistry; Genetics; Chemistry","score_opus":0.009679894184016628,"score_gpt":0.21521363712396085,"score_spread":0.20553374293994423,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1767701883","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":[0.004842472,0.0019368631,0.013138497,0.00041489874,0.00016709547,0.00022948324,0.9068577,0.046123143,0.026289796],"genre_scores_gemma":[0.010643822,0.0011203057,0.0071417857,0.00019157965,0.000047518246,0.00022532002,0.97256154,0.0017426782,0.0063253976],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9998479,0.000011172455,0.000016550475,0.00005349125,0.00005029204,0.000020505713],"domain_scores_gemma":[0.9996288,0.000121069315,0.0000437464,0.00006887811,0.00006984497,0.000067731205],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00021525213,0.0018402912,0.0013797395,0.003978903,0.00050933176,0.0013015356,0.002252305,0.0016290342,0.113956675],"category_scores_gemma":[0.0015969785,0.00054361345,0.00058982405,0.0042062444,0.00018367174,0.0011243862,0.00093573384,0.0011192135,0.08625615],"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":[0.0010936315,0.00019043744,0.00097386737,0.002115928,0.000098775025,0.00036157842,0.000057410594,0.0014542803,0.020449894,0.004398983,0.9026429,0.06616225],"study_design_scores_gemma":[0.00059443706,0.0000700242,0.0029978114,0.0001536447,0.00014261143,0.00037574882,0.00003656952,0.009875062,0.0133221,0.0053070015,0.9670643,0.000060715152],"about_ca_topic_score_codex":0.0040572234,"about_ca_topic_score_gemma":0.0049196035,"teacher_disagreement_score":0.113956675,"about_ca_system_score_codex":0.0005355347,"about_ca_system_score_gemma":0.000988748,"threshold_uncertainty_score":0.38122308},"labels":[],"label_agreement":null},{"id":"W1779770448","doi":"10.1023/a:1011348803324","title":"NMR Structure Determination and Structure-Based Functional Characterization of Conserved Hypothetical Protein MTH1175 from Methanobacterium thermoautotrophicum","year":2000,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":22,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Ontario Institute for Cancer Research","funders":"","keywords":"Thermotoga maritima; Methanobacterium; Structural genomics; Biology; Ribonuclease; Hyperthermophile; Protein family; Peptide sequence; Conserved sequence; Nuclear magnetic resonance spectroscopy; Protein structure; Biochemistry; Bacteria; Sequence alignment; Gene; Archaea; Genetics; Stereochemistry; Chemistry; RNA; Escherichia coli","score_opus":0.010704088225939913,"score_gpt":0.19143395201483077,"score_spread":0.18072986378889086,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1779770448","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":[0.9984982,0.00036005833,0.0006071689,0.000096672586,0.000009829113,0.000005286166,0.00022238013,0.000008857475,0.00019154357],"genre_scores_gemma":[0.99586225,0.00023850858,0.0017294912,0.00004244151,0.000008708653,0.000006311647,0.0016523362,0.0000073535552,0.0004525228],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99993134,0.000011973755,0.000007224629,0.000018235995,0.000015912588,0.000015274089],"domain_scores_gemma":[0.99984014,0.000024238645,0.00003392538,0.000013370885,0.00004062629,0.00004765593],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00019297696,0.00029075623,0.00025426084,0.00019078332,0.0003236987,0.00021380723,0.00029714627,0.0003410295,0.0004472549],"category_scores_gemma":[0.0003201273,0.00012489072,0.00033210331,0.00024782075,0.00019460522,0.00021571544,0.00015076566,0.00037010334,0.0001579814],"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":[0.00051321427,0.00006719501,0.0018581282,0.0000890557,0.000015720057,0.00026248043,0.00011721164,0.00043939217,0.9942247,0.00021326201,0.00009751525,0.0021021022],"study_design_scores_gemma":[0.00008766534,0.0005922061,0.060653232,0.000025855225,0.00012374758,0.0011633631,0.0004893212,0.008848923,0.9219903,0.00033259246,0.0056579933,0.000034727964],"about_ca_topic_score_codex":0.0023520824,"about_ca_topic_score_gemma":0.0020568587,"teacher_disagreement_score":0.0023520824,"about_ca_system_score_codex":0.00044630453,"about_ca_system_score_gemma":0.00030684163,"threshold_uncertainty_score":0.0046767592},"labels":[],"label_agreement":null},{"id":"W1820111519","doi":"10.1023/a:1024674220425","title":"Solution structure of the yeast ubiquitin-like modifier protein Hub1","year":2003,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":30,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium; University of Toronto; Ontario Institute for Cancer Research","funders":"Pacific Northwest National Laboratory; Battelle; National Institute of General Medical Sciences; Biological and Environmental Research; National Institutes of Health; U.S. Department of Energy","keywords":"Saccharomyces cerevisiae; Nuclear Overhauser effect; Two-dimensional nuclear magnetic resonance spectroscopy; Nuclear magnetic resonance spectroscopy; Yeast; Ubiquitin; Chemistry; Spectroscopy; Structural genomics; Nuclear magnetic resonance; Biochemistry; Protein structure; Stereochemistry; Physics; Gene","score_opus":0.012140963366007756,"score_gpt":0.19742696215736188,"score_spread":0.18528599879135413,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1820111519","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":[0.9901811,0.0017336718,0.0032329783,0.0006906801,0.0000686866,0.000009767963,0.0009606791,0.0002737186,0.0028487146],"genre_scores_gemma":[0.99295133,0.0010134605,0.0022336256,0.00005906956,0.000010219146,0.000007780628,0.0024651869,0.000049630347,0.0012097672],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.9999491,0.000003920398,0.0000015993301,0.000015625668,0.0000142597855,0.000015447116],"domain_scores_gemma":[0.99990356,0.00002881029,0.000014028754,0.000007098356,0.000020309386,0.000026155678],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00014162969,0.00036693795,0.0004806469,0.00021953371,0.0006921795,0.000578573,0.0007670255,0.0005269066,0.0030698136],"category_scores_gemma":[0.0004074327,0.00034794537,0.00018616808,0.00039750437,0.00027572433,0.00054130564,0.0002655767,0.0011322864,0.00042882445],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0032463095,0.0007024761,0.017466743,0.0016356949,0.00051029574,0.0030096755,0.0018282899,0.120540164,0.75156814,0.029561903,0.024283303,0.04564705],"study_design_scores_gemma":[0.00075684837,0.0012519527,0.028712368,0.00016945318,0.0003777297,0.0009485231,0.0025761423,0.436411,0.47015646,0.011870893,0.04642486,0.00034375355],"about_ca_topic_score_codex":0.0060673165,"about_ca_topic_score_gemma":0.006275788,"teacher_disagreement_score":0.0060673165,"about_ca_system_score_codex":0.0010191236,"about_ca_system_score_gemma":0.00072576886,"threshold_uncertainty_score":0.01206398},"labels":[],"label_agreement":null},{"id":"W1964082988","doi":"10.1007/s10969-009-9071-1","title":"Structural genomics target selection for the New York consortium on membrane protein structure","year":2009,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":46,"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","funders":"Genentech; National Institutes of Health; National Institute of General Medical Sciences; Magyar Tudományos Akadémia; Broad Institute; York University; University of Wisconsin-Madison","keywords":"Structural genomics; Transmembrane protein; Computational biology; Pipeline (software); Transmembrane domain; Biology; Genome; Membrane protein; Genomics; Selection (genetic algorithm); Protein structure; Bioinformatics; Genetics; Computer science; Gene; Biochemistry; Artificial intelligence; Membrane","score_opus":0.009241040380548059,"score_gpt":0.21255204995862745,"score_spread":0.20331100957807938,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1964082988","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":[0.40364516,0.002349365,0.49933106,0.001908437,0.00047121246,0.0067039286,0.045235474,0.022649175,0.017706277],"genre_scores_gemma":[0.21478876,0.0013940355,0.5833809,0.0010824783,0.00015679521,0.008766093,0.16845116,0.0054544047,0.016525382],"study_design_codex":"bench_or_experimental","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99849,0.00023288892,0.00013811667,0.0002974868,0.00067144266,0.0001701061],"domain_scores_gemma":[0.9978257,0.00053317915,0.0002064321,0.00037461697,0.00085549534,0.0002046105],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025134878,0.0010599308,0.0011678882,0.0012505793,0.0014739486,0.0010583699,0.0009061469,0.0007107354,0.007856937],"category_scores_gemma":[0.0037078024,0.0006422171,0.0009585813,0.001784078,0.00033210331,0.00065671327,0.001127857,0.001282135,0.0046509714],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.002223093,0.0003645483,0.008313504,0.00060841977,0.00011000308,0.0005714968,0.00036494414,0.0016588259,0.8558427,0.0034290215,0.04330334,0.083210036],"study_design_scores_gemma":[0.00046543515,0.00091541425,0.026361112,0.00007115257,0.00017942839,0.0013300152,0.00021989792,0.036088236,0.81086105,0.0020450798,0.121327735,0.00013541254],"about_ca_topic_score_codex":0.0036503891,"about_ca_topic_score_gemma":0.007942742,"teacher_disagreement_score":0.007856937,"about_ca_system_score_codex":0.0010970368,"about_ca_system_score_gemma":0.0024168151,"threshold_uncertainty_score":0.026284039},"labels":[],"label_agreement":null},{"id":"W1969747791","doi":"10.1007/s10969-011-9112-4","title":"Crystal structure of a putative transcriptional regulator SCO0520 from Streptomyces coelicolor A3(2) reveals an unusual dimer among TetR family proteins","year":2011,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Microbial Natural Products and Biosynthesis","field":"Medicine","cited_by":6,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Argonne National Laboratory; National Institute of General Medical Sciences; Biological and Environmental Research; National Institutes of Health; University of Chicago","keywords":"TetR; Streptomyces coelicolor; Dimer; Regulator; Repressor; Transcriptional regulation; Biology; DNA; Stereochemistry; Crystallography; Genetics; Chemistry; Streptomyces; Transcription factor; Gene; Bacteria","score_opus":0.025036856855131944,"score_gpt":0.22294152888660548,"score_spread":0.19790467203147355,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1969747791","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":[0.9923046,0.0018897373,0.001983363,0.00041349253,0.00006663275,0.000024480274,0.0010512933,0.00025706136,0.00200929],"genre_scores_gemma":[0.9905158,0.00071715534,0.0030755994,0.00019647049,0.000013793209,0.000022532875,0.0033835534,0.000036998255,0.0020380667],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999211,0.000009543946,0.0000053962285,0.000019996718,0.000027332979,0.000016661867],"domain_scores_gemma":[0.9998833,0.000018312598,0.000025941832,0.0000043932046,0.000010194324,0.000057739246],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00015217009,0.00044321612,0.000545068,0.00020535788,0.00038327771,0.00068835117,0.0005279799,0.00042261643,0.0014464065],"category_scores_gemma":[0.00020759291,0.00020171762,0.00022301519,0.00038080098,0.00025502077,0.0002869925,0.00027350255,0.0007562321,0.00032139948],"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":[0.0015636343,0.00030433177,0.007015784,0.0003979326,0.00013385025,0.002498043,0.00034230095,0.0030328322,0.97004336,0.0018552368,0.0039139614,0.008898803],"study_design_scores_gemma":[0.0015276235,0.0016096344,0.09656974,0.00015014455,0.00047159666,0.009574998,0.0016632972,0.09139894,0.73618686,0.0026968427,0.057886675,0.00026357727],"about_ca_topic_score_codex":0.003432459,"about_ca_topic_score_gemma":0.0047422303,"teacher_disagreement_score":0.003432459,"about_ca_system_score_codex":0.0008889955,"about_ca_system_score_gemma":0.0007073268,"threshold_uncertainty_score":0.0068249702},"labels":[],"label_agreement":null},{"id":"W1972107709","doi":"10.1007/s10969-014-9180-3","title":"The crystal structure of pyrimidine/thiamin biosynthesis precursor-like domain-containing protein CAE31940 from proteobacterium Bordetella bronchiseptica RB50, and evolutionary insight into the NMT1/THI5 family","year":2014,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Argonne National Laboratory; National Institute of General Medical Sciences; Biological and Environmental Research; National Institutes of Health; University of Chicago; U.S. Department of Energy","keywords":"Biology; Protein domain; Bordetella bronchiseptica; Yeast; Protein family; Biochemistry; Computational biology; Microbiology; Genetics; Bacteria; Gene","score_opus":0.006292172234910668,"score_gpt":0.1845157919505316,"score_spread":0.17822361971562095,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1972107709","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":[0.9852409,0.0037626883,0.0009071664,0.0007878051,0.00011907664,0.000027482864,0.0045424206,0.00021500044,0.004397474],"genre_scores_gemma":[0.9700253,0.0020986109,0.0033707714,0.00017772426,0.000020823525,0.00002486161,0.019232461,0.000052665167,0.004996696],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999422,0.000004472646,0.0000034238067,0.000012925478,0.000025173169,0.000011807093],"domain_scores_gemma":[0.99991965,0.000020506244,0.000017420361,0.000003830985,0.000014335845,0.000024298784],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00014959557,0.0003303892,0.00031488284,0.00019218738,0.00037901584,0.00039988098,0.0006126889,0.00052329735,0.0016217947],"category_scores_gemma":[0.00028612072,0.00018990363,0.0001670357,0.000403227,0.00020514202,0.00022100194,0.00014542276,0.0006515634,0.00030452939],"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":[0.0020667224,0.00040412243,0.01004807,0.0008402898,0.00015117969,0.0026899073,0.0004250796,0.005039734,0.9413768,0.0019466128,0.018186387,0.016825043],"study_design_scores_gemma":[0.0014129061,0.001018791,0.28860536,0.00023399542,0.00051754195,0.0064727576,0.0015951891,0.0767097,0.49946818,0.0020089166,0.1217501,0.00020661035],"about_ca_topic_score_codex":0.012620575,"about_ca_topic_score_gemma":0.014433811,"teacher_disagreement_score":0.012620575,"about_ca_system_score_codex":0.0011840978,"about_ca_system_score_gemma":0.00071585266,"threshold_uncertainty_score":0.025094211},"labels":[],"label_agreement":null},{"id":"W1976203775","doi":"10.1007/s10969-012-9134-6","title":"Three structural representatives of the PF06855 protein domain family from Staphyloccocus aureus and Bacillus subtilis have SAM domain-like folds and different functions","year":2012,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences","keywords":"Bacillus subtilis; Structural genomics; Protein family; Biology; Protein domain; Phylogenetic tree; Protein structure; Computational biology; Bacteria; Genetics; Biochemistry; Gene","score_opus":0.013236724642074357,"score_gpt":0.21551847655862802,"score_spread":0.20228175191655368,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1976203775","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":[0.9956975,0.0008429876,0.0020507616,0.000072524024,0.000021227861,0.000011713281,0.0003756092,0.000045233486,0.0008823311],"genre_scores_gemma":[0.99564004,0.00016677869,0.0021674503,0.00004485852,0.0000058490214,0.000011620726,0.0014521241,0.000007142375,0.00050417247],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999231,0.0000118802855,0.00000796627,0.000014051416,0.000025450016,0.000017494307],"domain_scores_gemma":[0.99983406,0.000032955933,0.00004467484,0.000012454761,0.000021045662,0.000054683016],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012393831,0.0003651656,0.0003445067,0.00045382965,0.00036000714,0.0002556535,0.00023450289,0.00036694884,0.0010291233],"category_scores_gemma":[0.0003070599,0.00008749724,0.0004044146,0.0005254083,0.00025819146,0.00025263868,0.00020023786,0.00042604556,0.00030375662],"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":[0.0016053146,0.00013708477,0.014469576,0.0002680746,0.000059518658,0.0011527338,0.00032302865,0.000647275,0.9648727,0.0011941529,0.00050183665,0.014768752],"study_design_scores_gemma":[0.0000998378,0.0011083921,0.28266695,0.000042380107,0.00019942259,0.010909688,0.00091969507,0.009372849,0.6767946,0.0014244447,0.016388642,0.000073176765],"about_ca_topic_score_codex":0.0006741979,"about_ca_topic_score_gemma":0.0004114845,"teacher_disagreement_score":0.0010291233,"about_ca_system_score_codex":0.0003242018,"about_ca_system_score_gemma":0.00018735298,"threshold_uncertainty_score":0.003442824},"labels":[],"label_agreement":null},{"id":"W1982802055","doi":"10.1007/s10969-011-9121-3","title":"Solution NMR structures reveal unique homodimer formation by a winged helix-turn-helix motif and provide first structures for protein domain family PF10771","year":2012,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences; Office of Science; National Institutes of Health","keywords":"Structural genomics; Helix-turn-helix; Dimer; Protein structure; DNA; Biology; DNA-binding domain; Protein family; Sequence motif; Crystallography; Helix (gastropod); Structural motif; Bacteroides thetaiotaomicron; Nuclear magnetic resonance spectroscopy; Chemistry; DNA-binding protein; Stereochemistry; Biochemistry; Genetics; Transcription factor; Gene; Bacteroides","score_opus":0.015098550140068038,"score_gpt":0.21601197488834323,"score_spread":0.2009134247482752,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1982802055","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":[0.9951226,0.00042440355,0.0023136584,0.00022945917,0.000026124277,0.000012845183,0.00032203493,0.00008455019,0.0014643138],"genre_scores_gemma":[0.9958384,0.00017126992,0.0020100635,0.00007917965,0.000009141109,0.000012649873,0.0011471793,0.00002157096,0.00071046234],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99993587,0.000007213715,0.000004645689,0.000016127295,0.000016557267,0.000019592479],"domain_scores_gemma":[0.9998622,0.000027374203,0.000029748044,0.000014542405,0.000021002184,0.00004509362],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012227401,0.00023336514,0.00032649585,0.00018821636,0.00032492308,0.00030116888,0.0001986587,0.00041915153,0.0011207887],"category_scores_gemma":[0.00019416834,0.000136143,0.00017069642,0.00020376939,0.00028628853,0.0003895101,0.00012336492,0.0005788532,0.00034190583],"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":[0.00027868777,0.00007805403,0.001516615,0.00006124054,0.00002037404,0.00029057005,0.00013512802,0.00043105695,0.9914316,0.0008570968,0.0011711974,0.003728401],"study_design_scores_gemma":[0.00007372304,0.00034484488,0.01841724,0.000010593771,0.000042276184,0.0015270372,0.00037327164,0.009680208,0.9601297,0.00088485825,0.0084746685,0.000041643783],"about_ca_topic_score_codex":0.00057031866,"about_ca_topic_score_gemma":0.0005304509,"teacher_disagreement_score":0.0011207887,"about_ca_system_score_codex":0.00045731114,"about_ca_system_score_gemma":0.00016654062,"threshold_uncertainty_score":0.0037493706},"labels":[],"label_agreement":null},{"id":"W1988950201","doi":"10.1007/s10969-005-5277-z","title":"NMR solution structure of Thermotoga maritima protein TM1509 reveals a Zn-metalloprotease-like tertiary structure","year":2005,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":23,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Ontario Institute for Cancer Research","funders":"National Institute of General Medical Sciences; National Science Foundation","keywords":"Thermotoga maritima; Consensus sequence; Sequence motif; Multiple sequence alignment; Structural genomics; Sequence alignment; Conserved sequence; Protein structure; Biology; Protein tertiary structure; Peptide sequence; Protein family; Biochemistry; Gene","score_opus":0.008369195511052585,"score_gpt":0.2012453206444952,"score_spread":0.19287612513344263,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1988950201","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":[0.99567235,0.00045779697,0.0016393323,0.0003660364,0.000033241577,0.000008181654,0.0005079122,0.00006830179,0.0012468421],"genre_scores_gemma":[0.9939027,0.0003406464,0.0021655094,0.00011496597,0.000013300588,0.0000068551803,0.0022087344,0.000020737902,0.0012266604],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99997723,0.0000020209357,0.0000015919125,0.000008217334,0.0000054344246,0.0000054979964],"domain_scores_gemma":[0.9999174,0.000011047631,0.000016994918,0.000006113732,0.000019776404,0.000028651359],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00007664287,0.00018093958,0.00018785473,0.00008660281,0.0002616841,0.00019657867,0.0002763947,0.00038355257,0.0015369386],"category_scores_gemma":[0.00017734223,0.000101862104,0.00016062116,0.00013076868,0.00016284523,0.00022054178,0.00010781645,0.00054200157,0.00036481553],"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":[0.00059651857,0.000089774585,0.0013734235,0.000099981684,0.000022666576,0.0003320629,0.000120940924,0.0007577816,0.9924414,0.0004868512,0.00096834206,0.0027102546],"study_design_scores_gemma":[0.00021780893,0.0009160231,0.06452468,0.000056416913,0.00016086744,0.0016732513,0.00070338824,0.03992681,0.8713527,0.0013281706,0.01905488,0.00008500424],"about_ca_topic_score_codex":0.0023909574,"about_ca_topic_score_gemma":0.0024475593,"teacher_disagreement_score":0.0023909574,"about_ca_system_score_codex":0.00036229708,"about_ca_system_score_gemma":0.00022829958,"threshold_uncertainty_score":0.0051415563},"labels":[],"label_agreement":null},{"id":"W1989259734","doi":"10.1007/s10969-012-9127-5","title":"Crystal structure of a putative isochorismatase hydrolase from Oleispira antarctica","year":2012,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Production and Characterization","field":"Biochemistry, Genetics and Molecular Biology","cited_by":24,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Argonne National Laboratory; National Institute of General Medical Sciences; Biological and Environmental Research; National Institutes of Health; University of Chicago","keywords":"Hydrolase; Structural genomics; Crystal structure; Chemistry; Biology; Crystallography; Protein structure; Enzyme; Biochemistry","score_opus":0.008212539600205014,"score_gpt":0.21491802162171178,"score_spread":0.20670548202150676,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1989259734","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":[0.99276066,0.0017470045,0.0009993013,0.00060344464,0.00008614533,0.000026152995,0.0015673373,0.00013052452,0.0020793818],"genre_scores_gemma":[0.9848675,0.0020414332,0.0021626526,0.00018837536,0.00002263111,0.000030135803,0.0072645647,0.00003378096,0.0033889269],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99994946,0.0000060516604,0.0000028890013,0.000008772317,0.000021379468,0.000011479304],"domain_scores_gemma":[0.99985516,0.00002996137,0.000028165407,0.000008096604,0.000032163665,0.000046487196],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00019867327,0.0005267702,0.0005141194,0.00022230443,0.00049644295,0.0007769616,0.00080054766,0.0004896676,0.0014346411],"category_scores_gemma":[0.00035417304,0.00022710356,0.00022131907,0.00047539527,0.0002634535,0.00029181264,0.00020534064,0.0010159439,0.00025635742],"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":[0.0038778943,0.00070950395,0.021812752,0.0013260115,0.0004653737,0.0056741345,0.0019568894,0.011324635,0.903976,0.004449031,0.014461083,0.029966556],"study_design_scores_gemma":[0.0027554207,0.0025456965,0.31633005,0.0003187451,0.0010629641,0.007223236,0.0066263597,0.14022434,0.41497976,0.0037483322,0.10384948,0.00033579537],"about_ca_topic_score_codex":0.011880289,"about_ca_topic_score_gemma":0.011922126,"teacher_disagreement_score":0.011880289,"about_ca_system_score_codex":0.0009240593,"about_ca_system_score_gemma":0.00094737887,"threshold_uncertainty_score":0.023622274},"labels":[],"label_agreement":null},{"id":"W1994346628","doi":"10.1007/s10969-007-9019-2","title":"Insights into the biology of Escherichia coli through structural proteomics","year":2007,"lang":"en","type":"review","venue":"Journal of Structural and Functional Genomics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; Queen's University; National Research Council Canada; Biotechnology Research Institute","funders":"","keywords":"Structural genomics; Biology; Escherichia coli; Computational biology; Proteomics; Function (biology); Functional genomics; Organism; Genomics; Structural biology; Model organism; Bacterial protein; Genome; Biochemistry; Genetics; Protein structure; Bacteria; Gene","score_opus":0.03257521711177249,"score_gpt":0.2860169434024889,"score_spread":0.25344172629071643,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1994346628","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":[0.00023333596,0.99615854,0.0013318217,0.00075139425,0.00054510223,0.000005754311,0.000021710994,0.000023254941,0.0009292069],"genre_scores_gemma":[0.0008382055,0.99670994,0.00094892125,0.0004371966,0.00025646234,0.0000058595583,0.000040531904,0.0000029554155,0.00075996894],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99972373,0.000028972421,0.000024774597,0.000045210178,0.00014746407,0.000029804234],"domain_scores_gemma":[0.9993666,0.00027188196,0.000068545865,0.000028344071,0.00020388006,0.00006084781],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00120713,0.0014903651,0.0017212835,0.0017965133,0.00029566465,0.0012928643,0.0019695733,0.0017546772,0.0015434904],"category_scores_gemma":[0.0009728253,0.000505065,0.00040269803,0.0022827196,0.0013771842,0.001907544,0.00084448064,0.002426719,0.002390291],"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":[0.00012823787,0.00008309632,0.00011453226,0.006130321,0.000056372355,0.00021010691,0.000031777232,0.000562463,0.01109618,0.007025027,0.027199583,0.9473623],"study_design_scores_gemma":[0.000041554606,0.000101441816,0.00070156576,0.0013445717,0.00008607917,0.0010435062,0.00004856433,0.00028034238,0.0067280885,0.005065389,0.9845244,0.00003439814],"about_ca_topic_score_codex":0.0012688182,"about_ca_topic_score_gemma":0.0015730425,"teacher_disagreement_score":0.0019695733,"about_ca_system_score_codex":0.0013446459,"about_ca_system_score_gemma":0.001314486,"threshold_uncertainty_score":0.009756148},"labels":[],"label_agreement":null},{"id":"W1996739157","doi":"10.1007/s10969-010-9095-6","title":"Fully automated high-quality NMR structure determination of small 2H-enriched proteins","year":2010,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of Diabetes and Digestive and Kidney Diseases; National Institute of General Medical Sciences; U.S. Public Health Service; National Institutes of Health","keywords":"Structural genomics; Crystal structure; Chemistry; Crystallography; Protein structure; Protonation; High resolution; Nuclear magnetic resonance spectroscopy; Stereochemistry; Biochemistry","score_opus":0.0076591753840881795,"score_gpt":0.2310563185159668,"score_spread":0.22339714313187864,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1996739157","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":[0.6128794,0.0010154019,0.37471172,0.00039501744,0.0000886133,0.00016317937,0.004063212,0.0050076973,0.0016757305],"genre_scores_gemma":[0.48985204,0.001108694,0.49719462,0.0001548336,0.00003565693,0.00031052914,0.008756856,0.0008725599,0.0017141209],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9993813,0.000110534296,0.000033428874,0.00015353253,0.00027416658,0.000046958572],"domain_scores_gemma":[0.9987913,0.0003384394,0.0001493586,0.00029089313,0.0003535724,0.000076543314],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010587656,0.0008813861,0.0009697035,0.00050711405,0.0005172022,0.00081670465,0.0013357748,0.0007133384,0.0013225953],"category_scores_gemma":[0.0016185627,0.0005764081,0.0007525298,0.000506065,0.00045826242,0.00087766186,0.0008609663,0.00127646,0.0009267579],"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":[0.0001763858,0.00005385166,0.0013076997,0.00018759185,0.00007712585,0.000114199065,0.00012383607,0.0062299133,0.97249585,0.0005788859,0.0012811584,0.017373635],"study_design_scores_gemma":[0.00007665334,0.00015976401,0.0055167703,0.000019608822,0.000045456854,0.00031767171,0.00011140218,0.08960542,0.89316523,0.00092022005,0.0099759,0.00008591581],"about_ca_topic_score_codex":0.0019032648,"about_ca_topic_score_gemma":0.0035420626,"teacher_disagreement_score":0.0019032648,"about_ca_system_score_codex":0.0005404194,"about_ca_system_score_gemma":0.0013183018,"threshold_uncertainty_score":0.0055993795},"labels":[],"label_agreement":null},{"id":"W1999196452","doi":"10.1007/s10969-010-9083-x","title":"Efficient condensed-phase production of perdeuterated soluble and membrane proteins","year":2010,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":18,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences; U.S. Public Health Service","keywords":"Membrane protein; Membrane; Escherichia coli; Chemistry; Protein structure; Biochemistry; Biophysics; Crystallography; Biology","score_opus":0.006285261081451179,"score_gpt":0.22156421216153346,"score_spread":0.2152789510800823,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1999196452","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":[0.97169614,0.0014432184,0.023421098,0.00022812796,0.00009508065,0.00011361219,0.00072964025,0.0002866987,0.0019863783],"genre_scores_gemma":[0.96939564,0.00116742,0.019999953,0.00004778378,0.0000376711,0.0000888871,0.0021051157,0.00022849823,0.00692908],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99987185,0.000024730229,0.000009483035,0.000020511523,0.000043511813,0.00002980103],"domain_scores_gemma":[0.9997825,0.000053522857,0.000049905353,0.00005144426,0.000031230324,0.00003139008],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00031994365,0.00056104304,0.0005345759,0.00024881252,0.0002603182,0.00051655015,0.00046828965,0.0003719919,0.0012666585],"category_scores_gemma":[0.0003056673,0.00025203385,0.00030728307,0.00044624516,0.00033828968,0.00048604386,0.0004763919,0.0009315304,0.00056170917],"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":[0.000081969585,0.000024542029,0.00003675334,0.000041390373,0.000006355202,0.000039616363,0.000030736544,0.00021693831,0.9973755,0.00045615053,0.00009235991,0.0015978212],"study_design_scores_gemma":[0.000022067208,0.00004801744,0.00013987903,0.000002331125,0.0000056124873,0.000036165115,0.000008467284,0.0005571794,0.99755836,0.000058180278,0.0015598943,0.000003842613],"about_ca_topic_score_codex":0.00048336943,"about_ca_topic_score_gemma":0.0008868982,"teacher_disagreement_score":0.0012666585,"about_ca_system_score_codex":0.00061452115,"about_ca_system_score_gemma":0.00035581,"threshold_uncertainty_score":0.0044587255},"labels":[],"label_agreement":null},{"id":"W2002288604","doi":"10.1007/s10969-011-9114-2","title":"The success of structural genomics","year":2011,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Genetics, Bioinformatics, and Biomedical Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":20,"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":"Structural genomics; Genomics; Computational biology; Genome Biology; Functional genomics; Data science; Biology; Computer science; Genome; Genetics; Protein structure; Gene; Biochemistry","score_opus":0.021238899856874823,"score_gpt":0.24365658385208339,"score_spread":0.22241768399520856,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2002288604","genre_codex":"commentary","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03067545,0.21981604,0.29797688,0.3195898,0.026737832,0.00020861163,0.0013692951,0.0032299391,0.100396186],"genre_scores_gemma":[0.3020982,0.21795912,0.33100572,0.06344124,0.01184057,0.0004552702,0.003552973,0.0021657096,0.06748117],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99540186,0.0019584869,0.00017589574,0.0007102738,0.0014582638,0.00029524887],"domain_scores_gemma":[0.98712385,0.0058684126,0.00058588723,0.0025496425,0.002738118,0.0011340886],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.015032339,0.0009117776,0.0012335177,0.0012156796,0.0015100418,0.0044139083,0.0017237853,0.0028849563,0.0069791824],"category_scores_gemma":[0.014901756,0.0005818692,0.0010452484,0.0013028038,0.0056688352,0.007258206,0.0027628713,0.008310334,0.004274639],"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":[0.00025862711,0.00016035003,0.0012221959,0.0012079915,0.00017395293,0.00030870334,0.00095253956,0.0021101737,0.023409052,0.5422237,0.089531325,0.33844134],"study_design_scores_gemma":[0.000048224116,0.00023912646,0.00083487935,0.00037477785,0.00004346311,0.00058191834,0.0003497814,0.0020985354,0.015904816,0.16775973,0.8116914,0.000073410934],"about_ca_topic_score_codex":0.0017504318,"about_ca_topic_score_gemma":0.0016693829,"teacher_disagreement_score":0.015032339,"about_ca_system_score_codex":0.0037916864,"about_ca_system_score_gemma":0.0036508194,"threshold_uncertainty_score":0.07949954},"labels":[],"label_agreement":null},{"id":"W2005913854","doi":"10.1007/s10969-009-9076-9","title":"Protein crystallization analysis on the World Community Grid","year":2010,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Cell Image Analysis Techniques","field":"Biochemistry, Genetics and Molecular Biology","cited_by":43,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; University Health Network","funders":"National Institutes of Health; National Institute of General Medical Sciences; Conquer Cancer Foundation","keywords":"Crystallization; Protein crystallization; Throughput; Computer science; Grid; Random forest; Set (abstract data type); Image (mathematics); Artificial intelligence; Pattern recognition (psychology); Mathematics; Engineering","score_opus":0.009274677604603711,"score_gpt":0.22991663857703923,"score_spread":0.22064196097243552,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2005913854","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":[0.45665053,0.0007064405,0.27419522,0.001267277,0.0002420154,0.00092917297,0.060605615,0.17215462,0.033249177],"genre_scores_gemma":[0.47050974,0.00047960272,0.40685025,0.00021567597,0.000033772034,0.00074407004,0.10957308,0.0062924894,0.0053012636],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99887234,0.00018806604,0.000074103125,0.0002855063,0.0004571816,0.00012286584],"domain_scores_gemma":[0.99831295,0.00017907922,0.00012034213,0.000569276,0.0006509472,0.00016741327],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017210542,0.0007421188,0.00086493534,0.0025471419,0.0011661418,0.00088064023,0.00170658,0.0003826474,0.0061388016],"category_scores_gemma":[0.0027134262,0.00035210178,0.000913203,0.0052779033,0.00036596452,0.0010382383,0.0012461594,0.0007497258,0.002835152],"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":[0.0033402708,0.0014609008,0.048850473,0.0006655753,0.00053477933,0.0019176184,0.00090642204,0.08515203,0.07260946,0.017116712,0.300944,0.46650183],"study_design_scores_gemma":[0.0005637062,0.0001766774,0.034255426,0.000054092394,0.00006426222,0.00047879128,0.00041033461,0.8082482,0.05899841,0.012120624,0.08449965,0.00012974185],"about_ca_topic_score_codex":0.026002299,"about_ca_topic_score_gemma":0.029602014,"teacher_disagreement_score":0.026002299,"about_ca_system_score_codex":0.0014133608,"about_ca_system_score_gemma":0.002047477,"threshold_uncertainty_score":0.051701903},"labels":[],"label_agreement":null},{"id":"W2006446553","doi":"10.1007/s10969-005-5243-9","title":"Automatic Classification and Pattern Discovery in High-throughput Protein Crystallization Trials","year":2005,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Advanced Proteomics Techniques and Applications","field":"Chemistry","cited_by":38,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences; Natural Sciences and Engineering Research Council of Canada","keywords":"Crystallization; Protein crystallization; Throughput; Identification (biology); Computer science; Data mining; Process (computing); Pattern recognition (psychology); Artificial intelligence; Biology; Chemistry","score_opus":0.027300690248202256,"score_gpt":0.268737495241601,"score_spread":0.24143680499339876,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2006446553","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":[0.8047468,0.0016882971,0.18514298,0.0009280584,0.000120233235,0.000659117,0.0018812722,0.0028397397,0.0019935523],"genre_scores_gemma":[0.86313415,0.0005064726,0.1311231,0.00019245307,0.000081430946,0.00032956238,0.0035004716,0.00012644922,0.0010058809],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.992867,0.0031912075,0.0008626164,0.0009786206,0.0016380218,0.00046262532],"domain_scores_gemma":[0.9802255,0.0138849495,0.00160852,0.0015005195,0.002422339,0.0003581988],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008098371,0.00070256967,0.0021194709,0.0032114964,0.00089710386,0.002775531,0.0019680152,0.0013719186,0.0007847055],"category_scores_gemma":[0.018107096,0.0004253598,0.00091326004,0.002436584,0.00068553677,0.0019454206,0.0010704395,0.0011036105,0.00055667514],"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":[0.005515095,0.0022637995,0.1247847,0.0010057742,0.00066011195,0.00043866507,0.00028263577,0.040138863,0.06589819,0.003236229,0.007871172,0.7479048],"study_design_scores_gemma":[0.0005069773,0.0008434904,0.043719843,0.000060561964,0.00045197384,0.0006663884,0.00023156028,0.8573135,0.08389306,0.008450214,0.003775958,0.0000864109],"about_ca_topic_score_codex":0.0018568132,"about_ca_topic_score_gemma":0.003427377,"teacher_disagreement_score":0.008098371,"about_ca_system_score_codex":0.0009017234,"about_ca_system_score_gemma":0.0018524748,"threshold_uncertainty_score":0.042828858},"labels":[],"label_agreement":null},{"id":"W2011846385","doi":"10.1007/s10969-014-9189-7","title":"Structural characterization of the putative ABC-type 2 transporter from Thermotoga maritima MSB8","year":2014,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Drug Transport and Resistance Mechanisms","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Argonne National Laboratory; National Institute of General Medical Sciences; Wellcome Trust; Biological and Environmental Research; National Institutes of Health; University of Chicago","keywords":"Thermotoga maritima; ATP-binding cassette transporter; Efflux; Permease; Biology; Biochemistry; Dimer; Cyclic nucleotide-binding domain; Transporter; Peptide sequence; Chemistry; Gene","score_opus":0.00995485286388935,"score_gpt":0.19622063338310833,"score_spread":0.18626578051921897,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2011846385","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":[0.99673426,0.000407365,0.0010279987,0.00020382648,0.000018427514,0.000015433397,0.00091102486,0.000050340775,0.0006313602],"genre_scores_gemma":[0.9880465,0.0003468096,0.0032384575,0.00010728844,0.000015532436,0.000019750714,0.005735072,0.00004296536,0.0024476114],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999535,0.000005787614,0.0000052185906,0.000011380463,0.00001619926,0.000007835049],"domain_scores_gemma":[0.99988925,0.00001469406,0.000027435628,0.00001122949,0.000034018016,0.000023463177],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000086459324,0.00036260978,0.00033759573,0.00021028573,0.0002688261,0.00030418928,0.00035822758,0.00040588417,0.0012283741],"category_scores_gemma":[0.00023195584,0.00015818805,0.0003732049,0.00034553127,0.0001374204,0.0002282972,0.00016674561,0.000554583,0.0006199677],"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":[0.00036809855,0.00005801381,0.0009980758,0.00006407198,0.000010216265,0.00032243563,0.000112530404,0.00040753363,0.9951633,0.00016377591,0.00017493562,0.0021569973],"study_design_scores_gemma":[0.00015085237,0.0008342566,0.075008295,0.0000528831,0.00016760822,0.0023033947,0.00109235,0.016290354,0.8743065,0.00057819457,0.029141124,0.00007428737],"about_ca_topic_score_codex":0.0046496117,"about_ca_topic_score_gemma":0.0037986776,"teacher_disagreement_score":0.0046496117,"about_ca_system_score_codex":0.00041520078,"about_ca_system_score_gemma":0.00031557857,"threshold_uncertainty_score":0.009245098},"labels":[],"label_agreement":null},{"id":"W2011850983","doi":"10.1007/s10969-012-9123-9","title":"Canadian macromolecular crystallography facility: a suite of fully automated beamlines","year":2012,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":15,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Canadian Light Source (Canada); University of Saskatchewan","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; University of Saskatchewan","keywords":"Beamline; Automation; Undulator; Suite; Computer science; Engineering; Physics; Optics; Beam (structure); Mechanical engineering; Geography","score_opus":0.008762474943783394,"score_gpt":0.2042243243349998,"score_spread":0.19546184939121639,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2011850983","genre_codex":"dataset","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.06924919,0.0071295053,0.27482846,0.01247278,0.0021751141,0.00591506,0.33656558,0.11857899,0.17308526],"genre_scores_gemma":[0.09374928,0.0049686814,0.67164135,0.0016748267,0.00031929102,0.0049313707,0.1601893,0.012867478,0.049658436],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9935137,0.0007503329,0.00018368034,0.0008650259,0.0037063442,0.0009808934],"domain_scores_gemma":[0.993028,0.0005503553,0.00030104007,0.0009954091,0.0041747554,0.0009505246],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006439545,0.0032795756,0.005628095,0.0047616786,0.010111627,0.0037192148,0.012891818,0.0021005485,0.06487568],"category_scores_gemma":[0.0058305445,0.003325529,0.001908999,0.010842515,0.001876289,0.003779844,0.0030979426,0.009867325,0.014743091],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":true,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009815549,0.00025383988,0.002347184,0.00057231286,0.0002486782,0.0001256318,0.0005509687,0.005072551,0.02999613,0.020916663,0.87966347,0.059271112],"study_design_scores_gemma":[0.0034792523,0.00024838548,0.010369655,0.00020640266,0.00041607654,0.00035898323,0.00023351064,0.057524398,0.05375283,0.00941613,0.863213,0.0007815002],"about_ca_topic_score_codex":0.6308701,"about_ca_topic_score_gemma":0.7438664,"teacher_disagreement_score":0.9693128,"about_ca_system_score_codex":0.030687196,"about_ca_system_score_gemma":0.06621446,"threshold_uncertainty_score":0.74260753},"labels":[],"label_agreement":null},{"id":"W2028079766","doi":"10.1007/s10969-013-9159-5","title":"Solution NMR structures provide first structural coverage of the large protein domain family PF08369 and complementary structural coverage of dark operative protochlorophyllide oxidoreductase complexes","year":2013,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Photosynthetic Processes and Mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences; National Institutes of Health","keywords":"Protochlorophyllide; Structural genomics; Oxidoreductase; Structural biology; Domain (mathematical analysis); Chemistry; Protein structure; Crystallography; Biochemistry; Enzyme","score_opus":0.007981530459450707,"score_gpt":0.21997556828644665,"score_spread":0.21199403782699594,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2028079766","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":[0.953638,0.002076654,0.024761936,0.001120376,0.000078978344,0.000035255405,0.01027324,0.0006901209,0.007325489],"genre_scores_gemma":[0.9554394,0.0012302345,0.0206106,0.0002680122,0.000043505388,0.00003521031,0.020601286,0.00016980387,0.0016019818],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999397,0.000005651191,0.000004204774,0.000014331088,0.000021133126,0.000015006167],"domain_scores_gemma":[0.99975103,0.00005202313,0.000039484443,0.000034380242,0.00006910439,0.000054021668],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00016323791,0.0005001158,0.0004006296,0.00036864498,0.00045945263,0.00030368913,0.000240053,0.00047502675,0.002332671],"category_scores_gemma":[0.00036244214,0.00023261408,0.0003237231,0.0005359478,0.00021706952,0.00058150163,0.00026185982,0.0011180765,0.0005924637],"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":[0.00014101643,0.00002455483,0.0005784745,0.00008159964,0.000010635735,0.00009423319,0.00008281788,0.0007761364,0.9935959,0.0005897629,0.0010334356,0.0029913879],"study_design_scores_gemma":[0.00009630689,0.00038470223,0.046276804,0.000047572004,0.00015301039,0.0018947296,0.00041470115,0.020802043,0.85419923,0.0042128693,0.071434386,0.000083560524],"about_ca_topic_score_codex":0.0018986001,"about_ca_topic_score_gemma":0.0029602256,"teacher_disagreement_score":0.002332671,"about_ca_system_score_codex":0.0004905054,"about_ca_system_score_gemma":0.0004629653,"threshold_uncertainty_score":0.0078034997},"labels":[],"label_agreement":null},{"id":"W2030909614","doi":"10.1007/s10969-005-1363-5","title":"High-throughput production of prokaryotic membrane proteins","year":2005,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":45,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Institute for Cancer Research; University of Toronto","funders":"Canada Research Chairs; Ontario Genomics Institute; Genome Canada","keywords":"Escherichia coli; Membrane protein; Biology; Transmembrane protein; Thermophile; Biochemistry; Thermotoga maritima; Integral membrane protein; Membrane; Gene; Enzyme","score_opus":0.00960853867886387,"score_gpt":0.20267470403158272,"score_spread":0.19306616535271884,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2030909614","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":[0.8052513,0.0066254707,0.17096251,0.0014609087,0.00052204565,0.00063057354,0.0072584455,0.0024725366,0.004816245],"genre_scores_gemma":[0.82294166,0.0059823697,0.14512655,0.00021133001,0.00014933186,0.00038622925,0.01591226,0.00054951233,0.008740765],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99890125,0.00023191591,0.00007817855,0.00013047809,0.000506699,0.00015145267],"domain_scores_gemma":[0.99959,0.00016136661,0.000043641114,0.00006688782,0.00006439941,0.000073864016],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00093267375,0.0010175059,0.0015730559,0.00071216497,0.00064933824,0.0017591815,0.0006767807,0.00086160353,0.0010236157],"category_scores_gemma":[0.0008765933,0.000470359,0.0010253442,0.0011289965,0.00035379393,0.0006730035,0.0010916372,0.0018761649,0.0015818909],"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":[0.00011259348,0.00006019026,0.000086855944,0.00011829565,0.000022398539,0.000062452746,0.000024805833,0.0004751211,0.9934115,0.00020578994,0.00027325077,0.005146756],"study_design_scores_gemma":[0.000030014793,0.00013553826,0.00063289306,0.00001279009,0.000036132664,0.0001366778,0.000025420028,0.0025132578,0.99220675,0.00020310894,0.0040537077,0.000013654627],"about_ca_topic_score_codex":0.0007519667,"about_ca_topic_score_gemma":0.0011130759,"teacher_disagreement_score":0.0017591815,"about_ca_system_score_codex":0.0008227998,"about_ca_system_score_gemma":0.0005767524,"threshold_uncertainty_score":0.0059698224},"labels":[],"label_agreement":null},{"id":"W2031800434","doi":"10.1007/s10969-005-9004-6","title":"The crystal structure of Rv0793, a hypothetical monooxygenase from M.␣tuberculosis","year":2006,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Biochemical and Molecular Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":18,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Lawrence Berkeley National Laboratory","keywords":"Mycobacterium tuberculosis; Tuberculosis; Polyketide; Biology; Streptomyces coelicolor; Microbiology; Oxidoreductase; Open reading frame; Bacteria; Computational biology; Function (biology); Escherichia coli; Monooxygenase; Structural genomics; Genetics; Protein structure; Gene; Streptomyces; Biochemistry; Peptide sequence; Biosynthesis; Cytochrome P450; Enzyme","score_opus":0.005991765149452562,"score_gpt":0.2057240224097127,"score_spread":0.19973225726026012,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2031800434","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":[0.97244,0.007514071,0.0023018857,0.0020085678,0.00043669846,0.000056241865,0.006921151,0.00036314767,0.007958287],"genre_scores_gemma":[0.9409754,0.0049481415,0.0088282125,0.0006823127,0.00008279705,0.0000851253,0.03520476,0.000111210415,0.009081935],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9998878,0.000013055459,0.000006399985,0.00001627093,0.000054213426,0.000022119877],"domain_scores_gemma":[0.9998567,0.00003981997,0.000023947083,0.000007565469,0.000043423108,0.000028510913],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00029771682,0.0003755741,0.00052897254,0.00021028671,0.00043782953,0.00067394,0.00059195654,0.00069240574,0.0015579549],"category_scores_gemma":[0.000528424,0.00023019088,0.00027924366,0.0004736387,0.0001885357,0.00026847594,0.00022131571,0.0013579885,0.00072353764],"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":[0.0029077583,0.00062683143,0.016876224,0.0012594671,0.00022109722,0.0025151572,0.000822631,0.003585952,0.8644152,0.005252684,0.04299888,0.058518097],"study_design_scores_gemma":[0.0011558125,0.0016586941,0.13902466,0.0005264389,0.00084522995,0.005760492,0.0022248973,0.052495614,0.5315187,0.00388613,0.26064652,0.00025686744],"about_ca_topic_score_codex":0.004105129,"about_ca_topic_score_gemma":0.0037707137,"teacher_disagreement_score":0.004105129,"about_ca_system_score_codex":0.0008451408,"about_ca_system_score_gemma":0.0006025054,"threshold_uncertainty_score":0.008162439},"labels":[],"label_agreement":null},{"id":"W2032362701","doi":"10.1007/s10969-011-9117-z","title":"Solution NMR structure of Dsy0195 homodimer from Desulfitobacterium hafniense: first structure representative of the YabP domain family of proteins involved in spore coat assembly","year":2011,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"Pacific Northwest National Laboratory; Biological and Environmental Research; National Institute of General Medical Sciences; University of Miami; U.S. Department of Energy","keywords":"Protein structure; Heteronuclear single quantum coherence spectroscopy; Biology; Domain (mathematical analysis); Spore; Biochemistry; Protein family; Nuclear magnetic resonance spectroscopy; Chemistry; Crystallography; Cell biology; Stereochemistry; Microbiology; Gene","score_opus":0.01804347108160921,"score_gpt":0.2053337360740678,"score_spread":0.1872902649924586,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2032362701","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":[0.99324363,0.0009809475,0.0016319932,0.0005035568,0.000048981907,0.000010886253,0.0016708086,0.00010882563,0.00180033],"genre_scores_gemma":[0.98906505,0.0004887213,0.0026340561,0.00013921564,0.000017765286,0.0000092925075,0.0060311025,0.000031144664,0.0015835522],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.999949,0.0000048107886,0.000003654979,0.00001745295,0.000011962439,0.000013167885],"domain_scores_gemma":[0.999861,0.000019784646,0.000031618125,0.0000070955903,0.000025439385,0.000055174394],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000143345,0.0003374226,0.00032312644,0.00024281301,0.00042440798,0.00028444905,0.00042504683,0.0004767651,0.0020565053],"category_scores_gemma":[0.00020910469,0.00018592627,0.00019588695,0.00034269554,0.00021554419,0.00033360938,0.00015627415,0.0008226564,0.00023646206],"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":[0.0009508581,0.00010771423,0.0021760876,0.00021575321,0.00006047386,0.0006470532,0.00027145442,0.0017601013,0.98700976,0.00079285644,0.0019303223,0.004077496],"study_design_scores_gemma":[0.00037355532,0.0013419768,0.11813523,0.00008908463,0.00027370217,0.0020117455,0.0014615299,0.05735831,0.786236,0.0008415502,0.031676892,0.00020044539],"about_ca_topic_score_codex":0.006255382,"about_ca_topic_score_gemma":0.00675514,"teacher_disagreement_score":0.006255382,"about_ca_system_score_codex":0.00084163004,"about_ca_system_score_gemma":0.000425043,"threshold_uncertainty_score":0.01243794},"labels":[],"label_agreement":null},{"id":"W2048205675","doi":"10.1007/s10969-007-9027-2","title":"The scientific impact of the Structural Genomics Consortium: a protein family and ligand-centered approach to medically-relevant human proteins","year":2007,"lang":"en","type":"review","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":65,"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":"Knut och Alice Wallenbergs Stiftelse; Karolinska Institutet; Wellcome Trust; VINNOVA; Stiftelsen för Strategisk Forskning; Canadian Institutes of Health Research; Genome Canada; Ontario Innovation Trust","keywords":"Structural genomics; Computational biology; Genomics; Biology; Protein family; Human proteins; Proteomics; Protein structure; Genome; Gene; Data science; Genetics; Computer science; Biochemistry","score_opus":0.049956712173686794,"score_gpt":0.2998952348970853,"score_spread":0.2499385227233985,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2048205675","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":[0.00027752857,0.9885477,0.0009980772,0.0044419565,0.002368054,0.000018179615,0.00001837096,0.000024791563,0.0033053875],"genre_scores_gemma":[0.0019252056,0.99076825,0.0014072545,0.0021741136,0.0011000679,0.000025037452,0.000035767353,0.0000067075134,0.0025575089],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9991799,0.00019617785,0.00007214667,0.00009382182,0.00040529927,0.00005260077],"domain_scores_gemma":[0.99863726,0.0004545922,0.000113211136,0.00005424351,0.00056220096,0.00017838183],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0024798964,0.00081120094,0.00081867434,0.002390953,0.0004677189,0.0012348484,0.00097389263,0.0020516934,0.0021713802],"category_scores_gemma":[0.0020864566,0.00027583266,0.00033659657,0.0022883585,0.0012399731,0.0020666933,0.00083449326,0.0019603148,0.0017595197],"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":[0.000076077085,0.000059035745,0.00013029865,0.0067960946,0.00004157159,0.00043409964,0.00012967463,0.0004957113,0.003496647,0.02241097,0.09390449,0.8720253],"study_design_scores_gemma":[0.000013008455,0.00007690005,0.00042082716,0.00088713516,0.00001726079,0.0010600989,0.0000459685,0.00006669917,0.0004970373,0.0024685112,0.9944351,0.000011409055],"about_ca_topic_score_codex":0.0016709515,"about_ca_topic_score_gemma":0.002197184,"teacher_disagreement_score":0.0024798964,"about_ca_system_score_codex":0.0014013614,"about_ca_system_score_gemma":0.002746743,"threshold_uncertainty_score":0.013115108},"labels":[],"label_agreement":null},{"id":"W2057310817","doi":"10.1007/s10969-012-9130-x","title":"Target selection for structural genomics based on combining fold recognition and crystallisation prediction methods: application to the human proteome","year":2012,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":6,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Karolinska Institutet; University of Toronto","keywords":"Human proteome project; Proteome; Structural genomics; Protein Data Bank (RCSB PDB); Computational biology; Protein crystallization; Protein Data Bank; Genomics; Biology; Crystallization; Computer science; Proteomics; Crystallography; Genome; Protein structure; Bioinformatics; Genetics; Chemistry; Gene; Biochemistry","score_opus":0.03411956348305664,"score_gpt":0.27346887615640875,"score_spread":0.2393493126733521,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2057310817","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":[0.59998554,0.00072051387,0.39313838,0.00036840522,0.00005260451,0.00018218502,0.000450696,0.0032739323,0.0018277296],"genre_scores_gemma":[0.68521345,0.00040124397,0.31161696,0.00015523986,0.0000340886,0.00018259513,0.0008368769,0.00037885035,0.0011806326],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999479,0.00019446245,0.00002174119,0.00009104845,0.00016908618,0.00004476767],"domain_scores_gemma":[0.999246,0.0004483792,0.000067461384,0.000045273162,0.00014428976,0.00004864133],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016670482,0.0007494268,0.001096342,0.0008711394,0.00055316946,0.00086412224,0.0004923236,0.00061913725,0.0013610525],"category_scores_gemma":[0.0015464209,0.00038247416,0.0006848109,0.0007689943,0.00027666386,0.00033379675,0.0007220175,0.000659071,0.0005716697],"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":[0.0018380064,0.0004983264,0.0073562125,0.0002444761,0.00013806723,0.00037923458,0.0001269935,0.021259133,0.756311,0.0014246933,0.0019301273,0.2084938],"study_design_scores_gemma":[0.00022601064,0.00047410338,0.007193109,0.000009803738,0.00014126589,0.0007277114,0.00008819108,0.52117425,0.465615,0.001853198,0.0024420922,0.00005529488],"about_ca_topic_score_codex":0.0006503323,"about_ca_topic_score_gemma":0.0009721331,"teacher_disagreement_score":0.0016670482,"about_ca_system_score_codex":0.00043663147,"about_ca_system_score_gemma":0.00038960442,"threshold_uncertainty_score":0.008816302},"labels":[],"label_agreement":null},{"id":"W2067847540","doi":"10.1007/s10969-009-9067-x","title":"Independently inducible system of gene expression for condensed single protein production (cSPP) suitable for high efficiency isotope enrichment","year":2009,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Virus-based gene therapy research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":16,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences; U.S. Public Health Service","keywords":"Chemistry; Protein biosynthesis; lac operon; Structural genomics; Biochemistry; Isotope; Stable isotope ratio; Gene expression; Gene; Protein structure","score_opus":0.019907680456250704,"score_gpt":0.2518646028660169,"score_spread":0.2319569224097662,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2067847540","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":[0.76781964,0.0015289811,0.22159423,0.0002505136,0.0002164734,0.00048058885,0.0020259167,0.0015042125,0.004579441],"genre_scores_gemma":[0.9032292,0.00096062117,0.08226857,0.0000845868,0.00006008224,0.00046190072,0.003434349,0.00040231086,0.00909854],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9997147,0.000041055908,0.00002078526,0.00007815571,0.00010776194,0.000037577203],"domain_scores_gemma":[0.9997564,0.000055207147,0.0000472373,0.00005162584,0.000048617654,0.000040914096],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002694866,0.00069818675,0.00044574114,0.000426739,0.0002474352,0.0003843485,0.0005710808,0.0004002042,0.0014609082],"category_scores_gemma":[0.00025770845,0.00025689643,0.00041581516,0.00050622225,0.0004117786,0.00041615914,0.00064998865,0.0010851161,0.0008544099],"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":[0.000029179273,0.0000095537325,0.000040304316,0.000024997396,0.0000022279744,0.000013224612,0.000006731225,0.000023061626,0.9986658,0.00019522627,0.000032540815,0.0009571965],"study_design_scores_gemma":[0.000007427235,0.000036725232,0.00030263638,0.0000015742264,0.000007922628,0.00005839923,0.0000029172866,0.00039430123,0.99779874,0.000030075571,0.0013551591,0.000004026623],"about_ca_topic_score_codex":0.0005162843,"about_ca_topic_score_gemma":0.0006770371,"teacher_disagreement_score":0.0014609082,"about_ca_system_score_codex":0.0005519105,"about_ca_system_score_gemma":0.0005204875,"threshold_uncertainty_score":0.0048872232},"labels":[],"label_agreement":null},{"id":"W2075697623","doi":"10.1007/s10969-005-9001-9","title":"Crystal structures of a phosphotransacetylase from Bacillus subtilis and its complex with acetyl phosphate","year":2005,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":26,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"Basic Energy Sciences; National Institute of General Medical Sciences; National Institutes of Health; U.S. Department of Energy","keywords":"Cofactor; Bacillus subtilis; Enzyme; Biochemistry; Dehydrogenase; Transferase; Phosphate; Acetyl-CoA; Oxidoreductase; Substrate (aquarium); Binding site; Stereochemistry; Isocitrate dehydrogenase; Active site; Chemistry; Biology; Bacteria","score_opus":0.015230629731186458,"score_gpt":0.20502133398984274,"score_spread":0.18979070425865627,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2075697623","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":[0.9900832,0.0036219445,0.0011914036,0.0007256183,0.00011727917,0.000024251045,0.0016621808,0.00012892984,0.0024451741],"genre_scores_gemma":[0.9794287,0.0037505091,0.0039517805,0.00018499864,0.00003295494,0.000035706194,0.009195885,0.00005112082,0.0033683947],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9997708,0.000021766336,0.000017223008,0.000033928587,0.00010653175,0.000049695704],"domain_scores_gemma":[0.9995509,0.00014494057,0.00006681278,0.000017399363,0.00007529198,0.00014461609],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003078355,0.0005541708,0.0008475431,0.00031954216,0.0007941735,0.0010398087,0.0008145277,0.0007368329,0.0017495204],"category_scores_gemma":[0.0007662621,0.00044985392,0.00027513978,0.0008392348,0.00036863305,0.00036519047,0.0002464123,0.0018337327,0.00025845278],"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":[0.006833609,0.0021401613,0.018899927,0.0015206492,0.00063063786,0.0047451453,0.001789314,0.011517538,0.8903023,0.008303821,0.02139813,0.031918664],"study_design_scores_gemma":[0.0020805143,0.0019048469,0.09653733,0.00020735072,0.0007526214,0.009022251,0.0030872503,0.06923919,0.7509919,0.0027883402,0.06297007,0.00041837574],"about_ca_topic_score_codex":0.014539009,"about_ca_topic_score_gemma":0.011469078,"teacher_disagreement_score":0.014539009,"about_ca_system_score_codex":0.0013199992,"about_ca_system_score_gemma":0.0011452979,"threshold_uncertainty_score":0.02890873},"labels":[],"label_agreement":null},{"id":"W2075773207","doi":"10.1007/s10969-005-3789-1","title":"Crystal structure of the hypothetical protein TA1238 from Thermoplasma acidophilum: A new type of helical super-bundle","year":2005,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":6,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium; University of Toronto; University Health Network","funders":"Argonne National Laboratory; National Institute of General Medical Sciences","keywords":"Thermoplasma acidophilum; Crystallography; Bundle; Crystal structure; Type (biology); Chemistry; Biology; Materials science; Biochemistry; Ecology; Enzyme","score_opus":0.015910516625773745,"score_gpt":0.21016764220136686,"score_spread":0.1942571255755931,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2075773207","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":[0.99478793,0.0011286782,0.0012041777,0.0003592327,0.00006454406,0.000019436351,0.0012456616,0.000113044276,0.001077336],"genre_scores_gemma":[0.9889501,0.0007842237,0.003216843,0.00012999489,0.000022095724,0.000029840112,0.004878524,0.000032954198,0.0019554906],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999484,0.000007980638,0.0000041139183,0.000011916386,0.000018243158,0.00000926764],"domain_scores_gemma":[0.99985516,0.000024600651,0.00003713422,0.000004945986,0.0000274725,0.000050639654],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012564876,0.00029523592,0.00046824143,0.00013314486,0.00036150886,0.00042121237,0.00043804848,0.00047727686,0.0014706329],"category_scores_gemma":[0.00026005475,0.00018919696,0.00015509175,0.00032597303,0.00020965468,0.00027333258,0.00016802402,0.0008377395,0.00033672818],"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":[0.0027508487,0.00030545518,0.006589145,0.0006786938,0.00013295036,0.0026001937,0.0004947164,0.0028799628,0.95914125,0.0026096299,0.009249387,0.0125677325],"study_design_scores_gemma":[0.0018787977,0.0017630225,0.13462244,0.00033130049,0.0006352764,0.011028537,0.0024128533,0.10610979,0.6392935,0.0043928856,0.09726272,0.00026883473],"about_ca_topic_score_codex":0.0015989786,"about_ca_topic_score_gemma":0.0017024988,"teacher_disagreement_score":0.0015989786,"about_ca_system_score_codex":0.0005022524,"about_ca_system_score_gemma":0.00036164504,"threshold_uncertainty_score":0.004919708},"labels":[],"label_agreement":null},{"id":"W2084106291","doi":"10.1007/s10969-010-9093-8","title":"Engineering of a wheat germ expression system to provide compatibility with a high throughput pET-based cloning platform","year":2010,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Viral Infectious Diseases and Gene Expression in Insects","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":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences","keywords":"Wheat germ; Structural genomics; Biology; Cloning (programming); Expression vector; Escherichia coli; Computational biology; Fusion protein; Cell-free protein synthesis; Protein biosynthesis; Biochemistry; Cell biology; Gene; Protein structure; Computer science","score_opus":0.005906233445938529,"score_gpt":0.20553292598271128,"score_spread":0.19962669253677276,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2084106291","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":[0.5502334,0.00090702827,0.42876738,0.0009159914,0.00038317803,0.0022141712,0.007025347,0.0034827706,0.006070692],"genre_scores_gemma":[0.56530565,0.0014712313,0.39410752,0.0005524792,0.00007379347,0.0016286172,0.01917812,0.0015470652,0.016135572],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9992586,0.00012597404,0.00011638189,0.00015190631,0.00023843104,0.000108646025],"domain_scores_gemma":[0.9996171,0.00009150607,0.00006253665,0.000089071305,0.00007970979,0.000060063245],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009471759,0.0006689108,0.0006828594,0.00053560233,0.00044464244,0.00092708593,0.00086499425,0.0008466387,0.0019793005],"category_scores_gemma":[0.0007227062,0.0006305943,0.0006053342,0.00072797236,0.00030033273,0.000541148,0.0007099102,0.0016478035,0.0032423693],"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":[0.000042332354,0.000044030145,0.00014939095,0.00006164869,0.000008500232,0.00011489406,0.0000465878,0.00029053827,0.9961098,0.0004011855,0.00033493494,0.00239614],"study_design_scores_gemma":[0.00003934322,0.00020293362,0.0017829443,0.000015629843,0.00003395018,0.00042313416,0.00004795552,0.00604577,0.9632709,0.00021294254,0.02790184,0.000022624496],"about_ca_topic_score_codex":0.0009887132,"about_ca_topic_score_gemma":0.0012958298,"teacher_disagreement_score":0.0019793005,"about_ca_system_score_codex":0.0005668295,"about_ca_system_score_gemma":0.0003926806,"threshold_uncertainty_score":0.006621361},"labels":[],"label_agreement":null},{"id":"W2088061860","doi":"10.1007/s10969-012-9136-4","title":"Crystal structure of a catalytically active GG(D/E)EF diguanylate cyclase domain from Marinobacter aquaeolei with bound c-di-GMP product","year":2012,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Bacterial biofilms and quorum sensing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":16,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"Brookhaven National Laboratory; National Institute of General Medical Sciences","keywords":"Allosteric regulation; GTP'; Enzyme; Chemistry; Second messenger system; Domain (mathematical analysis); Crystallography; Stereochemistry; Biochemistry","score_opus":0.005926118764505403,"score_gpt":0.19115358542107957,"score_spread":0.18522746665657416,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2088061860","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":[0.99469817,0.0011983035,0.0010765671,0.0004198546,0.000073624884,0.00002324999,0.00088596455,0.000109002576,0.0015152103],"genre_scores_gemma":[0.99106956,0.0006814054,0.002427261,0.00018325556,0.000015906622,0.000021680527,0.0030330077,0.000025953044,0.0025418662],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99992406,0.000007081598,0.000005183498,0.000018054267,0.000027235796,0.00001844642],"domain_scores_gemma":[0.99986994,0.000031591884,0.000022031945,0.000007297905,0.00001824604,0.000050892348],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00021181606,0.000350918,0.0004762141,0.0001216517,0.00039033155,0.00045887532,0.00069755793,0.00071627233,0.0010914032],"category_scores_gemma":[0.00031788528,0.00021903256,0.00015803578,0.0002557576,0.0002938616,0.00028527033,0.00021963587,0.0010747378,0.0002316577],"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":[0.0019866915,0.0006335173,0.0055864593,0.00052514544,0.00010612474,0.0040120734,0.00066337554,0.0041076946,0.9634861,0.003131046,0.0044396305,0.011322116],"study_design_scores_gemma":[0.001310489,0.0009099078,0.07945233,0.0001454303,0.00030336538,0.006073438,0.0020925126,0.054167595,0.80763507,0.0019741717,0.045722656,0.00021294058],"about_ca_topic_score_codex":0.0075051165,"about_ca_topic_score_gemma":0.0076383487,"teacher_disagreement_score":0.0075051165,"about_ca_system_score_codex":0.0013088525,"about_ca_system_score_gemma":0.0007884966,"threshold_uncertainty_score":0.014922857},"labels":[],"label_agreement":null},{"id":"W2089640001","doi":"10.1007/s10969-006-9008-x","title":"Crystal Structure of the Vitamin B12 Biosynthetic Cobaltochelatase, CbiXS, from Archaeoglobus Fulgidus","year":2006,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Porphyrin Metabolism and Disorders","field":"Biochemistry, Genetics and Molecular Biology","cited_by":36,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; University of Alberta","funders":"National Institute of General Medical Sciences; Biotechnology and Biological Sciences Research Council","keywords":"Cobalamin; Tetrapyrrole; Biosynthesis; Histidine; Biochemistry; Lyase; Chemistry; Biology; Stereochemistry; Enzyme; Vitamin B12","score_opus":0.0037118086300047225,"score_gpt":0.17770394040004917,"score_spread":0.17399213177004444,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2089640001","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":[0.990655,0.0027737669,0.0008536367,0.0009052952,0.00013107633,0.000021960755,0.0019025261,0.00021490315,0.0025418352],"genre_scores_gemma":[0.9864313,0.0017470258,0.002023089,0.00013361924,0.0000265774,0.0000214017,0.005227046,0.000039937986,0.0043499875],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99993837,0.000005561257,0.0000035776288,0.0000126540735,0.000026860047,0.000013010242],"domain_scores_gemma":[0.9998733,0.000027157537,0.000021137548,0.0000061304986,0.000031107007,0.000041161187],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00020685693,0.00033179353,0.00048436475,0.00019528123,0.00051451556,0.00063066406,0.0005248159,0.0004136192,0.001215069],"category_scores_gemma":[0.0003640972,0.00023097092,0.00013045619,0.00035849866,0.00024532198,0.00018374392,0.0001998501,0.0009064249,0.00016397519],"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":[0.004912575,0.0007016547,0.022813119,0.0016205222,0.00032498033,0.0042467057,0.001870353,0.011365115,0.855702,0.007299096,0.036078233,0.053065754],"study_design_scores_gemma":[0.0029835356,0.0014492905,0.2134564,0.00031661196,0.0006845298,0.006299269,0.004892266,0.09902056,0.52868205,0.0039363354,0.13794504,0.00033407452],"about_ca_topic_score_codex":0.015523223,"about_ca_topic_score_gemma":0.01645323,"teacher_disagreement_score":0.015523223,"about_ca_system_score_codex":0.0011385406,"about_ca_system_score_gemma":0.0010658005,"threshold_uncertainty_score":0.030865729},"labels":[],"label_agreement":null},{"id":"W2090727546","doi":"10.1007/s10969-011-9115-1","title":"Solution NMR structure of MED25(391–543) comprising the activator-interacting domain (ACID) of human mediator subunit 25","year":2011,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Herpesvirus Infections and Treatments","field":"Medicine","cited_by":23,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of Allergy and Infectious Diseases; National Institute of General Medical Sciences","keywords":"Transactivation; Heteronuclear molecule; Protein subunit; Activator (genetics); Biology; Mediator; Protein structure; Cell biology; Protein domain; Heteronuclear single quantum coherence spectroscopy; Molecular biology; Gene; Biochemistry; Transcription factor; Chemistry; Nuclear magnetic resonance spectroscopy; Stereochemistry","score_opus":0.03355596472384254,"score_gpt":0.260084532259546,"score_spread":0.22652856753570344,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2090727546","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":[0.9856937,0.001755604,0.0026729854,0.0011853487,0.00016617446,0.000034995912,0.0020282709,0.0002278269,0.006235175],"genre_scores_gemma":[0.986414,0.000849381,0.0041888678,0.00028807353,0.000048090016,0.000018904948,0.005504343,0.000039198945,0.0026492113],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999262,0.000010094619,0.0000045975326,0.00002053797,0.000014949102,0.000023549412],"domain_scores_gemma":[0.99982566,0.000039402148,0.000041672807,0.00000950061,0.00002168184,0.00006205844],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002076713,0.0004451919,0.0003245958,0.00021834238,0.00065371965,0.0004112551,0.00094331097,0.0008165688,0.0053839777],"category_scores_gemma":[0.000397583,0.00023421306,0.00026283207,0.00030994552,0.000380296,0.0005116333,0.00025888238,0.0010577907,0.00072359416],"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":[0.0041415696,0.00033049795,0.0031989699,0.0005306366,0.00018701344,0.0025322775,0.0005420863,0.009758434,0.9565917,0.0040833005,0.00859078,0.009512733],"study_design_scores_gemma":[0.0015816563,0.0024316248,0.055655345,0.00025436917,0.00071748224,0.0028863044,0.0018528558,0.087266706,0.74591386,0.0033058175,0.09765279,0.00048119345],"about_ca_topic_score_codex":0.0057762335,"about_ca_topic_score_gemma":0.0056772977,"teacher_disagreement_score":0.0057762335,"about_ca_system_score_codex":0.001224049,"about_ca_system_score_gemma":0.0006568412,"threshold_uncertainty_score":0.018011153},"labels":[],"label_agreement":null},{"id":"W2108076107","doi":"10.1007/s10969-008-9041-z","title":"Structure of an acetyl-CoA binding protein from Staphylococcus aureus representing a novel subfamily of GCN5-related N-acetyltransferase-like proteins","year":2008,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":17,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"Pacific Northwest National Laboratory; Battelle; National Institute of General Medical Sciences; Biological and Environmental Research; National Institutes of Health; U.S. Department of Energy","keywords":"Subfamily; Biology; Acetyltransferases; Cysteine; Acetyltransferase; Biochemistry; Residue (chemistry); Transferase; Stereochemistry; Enzyme; Acetylation; Chemistry; Gene","score_opus":0.01702220829124329,"score_gpt":0.2103631739380425,"score_spread":0.1933409656467992,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2108076107","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":[0.99259394,0.0018189225,0.002800132,0.00030328127,0.000055981516,0.00002611262,0.00071205373,0.00006426464,0.001625308],"genre_scores_gemma":[0.9897759,0.0006569687,0.0043297973,0.00017582257,0.000025103565,0.000011851437,0.0032320267,0.000018098395,0.001774489],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99994946,0.000006235157,0.0000034668874,0.0000142597355,0.000017037919,0.000009524922],"domain_scores_gemma":[0.99984777,0.000027547356,0.000035204892,0.00000960094,0.000021847582,0.00005811001],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00009288738,0.00045289466,0.00045537492,0.00023073649,0.00036914472,0.0003707489,0.00046287972,0.0007557379,0.00090251624],"category_scores_gemma":[0.00019588695,0.00013515068,0.0004903194,0.00035289602,0.0002465749,0.00022939338,0.00016558154,0.0006127882,0.0003638489],"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":[0.00060237525,0.000086012486,0.0015340975,0.00014875279,0.000037598475,0.0017185035,0.00014716371,0.00088998734,0.99093014,0.0006507726,0.0004099767,0.0028447194],"study_design_scores_gemma":[0.00020131108,0.0013727037,0.15543006,0.00012447538,0.000380814,0.009014362,0.0011589271,0.039335378,0.76812965,0.0012039712,0.023481362,0.00016696533],"about_ca_topic_score_codex":0.006565481,"about_ca_topic_score_gemma":0.0045134784,"teacher_disagreement_score":0.006565481,"about_ca_system_score_codex":0.0009356158,"about_ca_system_score_gemma":0.00039905042,"threshold_uncertainty_score":0.01305455},"labels":[],"label_agreement":null},{"id":"W2142745444","doi":"10.1007/s10969-011-9107-1","title":"Crystal structure of secretory protein Hcp3 from Pseudomonas aeruginosa","year":2011,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Vibrio bacteria research studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":42,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University Health Network; Structural Genomics Consortium; University of Toronto","funders":"Argonne National Laboratory; National Institute of General Medical Sciences","keywords":"Pseudomonas aeruginosa; Type VI secretion system; Biology; Secretion; Microbiology; Effector; Virulence; Flagellum; Cell envelope; Bacteria; Peptide sequence; Cell biology; Genetics; Gene; Biochemistry; Escherichia coli","score_opus":0.01597394300057982,"score_gpt":0.21333696824884563,"score_spread":0.19736302524826582,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2142745444","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":[0.9911014,0.0018329517,0.00092788145,0.0010304117,0.000092304166,0.000024384319,0.0017272036,0.00013047835,0.0031330667],"genre_scores_gemma":[0.9878612,0.001422237,0.0018839334,0.00028497388,0.000019391327,0.000033059296,0.004936668,0.000035257013,0.00352333],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99994135,0.000004827826,0.0000023643415,0.000008270824,0.000025608173,0.000017506643],"domain_scores_gemma":[0.99991333,0.000024526931,0.000014262921,0.0000035399087,0.000017019942,0.000027299317],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012601208,0.00030677376,0.00033377734,0.00013908751,0.0006078311,0.00044063674,0.0003692855,0.00039270477,0.0019824998],"category_scores_gemma":[0.00026158532,0.00014693958,0.00019679489,0.00038922703,0.00026195243,0.00023929661,0.00022695572,0.0010628848,0.00029402805],"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":[0.0017530092,0.00033727047,0.011619853,0.0007555026,0.000113599424,0.006272306,0.000721599,0.0049987254,0.9359464,0.0036996654,0.015172701,0.0186093],"study_design_scores_gemma":[0.0013441261,0.00095558504,0.15959275,0.00023917266,0.0002746398,0.00871243,0.003278263,0.09069655,0.6568659,0.0043028793,0.0735268,0.00021104407],"about_ca_topic_score_codex":0.011398963,"about_ca_topic_score_gemma":0.012866321,"teacher_disagreement_score":0.011398963,"about_ca_system_score_codex":0.0013953329,"about_ca_system_score_gemma":0.0010222207,"threshold_uncertainty_score":0.022665262},"labels":[],"label_agreement":null},{"id":"W2146732641","doi":"10.1007/s10969-011-9122-2","title":"Solution NMR structures reveal a distinct architecture and provide first structures for protein domain family PF04536","year":2011,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":8,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences; Office of Science; National Institutes of Health","keywords":"Protein structure; Structural similarity; Protein domain; Sequence alignment; Protein family; Structural genomics; Biology; Crystallography; Heteronuclear single quantum coherence spectroscopy; Residue (chemistry); Peptide sequence; Nuclear magnetic resonance spectroscopy; Chemistry; Computational biology; Biochemistry; Stereochemistry","score_opus":0.021306286739727398,"score_gpt":0.21110617341875962,"score_spread":0.18979988667903222,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2146732641","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":[0.98369193,0.0014473094,0.008406899,0.00063734833,0.00007292852,0.000031769632,0.0013542497,0.00019566144,0.004161854],"genre_scores_gemma":[0.98471916,0.0005854603,0.0075298967,0.00015421284,0.00002112866,0.000025329766,0.0053414446,0.00003683523,0.0015865188],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999502,0.0000063920515,0.0000040976056,0.000011503751,0.000012007476,0.000015838776],"domain_scores_gemma":[0.9998896,0.000024405439,0.0000205971,0.000009769146,0.000025908841,0.000029734903],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001296788,0.00035386105,0.00029702787,0.00022043793,0.00041728796,0.00036632616,0.00025514115,0.000547014,0.0014517385],"category_scores_gemma":[0.00024934945,0.00012255684,0.00023578867,0.00034234455,0.00028832429,0.00038998493,0.00014145154,0.0006845353,0.0005112779],"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":[0.00041256324,0.00006911806,0.002106773,0.00011553999,0.000023442233,0.00040156883,0.00018589316,0.00069254544,0.98460567,0.0016841305,0.00204525,0.007657406],"study_design_scores_gemma":[0.00008188717,0.0005844052,0.032858282,0.000028591665,0.000097517564,0.002936588,0.0005595614,0.0132940775,0.90228766,0.0030631218,0.04413681,0.00007152263],"about_ca_topic_score_codex":0.0013658707,"about_ca_topic_score_gemma":0.0009342747,"teacher_disagreement_score":0.0014517385,"about_ca_system_score_codex":0.0005165359,"about_ca_system_score_gemma":0.00033760633,"threshold_uncertainty_score":0.004856527},"labels":[],"label_agreement":null},{"id":"W2159028403","doi":"10.1007/s10969-009-9064-0","title":"Unique opportunities for NMR methods in structural genomics","year":2009,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":17,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Structural Genomics Consortium","funders":"National Center for Research Resources; National Institute of General Medical Sciences; National Institutes of Health","keywords":"Structural genomics; Structural biology; Genomics; Computational biology; Protein structure; Biology; Biochemistry; Genome; Gene","score_opus":0.02975699021587422,"score_gpt":0.2929266165796197,"score_spread":0.2631696263637455,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2159028403","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":[0.006266163,0.18277486,0.5982461,0.12900987,0.005265298,0.00011189718,0.0002771112,0.0010936363,0.076955035],"genre_scores_gemma":[0.15911591,0.1544499,0.641068,0.01554903,0.015145206,0.00071484275,0.0003875876,0.000517187,0.013052347],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9949785,0.0034652378,0.00012160748,0.00041259525,0.000765258,0.00025680693],"domain_scores_gemma":[0.9899706,0.00745264,0.00021631211,0.0012686349,0.0005052151,0.00058654847],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01873276,0.0011161786,0.0014549639,0.0027349852,0.0018479106,0.0040244632,0.0019194712,0.003671506,0.0060031405],"category_scores_gemma":[0.011867948,0.0007813516,0.00095830864,0.0016453218,0.010339934,0.013934073,0.006724849,0.0067883213,0.002059893],"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":[0.000055873494,0.00005621229,0.00041612374,0.00039728286,0.00004305123,0.00013813731,0.0004969878,0.0016074312,0.0020453634,0.90146077,0.013586418,0.07969648],"study_design_scores_gemma":[0.000024452938,0.000046218905,0.00021379234,0.00019579324,0.0000112533235,0.00022856258,0.00024469793,0.0030708539,0.0008203857,0.8663618,0.12873489,0.000047220205],"about_ca_topic_score_codex":0.00089871103,"about_ca_topic_score_gemma":0.0012332258,"teacher_disagreement_score":0.01873276,"about_ca_system_score_codex":0.0022437025,"about_ca_system_score_gemma":0.0022347304,"threshold_uncertainty_score":0.099069476},"labels":[],"label_agreement":null},{"id":"W2165120792","doi":"10.1007/s10969-012-9131-9","title":"Crystal structures of putative phosphoglycerate kinases from B. anthracis and C. jejuni","year":2012,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":7,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"National Institute of Allergy and Infectious Diseases","keywords":"Phosphoglycerate kinase; Campylobacter jejuni; Bacillus anthracis; Context (archaeology); Biology; Phosphoglycerate mutase; Protein structure; Biochemistry; Glycolysis; Enzyme; Bacteria; Genetics","score_opus":0.014995294211277906,"score_gpt":0.2272977872880978,"score_spread":0.2123024930768199,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2165120792","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":[0.9864423,0.0038192053,0.0010765216,0.0009684944,0.0001300215,0.00003994506,0.0035083147,0.00013902098,0.003876115],"genre_scores_gemma":[0.9694653,0.0026384003,0.0045116274,0.00021350628,0.000032290347,0.00006453492,0.01695624,0.000046941826,0.006071035],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9998914,0.000012324673,0.0000078983285,0.00001813409,0.00004386293,0.000026396423],"domain_scores_gemma":[0.9997794,0.00006671853,0.000045910674,0.000007975949,0.000042140775,0.000057849535],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00024967943,0.0006622678,0.0004803084,0.00034210345,0.00061114255,0.0006441158,0.00076558767,0.00069970015,0.002406819],"category_scores_gemma":[0.00057804293,0.00035890474,0.00025594444,0.0006429974,0.0002635143,0.0003794343,0.00025483264,0.001336287,0.00050042913],"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":[0.0075113187,0.0013701316,0.015341403,0.0020815842,0.00027863798,0.002364868,0.0019309003,0.01148966,0.90868556,0.008784213,0.019645693,0.02051605],"study_design_scores_gemma":[0.0031567586,0.0020509474,0.15847856,0.00037725698,0.0008049138,0.004839216,0.0046566385,0.09241546,0.6312672,0.0055281497,0.096115686,0.00030920556],"about_ca_topic_score_codex":0.005927786,"about_ca_topic_score_gemma":0.0067066834,"teacher_disagreement_score":0.005927786,"about_ca_system_score_codex":0.001616102,"about_ca_system_score_gemma":0.00074445555,"threshold_uncertainty_score":0.01178658},"labels":[],"label_agreement":null},{"id":"W4234286444","doi":"10.1007/s10969-004-3789-6","title":"Crystal structure of the hypothetical protein TA1238 from Thermoplasma acidophilum: a new type of helical super-bundle","year":2004,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Porphyrin Metabolism and Disorders","field":"Biochemistry, Genetics and Molecular Biology","cited_by":7,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Structural Genomics Consortium; University Health Network","funders":"Argonne National Laboratory","keywords":"Thermoplasma acidophilum; Antiparallel (mathematics); Coiled coil; Crystallography; Protein structure; Crystal structure; Biology; Sequence (biology); Stereochemistry; Chemistry; Biophysics; Biochemistry; Physics; Enzyme","score_opus":0.010008160577533023,"score_gpt":0.20289442635413768,"score_spread":0.19288626577660467,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4234286444","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":[0.99461347,0.0011832493,0.0011686557,0.00042349665,0.00006581382,0.000020517295,0.0012694908,0.00013170016,0.0011236772],"genre_scores_gemma":[0.9898463,0.0007373495,0.0030125892,0.00013935684,0.000021328977,0.00003044952,0.004328196,0.000034440283,0.0018499206],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9999435,0.000009517274,0.0000043687432,0.000012927344,0.00001945965,0.0000102638105],"domain_scores_gemma":[0.9998492,0.000027136122,0.000038779574,0.0000048421357,0.00002630163,0.000053715125],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012887987,0.0003078393,0.0004746547,0.00014594049,0.00040751544,0.00042260494,0.00046429492,0.0005008345,0.0016185635],"category_scores_gemma":[0.00027010293,0.00018969047,0.00015176782,0.00034366772,0.00021905679,0.0002588762,0.0001808504,0.00089944556,0.00034734877],"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":[0.0030269444,0.0003405988,0.007828437,0.0007862067,0.00015551735,0.0037762062,0.0005721076,0.004027774,0.9508999,0.0030033942,0.011247428,0.014335404],"study_design_scores_gemma":[0.0021408738,0.0017866374,0.14301899,0.00036891963,0.00066372677,0.0135092735,0.0027670134,0.13303035,0.59586066,0.0051946566,0.10134379,0.00031507495],"about_ca_topic_score_codex":0.0018417028,"about_ca_topic_score_gemma":0.0019227783,"teacher_disagreement_score":0.0018417028,"about_ca_system_score_codex":0.0005559892,"about_ca_system_score_gemma":0.00037264713,"threshold_uncertainty_score":0.0054146647},"labels":[],"label_agreement":null},{"id":"W426015217","doi":"10.1023/b:jsfg.0000029244.65028.71","title":"2003 NIH Protein Structure Initiative Workshop in Protein Production and Crystallization for Structural and Functional Genomics","year":2004,"lang":"en","type":"article","venue":"Journal of Structural and Functional Genomics","topic":"Enzyme Structure and Function","field":"Materials Science","cited_by":11,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Structural Genomics Consortium","funders":"National Institute of General Medical Sciences","keywords":"Structural genomics; Genomics; Functional genomics; Crystallization; Computational biology; Biology; Protein structure; Genetics; Genome; Engineering; Biochemistry; Gene; Chemical engineering","score_opus":0.01934834248250878,"score_gpt":0.22376160785547122,"score_spread":0.20441326537296245,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W426015217","genre_codex":"commentary","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.11525555,0.039703045,0.18435808,0.5215753,0.04108751,0.0019593309,0.008489569,0.0043721483,0.08319955],"genre_scores_gemma":[0.27623168,0.042706106,0.29313776,0.056641337,0.010002695,0.0018402183,0.031552866,0.003123734,0.2847636],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9982894,0.00039873194,0.00010414537,0.0002351352,0.00060323917,0.00036939819],"domain_scores_gemma":[0.9968939,0.0003102492,0.00014878425,0.00025326826,0.0012020036,0.0011917893],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0128928255,0.0018419262,0.0017515728,0.0014729579,0.0032892772,0.004277388,0.0037322715,0.004995485,0.008302627],"category_scores_gemma":[0.003792248,0.0012082383,0.001135191,0.0015226513,0.0017812252,0.003288033,0.003769532,0.0071724937,0.0031783823],"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":[0.0021549445,0.0010013399,0.0019650806,0.00042690104,0.00008858001,0.00077866187,0.001099775,0.0030178414,0.060435016,0.020639282,0.78925484,0.11913765],"study_design_scores_gemma":[0.00021516846,0.0002425365,0.0047658086,0.0002200361,0.0001059507,0.0004516585,0.00085480756,0.004207029,0.055503078,0.009959081,0.92339617,0.00007866217],"about_ca_topic_score_codex":0.032285366,"about_ca_topic_score_gemma":0.07408874,"teacher_disagreement_score":0.032285366,"about_ca_system_score_codex":0.0064175255,"about_ca_system_score_gemma":0.010386241,"threshold_uncertainty_score":0.068184614},"labels":[],"label_agreement":null}]}