{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":54,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":54,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"2a78974ab50e","filters":{"venue":"NAR Genomics and Bioinformatics"}},"results":[{"id":"W4310786235","doi":"10.1093/nargab/lqac092","title":"PBSIM3: a simulator for all types of PacBio and ONT long reads","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":106,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Institute of Genetics; Japan Society for the Promotion of Science","keywords":"Computer science; Nanopore sequencing; High fidelity; Fidelity; Throughput; DNA sequencing; Operating system; Engineering; Wireless; Telecommunications","authors":[{"name":"Yukiteru Ono","is_ca":false},{"name":"Michiaki Hamada","is_ca":false},{"name":"Kiyoshi Asai","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01094207650555211,"gpt":0.2306906389217082,"spread":0.2197485624161561,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001939683,0.0001176894,0.0001683937,0.00003508116,0.0001446885,0.00001594242,0.00009250332,0.00005433556,0.00000315455],"category_scores_gemma":[0.00001866957,0.0001102489,0.00005237097,0.00003010825,0.00006463929,0.000001078855,0.0002893498,0.00004535076,2.69194e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001175616,"about_ca_system_score_gemma":0.00004542494,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009415555,"about_ca_topic_score_gemma":0.00001303653,"domain_scores_codex":[0.9993767,0.000009986746,0.0002558019,0.0001301871,0.00006550111,0.0001618233],"domain_scores_gemma":[0.9995914,0.00002092368,0.0001257915,0.0001594577,0.00004917385,0.00005325833],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001479566,0.0004292693,0.0378825,0.001489705,0.001973802,0.000004538107,0.01138991,0.008140444,0.8237277,0.003429534,0.00621766,0.1038353],"study_design_scores_gemma":[0.005555527,0.005947318,0.01302777,0.0000249909,0.00041421,0.0001376391,0.00564544,0.0646591,0.07997321,0.001723969,0.8213864,0.001504439],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9955359,0.002903262,0.000594012,0.0000874671,0.00008986152,0.0002932575,0.0002834658,0.000001997186,0.0002107479],"genre_scores_gemma":[0.9922605,0.001182208,0.005990548,0.000290722,0.0000543862,0.00002063138,0.00007318408,0.00001388758,0.0001139498],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8151687,"threshold_uncertainty_score":0.4495822,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3016947700","doi":"10.1093/nargab/lqaa024","title":"A minimum reporting standard for multiple sequence alignments","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":76,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"Commonwealth Scientific and Industrial Research Organisation","keywords":"Completeness (order theory); Sequence (biology); Transparency (behavior); nobody; Computer science; Simple (philosophy); Data mining; Mathematics; Biology; Computer security; Genetics","authors":[{"name":"Thomas K. F. Wong","is_ca":false},{"name":"Subha Kalyaanamoorthy","is_ca":true},{"name":"Karen Meusemann","is_ca":false},{"name":"David Yeates","is_ca":false},{"name":"Bernhard Misof","is_ca":false},{"name":"Lars S. Jermiin","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04466610776664429,"gpt":0.2708863985655486,"spread":0.2262202907989043,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002127024,0.0001527891,0.000201356,0.00001515837,0.0001465586,0.00004581576,0.0001117119,0.00008829702,0.000001213519],"category_scores_gemma":[0.0002744563,0.000140872,0.00007830057,0.00003658636,0.00005559109,0.00000172292,0.0001604439,0.0000419805,0.000001972458],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001318466,"about_ca_system_score_gemma":0.00006973864,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000002671768,"about_ca_topic_score_gemma":0.000003918645,"domain_scores_codex":[0.9988943,0.000006344367,0.0005913138,0.0001981981,0.0000801036,0.0002297796],"domain_scores_gemma":[0.9992102,0.00001660246,0.000414841,0.0001526387,0.000082878,0.0001228499],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000342903,0.00002176551,0.00905928,0.0002870772,0.000218172,0.000002594365,0.002304829,0.0002007871,0.9580724,0.0001117198,0.003915299,0.02546323],"study_design_scores_gemma":[0.004236371,0.002754496,0.001406572,0.00003487189,0.0001355252,0.00006219574,0.003865041,0.06669134,0.2307993,0.000467188,0.6883497,0.001197343],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9895632,0.000589063,0.008198183,0.0004545423,0.00009824052,0.0004351753,0.0003068528,0.000006002649,0.0003487511],"genre_scores_gemma":[0.9434174,0.0006636926,0.05430816,0.001253177,0.0001927987,0.00002135177,0.00008412558,0.00001924697,0.00004000689],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.727273,"threshold_uncertainty_score":0.5744593,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3032790130","doi":"10.1093/nargab/lqaa035","title":"Where are G-quadruplexes located in the human transcriptome?","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":36,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; Université de Sherbrooke","keywords":"Transcriptome; Computational biology; Non-coding RNA; Biology; RNA splicing; Alternative splicing; RNA; Untranslated region; G-quadruplex; Nucleic acid structure; Gene; RNA-Seq; microRNA; Genetics; Gene expression; Messenger RNA; DNA","authors":[{"name":"Anaïs Vannutelli","is_ca":true},{"name":"Sarah Belhamiti","is_ca":true},{"name":"Jean‐Michel Garant","is_ca":true},{"name":"Aïda Ouangraoua","is_ca":true},{"name":"Jean‐Pierre Perreault","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01710935024960775,"gpt":0.2270001340014276,"spread":0.2098907837518198,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001932555,0.0001295862,0.0001295804,0.00002230435,0.00009976266,0.00006302826,0.0002256489,0.0001204122,0.00001044589],"category_scores_gemma":[0.00001410808,0.00009317725,0.00004818161,0.00007538281,0.00003760487,0.000005929218,0.00004140024,0.00008691943,0.000007855047],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00000526717,"about_ca_system_score_gemma":0.00002116959,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00000798911,"about_ca_topic_score_gemma":0.00003607923,"domain_scores_codex":[0.9993388,0.00003134314,0.0002493599,0.0001223891,0.00008966379,0.0001684244],"domain_scores_gemma":[0.9996611,0.000006733888,0.00008390029,0.0001660021,0.00002170443,0.00006051704],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0000731031,0.00005376299,0.002342903,0.0003869095,0.00005118288,0.000007361517,0.00539409,0.00005641234,0.9767789,0.0007935062,0.002388621,0.01167327],"study_design_scores_gemma":[0.003634442,0.00179137,0.006259435,0.0001817394,0.0001072557,0.00009702126,0.0315468,0.008754714,0.4399872,0.0007712932,0.5052656,0.001603104],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9941167,0.001310582,0.001792915,0.001713343,0.00003248709,0.0002920787,0.00003908306,0.000009767176,0.0006930889],"genre_scores_gemma":[0.9952216,0.001299219,0.001519534,0.001760914,0.00009724507,0.00001131686,0.00004636641,0.00001235235,0.0000315107],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5367916,"threshold_uncertainty_score":0.3799658,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3110037558","doi":"10.1093/nargab/lqaa098","title":"TOUCAN: a framework for fungal biosynthetic gene cluster discovery","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Microbial Natural Products and Biosynthesis","field":"Medicine","cited_by":36,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Concordia University; Université du Québec à Montréal","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Université du Québec à Montréal","keywords":"Biology; Computational biology; Scope (computer science); Computer science","authors":[{"name":"Hayda Almeida","is_ca":true},{"name":"Sylvester Palys","is_ca":true},{"name":"Adrian Tsang","is_ca":true},{"name":"Abdoulaye Baniré Diallo","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02021586546052363,"gpt":0.2393252859033765,"spread":0.2191094204428528,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001254325,0.0001748756,0.0003037911,0.0000442158,0.0001068999,0.0001046803,0.00008695835,0.0001607987,0.000024149],"category_scores_gemma":[0.0002132198,0.0001227144,0.0001133689,0.0001060025,0.00006515248,0.0001261106,0.00008572509,0.0001714907,0.00002069784],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003335489,"about_ca_system_score_gemma":0.00007321173,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00000254205,"about_ca_topic_score_gemma":7.931511e-7,"domain_scores_codex":[0.9991142,0.000006621503,0.0003487145,0.0001779744,0.0001017243,0.0002507446],"domain_scores_gemma":[0.9993963,0.00007609214,0.0001087971,0.0001757371,0.00006125811,0.0001818226],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.006765617,0.0005321872,0.001215103,0.009913179,0.001203291,0.00004286566,0.01543056,0.00004639752,0.493205,0.01399273,0.05162929,0.4060237],"study_design_scores_gemma":[0.004382943,0.002733241,0.0005266552,0.0003941217,0.0007950053,0.0004509915,0.002085057,0.06184226,0.3664204,0.001807404,0.5571312,0.001430706],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9370061,0.001121539,0.02210792,0.03694769,0.0003488349,0.001497252,0.0005052956,0.0000745588,0.0003908097],"genre_scores_gemma":[0.7532796,0.0007378689,0.2304297,0.01431249,0.0008647388,0.000002528094,0.00005803795,0.00003192806,0.0002831343],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5055019,"threshold_uncertainty_score":0.5004146,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3129861733","doi":"10.1093/nargab/lqab011","title":"Single-cell mapper (scMappR): using scRNA-seq to infer the cell-type specificities of differentially expressed genes","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":34,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Princess Margaret Cancer Centre; University of Toronto; SickKids Foundation; University Health Network; Canadian Institute for Advanced Research; Vector Institute","funders":"Canadian Institutes of Health Research; Hospital for Sick Children; Natural Sciences and Engineering Research Council of Canada; Ontario Ministry of Research, Innovation and Science","keywords":"RNA-Seq; RNA; Biology; Cell type; Computational biology; Gene expression; Gene; Population; Cell; Genetics; Transcriptome","authors":[{"name":"Dustin Sokolowski","is_ca":true},{"name":"Mariela Faykoo-Martinez","is_ca":true},{"name":"Lauren Erdman","is_ca":true},{"name":"Huayun Hou","is_ca":true},{"name":"Cadia Chan","is_ca":true},{"name":"Helen Zhu","is_ca":true},{"name":"Melissa M. Holmes","is_ca":true},{"name":"Anna Goldenberg","is_ca":true},{"name":"Michael D. Wilson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02578047590593834,"gpt":0.2153481003316333,"spread":0.189567624425695,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001227001,0.0002229137,0.0002429176,0.00005119816,0.0001499445,0.0001132864,0.0002221361,0.0001696223,0.00002224511],"category_scores_gemma":[0.00001793584,0.0001745594,0.0001058025,0.0001197101,0.000109766,0.000009550622,0.0002018094,0.0001008919,0.000004932998],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001791009,"about_ca_system_score_gemma":0.0001454727,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001709937,"about_ca_topic_score_gemma":0.00001968836,"domain_scores_codex":[0.9988781,0.00003163853,0.000465545,0.0001977311,0.0001569527,0.0002699835],"domain_scores_gemma":[0.9991347,0.00002063576,0.0001470373,0.0003905957,0.000208648,0.00009836639],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00007152117,0.0001055428,0.0005265069,0.0001393949,0.00003504162,0.000001440811,0.0008570526,0.0003809005,0.994944,0.00003166779,0.000320573,0.00258635],"study_design_scores_gemma":[0.0005132701,0.0002053822,0.0001229662,0.00002277848,0.00004846488,0.00001280952,0.001503679,0.002893176,0.9634228,0.00003282783,0.03094761,0.0002741817],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9885974,0.001474636,0.007716518,0.00007481009,0.0003389402,0.0001908053,0.00007488985,0.000007797022,0.001524207],"genre_scores_gemma":[0.9732398,0.00142142,0.02380553,0.0004890875,0.0002340743,0.000002682481,0.0001215133,0.0000358655,0.0006500068],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03152115,"threshold_uncertainty_score":0.7118328,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4206004774","doi":"10.1093/nargab/lqab122","title":"A comparison on predicting functional impact of genomic variants","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Rare Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":32,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"National Key Research and Development Program of China","keywords":"Single-nucleotide polymorphism; Missense mutation; Genomics; Human genome; Functional genomics; SNP","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.01467209410687007,"gpt":0.2587007337804651,"spread":0.244028639673595,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000179576,0.0001355526,0.0001682583,0.00007375262,0.0002271228,0.00002239201,0.0001302637,0.00005264839,0.00007446003],"category_scores_gemma":[0.00001536435,0.0001228824,0.0001287893,0.00005223996,0.0000423073,0.000003470783,0.0002543289,0.0001035388,0.000005053489],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005575887,"about_ca_system_score_gemma":0.000181795,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001549236,"about_ca_topic_score_gemma":0.000001501502,"domain_scores_codex":[0.9991868,0.00002159598,0.0003419013,0.0001429946,0.0001278334,0.0001788821],"domain_scores_gemma":[0.9994465,0.00001284617,0.0002115052,0.0002041701,0.00004033,0.00008465695],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","study_design_scores_codex":[0.002965881,0.001431263,0.1676575,0.0002888046,0.001454385,0.00001365346,0.003129795,0.1110661,0.6608862,0.001053118,0.02373825,0.02631509],"study_design_scores_gemma":[0.008230046,0.01484919,0.5519726,0.00004310327,0.0003493431,0.0006890295,0.006393644,0.3424447,0.01220047,0.0004912027,0.0601657,0.002171021],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9975978,0.0003788065,0.0002389278,0.00001660843,0.0001400051,0.000161874,0.0004766779,0.000005890998,0.0009834771],"genre_scores_gemma":[0.9983737,0.000107203,0.0008446917,0.000117238,0.00008784464,0.000009702485,0.0003635516,0.00001503704,0.00008096707],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6486857,"threshold_uncertainty_score":0.5010998,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3089247545","doi":"10.1093/nargab/lqaa066","title":"The expression of long non-coding RNAs is associated with H3Ac and H3K4me2 changes regulated by the HDA6-LDL1/2 histone modification complex in <i>Arabidopsis</i>","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Cancer-related molecular mechanisms research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":31,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University; Agriculture and Agri-Food Canada","funders":"Agriculture and Agri-Food Canada; National Taiwan University; Ministry of Science and Technology of the People's Republic of China","keywords":"Biology; Long non-coding RNA; Arabidopsis; Histone; Histone deacetylase; Regulation of gene expression; Genetics; RNA splicing; Transcription factor; Gene expression; Alternative splicing; Gene; RNA; Cell biology; Messenger RNA; Mutant","authors":[{"name":"Fu‐Yu Hung","is_ca":true},{"name":"Chen Chen","is_ca":true},{"name":"Ming‐Ren Yen","is_ca":false},{"name":"Jo‐Wei Allison Hsieh","is_ca":false},{"name":"Chenlong Li","is_ca":true},{"name":"Yuan‐Hsin Shih","is_ca":false},{"name":"Fang-Fang Chen","is_ca":false},{"name":"Pao‐Yang Chen","is_ca":false},{"name":"Yuhai Cui","is_ca":true},{"name":"Keqiang Wu","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01933077866174409,"gpt":0.253167448844162,"spread":0.2338366701824179,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002927667,0.0001350757,0.0001560712,0.0000287595,0.0001786376,0.00005254753,0.0001810442,0.0001405631,0.000002152473],"category_scores_gemma":[0.00004930587,0.00008699091,0.00002544504,0.0001556659,0.0001569383,0.000007395352,0.0001523328,0.0001480722,5.070677e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003597192,"about_ca_system_score_gemma":0.00005388422,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001949467,"about_ca_topic_score_gemma":0.00004621494,"domain_scores_codex":[0.9991521,0.00003839748,0.0002652399,0.0001619224,0.0001773219,0.000205],"domain_scores_gemma":[0.9993853,0.00002771392,0.0002056037,0.0002190768,0.00009188867,0.00007043603],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001496609,0.00001696762,0.00002734291,0.00004782427,0.00004230937,7.735322e-7,0.001227884,0.00007758209,0.9889562,0.00001270317,0.002279883,0.007160932],"study_design_scores_gemma":[0.001206491,0.0003890343,0.0004162789,0.00005898449,0.00002806893,0.00001182368,0.0008895661,0.05684981,0.9333024,0.00001688511,0.006628284,0.0002023755],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9777418,0.002346358,0.01557462,0.003189716,0.00003483387,0.0007637694,0.000127236,0.00001040689,0.0002111986],"genre_scores_gemma":[0.9957553,0.002912063,0.000445715,0.0005889251,0.00002311156,0.00001446287,0.0001732552,0.00001814824,0.00006903797],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.05677223,"threshold_uncertainty_score":0.3547387,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3001111520","doi":"10.1093/nargab/lqaa002","title":"Bayesian correlation is a robust gene similarity measure for single-cell RNA-seq data","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":24,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital; University of Ottawa","funders":"University of Bern; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Similarity (geometry); Bayesian probability; Correlation; Pearson product-moment correlation coefficient; Mathematics; Artificial intelligence; RNA-Seq; Similarity measure; Pattern recognition (psychology); Computational biology; Computer science; Gene; Statistics; Biology; Gene expression; Transcriptome; Genetics","authors":[{"name":"Daniel Sánchez‐Taltavull","is_ca":false},{"name":"Theodore J. Perkins","is_ca":true},{"name":"Noëlle Dommann","is_ca":false},{"name":"Nicolas Mélin","is_ca":false},{"name":"Adrian Keogh","is_ca":false},{"name":"Daniel Candinas","is_ca":false},{"name":"Deborah Stroka","is_ca":false},{"name":"Guido Beldi","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0596890529347593,"gpt":0.2264212591935154,"spread":0.1667322062587561,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002103468,0.0002247615,0.0002169243,0.0000282471,0.0001738675,0.0001073813,0.0003541421,0.0002744825,0.000007250058],"category_scores_gemma":[0.00006507085,0.0002193361,0.00008265318,0.00007640722,0.00006201374,0.00002264227,0.0001816147,0.0001287958,0.000005066565],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001908356,"about_ca_system_score_gemma":0.000108613,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000008907346,"about_ca_topic_score_gemma":0.00001633594,"domain_scores_codex":[0.9988462,0.00001539847,0.000416419,0.0003250152,0.0001325102,0.0002644153],"domain_scores_gemma":[0.9990831,0.00001663861,0.0001656956,0.0004371335,0.0001103388,0.0001870638],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005669924,0.0002267993,0.001289614,0.0005913738,0.0001702044,0.000002336572,0.002048346,0.001568281,0.9480886,0.00004066479,0.01507573,0.03033108],"study_design_scores_gemma":[0.002512846,0.00105121,0.00003918904,0.00001905934,0.0001620592,0.00002058319,0.0003725102,0.7468775,0.1655339,0.00005887504,0.08268502,0.0006672412],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1946941,0.001258517,0.7971661,0.00169213,0.0003725665,0.0009540027,0.00196689,0.00004577518,0.001849982],"genre_scores_gemma":[0.8793997,0.0006515055,0.1118795,0.004312526,0.0005782457,0.000008865947,0.003006019,0.00005949318,0.0001041539],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7825547,"threshold_uncertainty_score":0.8944268,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4206771039","doi":"10.1093/nargab/lqab123","title":"FILER: a framework for harmonizing and querying large-scale functional genomics knowledge","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":20,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Institute on Aging; Weston Brain Institute","keywords":"Scale (ratio); Genomics; Computer science; Functional genomics; Data science; Computational biology; Biology; Genome; Geography; Genetics; Cartography; Gene","authors":[{"name":"Pavel P. Kuksa","is_ca":false},{"name":"Yuk Yee Leung","is_ca":false},{"name":"Prabhakaran Gangadharan","is_ca":false},{"name":"Živadin Katanić","is_ca":false},{"name":"Lauren Kleidermacher","is_ca":false},{"name":"Alexandre Amlie‐Wolf","is_ca":false},{"name":"Chien‐Yueh Lee","is_ca":false},{"name":"Liming Qu","is_ca":false},{"name":"Emily Greenfest‐Allen","is_ca":false},{"name":"Otto Valladares","is_ca":false},{"name":"Li‐San Wang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02949434952622089,"gpt":0.253449076832666,"spread":0.2239547273064452,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002599908,0.0001233847,0.0001141276,0.00005866109,0.0005279731,0.00005782668,0.00009454595,0.00009720494,0.00002039949],"category_scores_gemma":[0.00002274919,0.0001266418,0.00005375646,0.00006520809,0.00003450305,0.000007038958,0.0002661897,0.0001220558,0.00000171117],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004422739,"about_ca_system_score_gemma":0.0001088129,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":9.186356e-7,"about_ca_topic_score_gemma":0.000003912182,"domain_scores_codex":[0.9992756,0.00001731915,0.0002396605,0.0001931835,0.0000735645,0.00020073],"domain_scores_gemma":[0.9995478,0.00002125166,0.0001233149,0.0001766307,0.00004946756,0.00008147053],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001120972,0.0004834605,0.0030102,0.0007929741,0.0003341501,0.000001183369,0.01530901,0.001402359,0.6901258,0.01367811,0.1001508,0.1735911],"study_design_scores_gemma":[0.0009210825,0.0002275661,0.0004949517,0.00001074988,0.000028697,0.00003029364,0.004579325,0.03846842,0.008711496,0.001111632,0.945112,0.0003037577],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7343967,0.004823918,0.2573004,0.0006148286,0.0006847164,0.0006641831,0.0005007182,0.00002844986,0.0009861079],"genre_scores_gemma":[0.8398544,0.002745572,0.1516568,0.002588888,0.0006680102,0.0003300933,0.001043858,0.00006771441,0.001044662],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8449613,"threshold_uncertainty_score":0.5164302,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4392844094","doi":"10.1093/nargab/lqae005","title":"Functional domain annotation by structural similarity","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":19,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"Canadian Institutes of Health Research","keywords":"Annotation; Computational biology; Structural similarity; In silico; Protein domain; Domain (mathematical analysis); Similarity (geometry); Structural alignment; Sequence alignment; Proteome; UniProt; Biology; Sequence (biology); Protein sequencing; Benchmark (surveying); Computer science; Bioinformatics; Genetics; Peptide sequence; Artificial intelligence; Geography","authors":[{"name":"Poorya Mirzavand Borujeni","is_ca":true},{"name":"Reza Salavati","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.007423367596102924,"gpt":0.2145816978091313,"spread":0.2071583302130283,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001240809,0.0001366403,0.00009596797,0.0000331078,0.0001368141,0.0001010268,0.00006290289,0.0001029455,0.000008957792],"category_scores_gemma":[0.000008404623,0.0001183947,0.00005125804,0.00005139911,0.00006901922,0.00000281225,0.00008944294,0.00007163046,0.000008939202],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001659628,"about_ca_system_score_gemma":0.00004992689,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004626226,"about_ca_topic_score_gemma":0.00000609886,"domain_scores_codex":[0.9993743,0.000009118638,0.0002129109,0.0001597689,0.00008081996,0.0001630992],"domain_scores_gemma":[0.9997356,0.00001033474,0.00003837921,0.0001181725,0.00003904564,0.00005846352],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001108648,0.00002850975,0.002288661,0.0003228757,0.0004314052,0.000003333718,0.001752411,0.0002585745,0.8467063,0.003203998,0.07677342,0.0681197],"study_design_scores_gemma":[0.001047937,0.0005647213,0.009281103,0.00002698504,0.00008542697,0.0001408026,0.001395966,0.02974894,0.03026578,0.005683051,0.9208325,0.0009268547],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9897022,0.005393073,0.003058172,0.0002952211,0.0003444497,0.0001283699,0.0002473589,0.000008429623,0.0008227013],"genre_scores_gemma":[0.9894115,0.001330721,0.007957944,0.0003786612,0.0002442732,0.000007161283,0.00036528,0.0000165605,0.0002879389],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.844059,"threshold_uncertainty_score":0.4827995,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4289712796","doi":"10.1093/nargab/lqac057","title":"Mining bacterial NGS data vastly expands the complete genomes of temperate phages","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Bacteriophages and microbial interactions","field":"Environmental Science","cited_by":19,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Guelph; Simon Fraser University; University of British Columbia","funders":"National Institutes of Health; National Key Research and Development Program of China; Hebei Provincial Key Research Projects; U.S. National Library of Medicine; National Natural Science Foundation of China","keywords":"Bacterial genome size; Genome; Temperate climate; Biology; Computational biology; Evolutionary biology; Data science; Computer science; Genetics; Ecology; Gene","authors":[{"name":"Xianglilan Zhang","is_ca":false},{"name":"Ruohan Wang","is_ca":false},{"name":"Xiangcheng Xie","is_ca":false},{"name":"Yunjia Hu","is_ca":false},{"name":"Jianping Wang","is_ca":false},{"name":"Qiang Sun","is_ca":false},{"name":"Xikang Feng","is_ca":false},{"name":"Shanwei Tong","is_ca":true},{"name":"Wei Yan","is_ca":false},{"name":"Huiqi Wen","is_ca":false},{"name":"Mengyao Wang","is_ca":false},{"name":"Shixiang Zhai","is_ca":false},{"name":"Cheng Sun","is_ca":false},{"name":"Fangyi Wang","is_ca":false},{"name":"Qi Niu","is_ca":false},{"name":"Andrew M. Kropinski","is_ca":true},{"name":"Yujun Cui","is_ca":false},{"name":"Xiaofang Jiang","is_ca":false},{"name":"Shaoliang Peng","is_ca":false},{"name":"Shuai Cheng Li","is_ca":false},{"name":"Yigang Tong","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03181369653655028,"gpt":0.2370182486495584,"spread":0.2052045521130082,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0002793276,0.00009693257,0.0001296221,0.00002240554,0.0004361524,0.00007398361,0.0004387923,0.00002035768,0.001305719],"category_scores_gemma":[0.000005558555,0.00007207266,0.00003081448,0.00008631877,0.0001240226,0.0002323875,0.001342103,0.0001049455,0.00002007714],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004780725,"about_ca_system_score_gemma":0.00001663006,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008741548,"about_ca_topic_score_gemma":0.00002778544,"domain_scores_codex":[0.9992994,0.00002375121,0.0003051537,0.0001144244,0.0001029438,0.0001543245],"domain_scores_gemma":[0.9994037,0.00003591308,0.0001714118,0.0003452994,0.000005739877,0.00003792019],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00007717228,0.00008657646,0.000908748,0.00004129412,0.00008348624,0.000002160025,0.01235697,0.0003969886,0.917134,0.00003751049,0.01100761,0.05786742],"study_design_scores_gemma":[0.0008212293,0.000287607,0.01188178,0.000009754047,0.00007917752,0.000162468,0.009766753,0.03686192,0.003122355,0.00002529933,0.9366094,0.0003722865],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9974068,0.00006407665,0.00007416951,0.0002349149,0.0002547113,0.0001470172,0.0007132185,0.000008651527,0.001096484],"genre_scores_gemma":[0.9922462,0.000261908,0.00644776,0.0003756626,0.00005941499,0.000006776918,0.0002663664,0.000011285,0.0003246505],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9256018,"threshold_uncertainty_score":0.9996072,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4214852607","doi":"10.1093/nargab/lqac010","title":"G-quadruplex occurrence and conservation: more than just a question of guanine–cytosine content","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"DNA and Nucleic Acid Chemistry","field":"Biochemistry, Genetics and Molecular Biology","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"Fonds de recherche du Québec – Nature et technologies; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Guanine; Cytosine; G-quadruplex; Content (measure theory); Chemistry; DNA; Mathematics; Biochemistry; Gene; Mathematical analysis","authors":[{"name":"Anaïs Vannutelli","is_ca":true},{"name":"Jean‐Pierre Perreault","is_ca":true},{"name":"Aïda Ouangraoua","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02089995656188113,"gpt":0.2418839903462985,"spread":0.2209840337844174,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001535272,0.0001048587,0.0001242134,0.0000233636,0.0001214028,0.00001705225,0.0000947613,0.00005880575,0.00001240776],"category_scores_gemma":[0.00002428055,0.000101303,0.00003043686,0.00005044604,0.000113194,0.000006103194,0.0002161985,0.00007456717,4.576046e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001356463,"about_ca_system_score_gemma":0.0000518116,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001710062,"about_ca_topic_score_gemma":0.000002679817,"domain_scores_codex":[0.9994032,0.0000107756,0.0002607122,0.0001125752,0.0001001464,0.0001125754],"domain_scores_gemma":[0.9995624,0.00000727132,0.0001602166,0.0001542175,0.00006074896,0.00005510985],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0003339809,0.00014707,0.02006092,0.0005135091,0.0001218908,0.000003646128,0.001780324,0.00008101481,0.9516412,0.0006477706,0.009308203,0.01536047],"study_design_scores_gemma":[0.006164173,0.002432419,0.01982961,0.0001072477,0.0002643759,0.0008462613,0.02481801,0.04999952,0.5608061,0.0001989304,0.3328307,0.001702658],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9979165,0.0007187824,0.0002818938,0.000393419,0.00005864454,0.0001091368,0.0002649784,0.000006088749,0.0002505567],"genre_scores_gemma":[0.9959711,0.0005870505,0.00245445,0.000391544,0.00003841671,0.000009027929,0.0003967132,0.00000653523,0.0001451068],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3908351,"threshold_uncertainty_score":0.4131016,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2964367788","doi":"10.1093/nargab/lqz003","title":"Open chromatin structure in PolyQ disease-related genes: a potential mechanism for CAG repeat expansion in the normal human population","year":2019,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":15,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Azrieli Foundation; Israel Science Foundation","keywords":"Chromatin; Gene; Biology; Epigenetics; Genetics; Genome; ChIA-PET; Population; Chromatin remodeling","authors":[{"name":"Matan Sorek","is_ca":false},{"name":"Lea R Z Cohen","is_ca":false},{"name":"Eran Meshorer","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.00761182982789455,"gpt":0.2540048161247481,"spread":0.2463929862968536,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003119403,0.0001200829,0.0001347732,0.00007178931,0.0001219345,0.0001185825,0.0002913426,0.0001267443,0.000007237966],"category_scores_gemma":[0.00002033025,0.00008850286,0.00004264835,0.00007368912,0.00001795518,0.00002341465,0.0001908352,0.00009652572,0.000001527389],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000279831,"about_ca_system_score_gemma":0.00005563412,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002116973,"about_ca_topic_score_gemma":0.0003006272,"domain_scores_codex":[0.9991199,0.00003327834,0.0003405178,0.0001568315,0.0001148908,0.0002345847],"domain_scores_gemma":[0.9995513,0.000007455315,0.0001004537,0.0002506781,0.00003248456,0.00005765106],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002822618,0.00005106949,0.006973095,0.0001993477,0.00002684882,0.000004980589,0.0008978452,0.000811689,0.982035,0.001648323,0.0001054212,0.006964071],"study_design_scores_gemma":[0.01747927,0.002856322,0.1752638,0.0003371172,0.0001000172,0.0001599405,0.0106241,0.7085329,0.0694909,0.01144683,0.001784992,0.001923721],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9983624,0.0001284578,0.0002008662,0.0001141263,0.00004960461,0.001007561,0.00006587157,0.000002804223,0.00006835131],"genre_scores_gemma":[0.9972115,0.000111513,0.001601381,0.000186605,0.00003825006,0.00001758174,0.0006830694,0.00001174858,0.0001383065],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9125441,"threshold_uncertainty_score":0.3609042,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4211146659","doi":"10.1093/nargab/lqab125","title":"An improved molecular inversion probe based targeted sequencing approach for low variant allele frequency","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Cancer Genomics and Diagnostics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Toronto General Hospital; University of Toronto; Princess Margaret Cancer Centre; University Health Network","funders":"Rising Tide Foundation; Applebaum Foundation; Princess Margaret Cancer Foundation","keywords":"Allele frequency; Allele; Genetics; Inversion (geology); Computational biology; Biology; Paleontology; Gene","authors":[{"name":"Tamir Biezuner","is_ca":false},{"name":"Yardena Brilon","is_ca":false},{"name":"Asaf Ben Arye","is_ca":false},{"name":"Barak Oron","is_ca":false},{"name":"Aditee Kadam","is_ca":false},{"name":"Adi Danin","is_ca":false},{"name":"Nili Furer","is_ca":false},{"name":"Mark D. Minden","is_ca":true},{"name":"Dennis Dong Hwan Kim","is_ca":true},{"name":"Shiran Shapira","is_ca":false},{"name":"Nadir Arber","is_ca":false},{"name":"John E. Dick","is_ca":true},{"name":"Paaladinesh Thavendiranathan","is_ca":true},{"name":"Yoni Moskovitz","is_ca":false},{"name":"Nathali Kaushansky","is_ca":false},{"name":"Noa Chapal-Ilani","is_ca":false},{"name":"Liran I. Shlush","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.007255798986068817,"gpt":0.2040845625466862,"spread":0.1968287635606174,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000387085,0.0002267857,0.0002022328,0.00008236592,0.0003844125,0.00008496959,0.0002869929,0.0001322929,0.00001109758],"category_scores_gemma":[0.00004330495,0.0002386094,0.0001061209,0.0001006206,0.00005380197,0.0000122841,0.0001982335,0.0001393058,6.015651e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001482345,"about_ca_system_score_gemma":0.0004716517,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000443672,"about_ca_topic_score_gemma":0.000007676655,"domain_scores_codex":[0.9987772,0.00003230918,0.0003823807,0.000321181,0.00013454,0.0003524089],"domain_scores_gemma":[0.9990939,0.00001460159,0.0002005221,0.0004251287,0.0001016473,0.0001642348],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001969915,0.0001627071,0.00009154027,0.0002272552,0.00007628889,0.00000451802,0.000456731,0.01712689,0.9769672,0.0007690901,0.0008089374,0.00311187],"study_design_scores_gemma":[0.001899589,0.001561585,0.00001633037,0.000004395451,0.00005855114,0.00003138073,0.001071784,0.9215221,0.06439956,0.0005302168,0.008353868,0.0005505928],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5102364,0.0008345564,0.4854853,0.0001490057,0.0002210689,0.001573757,0.001004381,0.00003436422,0.0004611577],"genre_scores_gemma":[0.632171,0.0001193806,0.361926,0.001913675,0.0001234832,0.0001783811,0.003491437,0.00005060981,0.00002600391],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9125676,"threshold_uncertainty_score":0.973021,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3038449881","doi":"10.1093/nargab/lqaa046","title":"On the prediction of DNA-binding preferences of C2H2-ZF domains using structural models: application on human CTCF","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"BC Children's Hospital; University of British Columbia","funders":"European Regional Development Fund; Ministerio de Ciencia e Innovación","keywords":"Zinc finger; CTCF; DNA; Transcription factor; Computational biology; Binding site; Biology; DNA-binding protein; Chromatin; Genetics; Gene","authors":[{"name":"Alberto Meseguer","is_ca":false},{"name":"Filip Årman","is_ca":false},{"name":"Oriol Fornés","is_ca":true},{"name":"Rubén Molina-Fernández","is_ca":false},{"name":"Jaume Bonet","is_ca":false},{"name":"Narcís Fernández‐Fuentes","is_ca":false},{"name":"Baldo Oliva","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04841538685684868,"gpt":0.2496151709710039,"spread":0.2011997841141553,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001197878,0.00009659462,0.0001149264,0.00002416112,0.0001036971,0.00001466738,0.0001256864,0.00009457122,0.000002879774],"category_scores_gemma":[0.00001725234,0.00006691779,0.00004197163,0.00004401351,0.00005012449,0.000006859525,0.00004869004,0.0000494138,6.008881e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000008625849,"about_ca_system_score_gemma":0.00002237644,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004544407,"about_ca_topic_score_gemma":0.00000119085,"domain_scores_codex":[0.9994147,0.00001899854,0.0002734354,0.00009843292,0.0001078213,0.00008664189],"domain_scores_gemma":[0.999529,0.00001528464,0.000231873,0.000154125,0.00003532717,0.00003436852],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.000055377,0.000008398561,0.00004368558,0.00006542575,0.00003043671,2.985455e-8,0.0005715993,0.0009022821,0.9890482,0.005876251,0.00001891926,0.003379432],"study_design_scores_gemma":[0.0002443378,0.0006405478,0.00005441788,0.00003117851,0.00002557786,0.000001852384,0.0006623276,0.1018816,0.892814,0.003388271,0.0001531374,0.0001026885],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9917113,0.00003186824,0.007353695,0.00005093926,0.0000242948,0.0002572371,0.0001060549,0.000004005415,0.0004606448],"genre_scores_gemma":[0.9968095,0.00007794092,0.002917106,0.00008256146,0.00004825135,0.000004639796,0.00004920793,0.000007111232,0.00000368129],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1009794,"threshold_uncertainty_score":0.2728828,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3164889407","doi":"10.1093/nargab/lqab037","title":"Detection of homozygous and hemizygous complete or partial exon deletions by whole-exome sequencing","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomic variations and chromosomal abnormalities","field":"Biochemistry, Genetics and Molecular Biology","cited_by":13,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Center for Research Resources; National Institutes of Health; Fonds de Recherche en Santé Respiratoire; Hospital for Sick Children; Institut National de la Santé et de la Recherche Médicale; SCOR Corporate Foundation for Science; Institut des maladies génétiques Imagine; St. Giles Foundation; Université de Paris; Rockefeller University; Yale University; National Center for Advancing Translational Sciences; National Human Genome Research Institute; Agence Nationale de la Recherche; National Institute of Allergy and Infectious Diseases; Howard Hughes Medical Institute","keywords":"Exome sequencing; Exon; Genetics; Biology; Computational biology; Exome; Copy-number variation; Gene; Mutation; Genome","authors":[{"name":"Benedetta Bigio","is_ca":false},{"name":"Yoann Seeleuthner","is_ca":false},{"name":"Gaspard Kerner","is_ca":false},{"name":"Mélanie Migaud","is_ca":false},{"name":"Jérémie Rosain","is_ca":false},{"name":"Bertrand Boisson","is_ca":false},{"name":"Carla Nasca","is_ca":false},{"name":"Anne Puel","is_ca":false},{"name":"Jacinta Bustamante","is_ca":false},{"name":"Jean‐Laurent Casanova","is_ca":false},{"name":"Laurent Abel","is_ca":false},{"name":"Aurélie Cobat","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01297703630843684,"gpt":0.2067800720448492,"spread":0.1938030357364123,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001025628,0.0001270484,0.0001645598,0.00003063525,0.0001605703,0.00005161333,0.00005838561,0.000120041,0.000009224273],"category_scores_gemma":[0.00002652638,0.0001156809,0.00004230574,0.00006581244,0.00008306733,0.00001099741,0.0001125408,0.00006502837,0.000002093444],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002525963,"about_ca_system_score_gemma":0.0001194606,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002675683,"about_ca_topic_score_gemma":0.00005906836,"domain_scores_codex":[0.9992536,0.00001901268,0.0003449834,0.0001390481,0.00007170565,0.0001717008],"domain_scores_gemma":[0.9995108,0.00001317388,0.0001333499,0.0001724964,0.00009810007,0.00007212194],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00005009042,0.00002918037,0.000193406,0.0001430734,0.00006839733,0.000004503248,0.0007053027,0.00007087647,0.9919031,0.00009833416,0.0002143339,0.006519455],"study_design_scores_gemma":[0.001888176,0.0008603061,0.0004925348,0.00004650935,0.0001350108,0.001625904,0.00756713,0.01555801,0.8319003,0.0001755481,0.1390812,0.0006693436],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9910771,0.0009060445,0.007189871,0.00006743935,0.00007281495,0.0001050014,0.0003413194,0.000007667777,0.0002327452],"genre_scores_gemma":[0.9927703,0.001161625,0.005273533,0.0001372858,0.00006956229,0.000007203274,0.0002699058,0.00001160757,0.0002989442],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1600028,"threshold_uncertainty_score":0.4717332,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4391994538","doi":"10.1093/nargab/lqae018","title":"CDBProm: the Comprehensive Directory of Bacterial Promoters","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Izaak Walton Killam Health Centre; Dalhousie University","funders":"Dirección General de Asuntos del Personal Académico, Universidad Nacional Autónoma de México; Research Nova Scotia; Universidad Nacional Autónoma de México; Dalhousie Medical Research Foundation; Dalhousie University; Canadian Institutes of Health Research; Li Ka Shing Foundation","keywords":"Promoter; Genome; Bacterial genome size; Biology; Computational biology; Genetics; Gene; Bacterial artificial chromosome; DNA sequencing; In silico; Directory; Computer science; Gene expression","authors":[{"name":"Gustavo Sganzerla Martinez","is_ca":true},{"name":"Ernesto Pérez‐Rueda","is_ca":false},{"name":"Anuj Kumar","is_ca":true},{"name":"Mansi Dutt","is_ca":true},{"name":"C. Maya","is_ca":false},{"name":"Leonardo Ledesma","is_ca":false},{"name":"Pedro Lenz Casa","is_ca":false},{"name":"Aditya Kumar","is_ca":true},{"name":"Scheila de Ávila e Silva","is_ca":false},{"name":"David J. Kelvin","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01305627588397154,"gpt":0.2351026602278646,"spread":0.222046384343893,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001107208,0.0001194166,0.0001249705,0.00002881756,0.00007561206,0.0000462332,0.0001143579,0.00006584923,0.000005028532],"category_scores_gemma":[0.000009138992,0.0000797762,0.0000657058,0.000047492,0.0001463366,0.000001239515,0.0001452854,0.00005611575,0.000003691828],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00000743691,"about_ca_system_score_gemma":0.0000595677,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005675889,"about_ca_topic_score_gemma":0.000004160218,"domain_scores_codex":[0.9994423,0.00001349636,0.0002323113,0.0001156162,0.00006292976,0.000133392],"domain_scores_gemma":[0.9996601,0.00002030799,0.00005629536,0.0001827904,0.00004665464,0.00003387392],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005123231,0.0000192851,0.0002060902,0.0002942218,0.0002882966,0.000001192387,0.002057753,0.00003549158,0.9592586,0.0004570855,0.003598394,0.03373235],"study_design_scores_gemma":[0.0004890012,0.0005241609,0.003595556,0.00005363549,0.0001076557,0.00005841663,0.001304454,0.005139802,0.08962239,0.0003281146,0.8983644,0.0004123861],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9918693,0.006272159,0.000164998,0.0001295576,0.0003942292,0.0002002703,0.00008818886,0.000004320326,0.000876934],"genre_scores_gemma":[0.993361,0.003530567,0.002549842,0.0001296835,0.0002163871,0.00000781871,0.00004303424,0.00001553875,0.0001461096],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.894766,"threshold_uncertainty_score":0.3253179,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4292615690","doi":"10.1093/nargab/lqac058","title":"A computational approach to rapidly design peptides that detect SARS-CoV-2 surface protein S","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"SARS-CoV-2 and COVID-19 Research","field":"Medicine","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Agriculture and Agri-Food Canada; Ottawa Hospital; University of Regina; Carleton University; University of Ottawa","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Coronavirus disease 2019 (COVID-19); 2019-20 coronavirus outbreak; Computational biology; Pandemic; Peptide; Coronavirus; Surface plasmon resonance; Biology; Virology; Computer science; Medicine; Biochemistry; Disease; Nanotechnology; Infectious disease (medical specialty); Pathology; Materials science","authors":[{"name":"Maryam Hajikarimlou","is_ca":true},{"name":"Mohsen Hooshyar","is_ca":true},{"name":"Mohamed Taha Moutaoufik","is_ca":true},{"name":"Khaled A. Aly","is_ca":true},{"name":"Taha Azad","is_ca":true},{"name":"Sarah Takallou","is_ca":true},{"name":"Sasi Kumar Jagadeesan","is_ca":true},{"name":"Sadhna Phanse","is_ca":true},{"name":"Kamaleldin B. Said","is_ca":true},{"name":"Bahram Samanfar","is_ca":true},{"name":"John C. Bell","is_ca":true},{"name":"Frank Dehne","is_ca":true},{"name":"Mohan Babu","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06575502002795491,"gpt":0.3012700700699599,"spread":0.235515050042005,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008270679,0.0001988578,0.0002965267,0.0001958706,0.0003439202,0.0001079,0.0002074233,0.00006961132,0.000003911372],"category_scores_gemma":[0.00007023427,0.0001773009,0.00008383995,0.0003308381,0.00007265033,0.00009691171,0.0002889585,0.0003193078,0.00003776175],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002179991,"about_ca_system_score_gemma":0.0005700398,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004425444,"about_ca_topic_score_gemma":0.00000261613,"domain_scores_codex":[0.9984398,0.00006115199,0.0003595391,0.0002103703,0.0005566064,0.000372567],"domain_scores_gemma":[0.9993864,0.000102043,0.0001032283,0.00024349,0.00009214795,0.00007263164],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.005859821,0.001648844,0.005116261,0.00382116,0.00102621,0.0001351879,0.03461766,0.01557564,0.8054624,0.001939818,0.027966,0.09683103],"study_design_scores_gemma":[0.003470935,0.001883652,0.0002875887,0.00005443679,0.00007052117,0.0004487055,0.003767367,0.6855524,0.1695926,0.0008825416,0.1333174,0.0006718508],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9438411,0.0002724324,0.05116668,0.0002454671,0.00004364816,0.001794441,0.00005073915,0.00006305819,0.002522411],"genre_scores_gemma":[0.6594419,0.00000896282,0.3123108,0.02791445,0.0000622056,0.0001327824,0.0000309134,0.00004356097,0.00005442698],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6699767,"threshold_uncertainty_score":0.723012,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399656493","doi":"10.1093/nargab/lqae068","title":"Structure-based learning to predict and model protein–DNA interactions and transcription-factor co-operativity in <i>cis</i>-regulatory elements","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"BC Children's Hospital; University of British Columbia","funders":"Agencia Estatal de Investigación; Ministerio de Ciencia e Innovación; Generalitat de Catalunya; Human Frontier Science Program","keywords":"Transcription factor; DNA; Computational biology; Transcription (linguistics); Genetics; Biology; Gene","authors":[{"name":"Oriol Fornés","is_ca":true},{"name":"Alberto Meseguer","is_ca":false},{"name":"Joachim Aguirre-Plans","is_ca":false},{"name":"Patrick Gohl","is_ca":false},{"name":"Patricia Mirela Bota","is_ca":false},{"name":"Rubén Molina-Fernández","is_ca":false},{"name":"Jaume Bonet","is_ca":false},{"name":"Altair Chinchilla-Hernandez","is_ca":false},{"name":"Ferran Pegenaute","is_ca":false},{"name":"Oriol Gallego","is_ca":false},{"name":"Narcis Fernandez-Fuentes","is_ca":false},{"name":"Baldo Oliva","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.009518222244531825,"gpt":0.2376803225427711,"spread":0.2281621002982393,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001604914,0.0001536933,0.0001309495,0.00008906546,0.0001050019,0.0001322299,0.0000516494,0.0001045443,0.000008231008],"category_scores_gemma":[0.00001687442,0.0001363767,0.00002614349,0.00004437333,0.00003650101,0.00002207771,0.00003819708,0.0001193939,0.000001160463],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000211229,"about_ca_system_score_gemma":0.00007167393,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004581879,"about_ca_topic_score_gemma":0.00003370727,"domain_scores_codex":[0.9992884,0.00002461744,0.0002550432,0.0001892119,0.00008364513,0.0001590717],"domain_scores_gemma":[0.999728,0.000009338351,0.00003866473,0.0001068062,0.00001986497,0.00009736467],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00005584792,0.00001098373,0.0002733296,0.0001577139,0.00002480716,7.24963e-7,0.0006918469,0.0004737445,0.9702142,0.00007458242,0.00003769484,0.02798454],"study_design_scores_gemma":[0.000733742,0.0005101115,0.0006368145,0.0001547301,0.00002870567,0.00001940602,0.0003408661,0.2895376,0.6894658,0.0002747208,0.01785967,0.0004378499],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9842872,0.0005177411,0.01433685,0.0001057206,0.00004839286,0.0004571863,0.0001708975,0.00001348535,0.00006254443],"genre_scores_gemma":[0.984235,0.0001884037,0.01518431,0.0001576214,0.00003707567,0.00002462083,0.00005794323,0.00001564772,0.00009935999],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2890639,"threshold_uncertainty_score":0.556128,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4220964301","doi":"10.1093/nargab/lqac018","title":"Data-driven identification of inherent features of eukaryotic stress-responsive genes","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"National Cancer Institute; National Institutes of Health; Boehringer Ingelheim Fonds","keywords":"Gene; Computational biology; Biology; Saccharomyces cerevisiae; Yeast; Gene expression; Genetics","authors":[{"name":"Pablo Latorre","is_ca":false},{"name":"René Böttcher","is_ca":false},{"name":"Mariona Nadal‐Ribelles","is_ca":false},{"name":"Constance H. Li","is_ca":true},{"name":"Carme Solé","is_ca":false},{"name":"Gerard Martínez-Cebrián","is_ca":false},{"name":"Paul C. Boutros","is_ca":true},{"name":"Francesc Posas","is_ca":false},{"name":"Eulàlia de Nadal","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0202331987383453,"gpt":0.2752859193507044,"spread":0.2550527206123591,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002798913,0.0000772061,0.0001260352,0.00007016124,0.00009690121,0.00001592937,0.0003292596,0.00004018711,0.000007836014],"category_scores_gemma":[0.00004906905,0.00007137768,0.00003358058,0.0000730995,0.00007117524,0.000007876406,0.0006350122,0.00007493375,6.090096e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001268631,"about_ca_system_score_gemma":0.000107068,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001926907,"about_ca_topic_score_gemma":0.0000174152,"domain_scores_codex":[0.9992034,0.00003568692,0.0003484682,0.0001181887,0.0001710901,0.0001231793],"domain_scores_gemma":[0.9992138,0.00001568411,0.0002339162,0.0004120163,0.0000801763,0.00004443952],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.000184338,0.0000733666,0.002917611,0.0002084044,0.0001051679,0.000001060782,0.0003232762,0.001666601,0.9707094,0.0001435353,0.001167829,0.02249938],"study_design_scores_gemma":[0.002030853,0.001837462,0.03600765,0.00004259186,0.0001221445,0.00008502517,0.01075093,0.1030045,0.8282084,0.0001284883,0.01717402,0.0006079052],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9965376,0.001334374,0.0004858712,0.00004843671,0.00004917974,0.0001852731,0.001276633,0.000002283786,0.00008037517],"genre_scores_gemma":[0.9955505,0.001151115,0.001986881,0.00003281546,0.00003009308,0.00000687178,0.0009638694,0.000008445398,0.0002694133],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.142501,"threshold_uncertainty_score":0.2910697,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4378713713","doi":"10.1093/nargab/lqad052","title":"Refining the genomic determinants underlying escape from X-chromosome inactivation","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genetic and Clinical Aspects of Sex Determination and Chromosomal Abnormalities","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"BC Children's Hospital; University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"X-inactivation; Biology; Genetics; Gene; Transgene; X chromosome; Transcription (linguistics); Transcription factor; Phenotype; Chromosome; Context (archaeology)","authors":[{"name":"Samantha B. Peeters","is_ca":true},{"name":"Tiffany Leung","is_ca":true},{"name":"Oriol Fornés","is_ca":true},{"name":"Rachelle A. Farkas","is_ca":true},{"name":"Wyeth W. Wasserman","is_ca":true},{"name":"Carolyn J. Brown","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02274888419798342,"gpt":0.2552338556071225,"spread":0.2324849714091391,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000233111,0.0001429464,0.0001368434,0.00003111534,0.0006238763,0.00007917364,0.0002446078,0.0000697659,0.00007055928],"category_scores_gemma":[0.0000243549,0.0001143816,0.00006319468,0.00006008073,0.00008939668,0.00001024784,0.0004003317,0.0001564408,0.000008591762],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003425895,"about_ca_system_score_gemma":0.00006793106,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000347404,"about_ca_topic_score_gemma":0.00002218595,"domain_scores_codex":[0.9990566,0.00004348197,0.0003721556,0.0001744534,0.0001551521,0.000198145],"domain_scores_gemma":[0.9994312,0.00004024815,0.0001941372,0.0002411873,0.00003251785,0.00006070337],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0005218573,0.0003364911,0.01881542,0.0001580647,0.0003823954,0.00002002505,0.007412701,0.001455904,0.4686413,0.001103822,0.003451434,0.4977006],"study_design_scores_gemma":[0.005404233,0.003864079,0.0779181,0.00005932728,0.0002536098,0.0004719048,0.0585863,0.06041415,0.0607601,0.003795194,0.7260005,0.002472438],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.997607,0.0004278434,0.000221886,0.0002868027,0.0001779711,0.0001076719,0.00007983587,0.0000114144,0.001079585],"genre_scores_gemma":[0.9963077,0.0002769186,0.001406802,0.001189154,0.0001147147,0.00002353023,0.0001419306,0.00001519105,0.0005240733],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7225491,"threshold_uncertainty_score":0.4798416,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4408249210","doi":"10.1093/nargab/lqaf016","title":"BacTermFinder: a comprehensive and general bacterial terminator finder using a CNN ensemble","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Memorial University of Newfoundland","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada","keywords":"Terminator (solar); False positive paradox; Computer science; Bacterial transcription; Bacterial genome size; Convolutional neural network; Artificial intelligence; Computational biology; Bacterial protein; Machine learning; Biology; Data mining; RNA; Genome; Gene; Genetics; RNA polymerase; Physics","authors":[{"name":"Lourdes Peña‐Castillo","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0124910237695655,"gpt":0.2440232268859487,"spread":0.2315322031163832,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00007776398,0.0002096111,0.0002285163,0.00006878284,0.0002037168,0.00009234568,0.00009778196,0.0001632328,0.000003047825],"category_scores_gemma":[0.00001251631,0.000195346,0.00005408434,0.00005238589,0.0001227115,0.000002855758,0.0003377233,0.00007064141,0.000001332739],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001842663,"about_ca_system_score_gemma":0.0001063781,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001544657,"about_ca_topic_score_gemma":0.000009802726,"domain_scores_codex":[0.9991635,0.00001721713,0.0003033082,0.0002110235,0.0000551886,0.00024973],"domain_scores_gemma":[0.999528,0.00001411301,0.00009865446,0.0002087923,0.000075891,0.00007454213],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000104864,0.00002943683,0.002809378,0.0001524292,0.0002069325,0.000001979118,0.0004728647,0.0000372805,0.9746245,0.0002815092,0.0007525706,0.02052624],"study_design_scores_gemma":[0.008736446,0.001210317,0.06364343,0.0001661013,0.0005285835,0.0003817885,0.003128974,0.05779716,0.2261533,0.001810291,0.6338975,0.002546029],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9969257,0.001224368,0.0005960703,0.00007252605,0.000249008,0.0002220387,0.00005486259,0.000003541589,0.0006519571],"genre_scores_gemma":[0.9747844,0.0019127,0.02181187,0.0008822095,0.0001849781,0.000007761992,0.00004978449,0.00001772397,0.000348614],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7484712,"threshold_uncertainty_score":0.796598,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4400562029","doi":"10.1093/nargab/lqae079","title":"Optimizing hybrid ensemble feature selection strategies for transcriptomic biomarker discovery in complex diseases","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":9,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université Laval","funders":"Natural Sciences and Engineering Research Council of Canada; Agence Nationale de la Recherche","keywords":"Feature selection; Robustness (evolution); Computer science; Biomarker discovery; Feature (linguistics); Data mining; Machine learning; Artificial intelligence; Computational biology; Biology; Gene","authors":[{"name":"Elsa Claude","is_ca":true},{"name":"Mickaël Leclercq","is_ca":true},{"name":"Patricia Thébault","is_ca":false},{"name":"Arnaud Droit","is_ca":true},{"name":"Raluca Uricaru","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01872548788231583,"gpt":0.2641647738759884,"spread":0.2454392859936726,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00008140791,0.0001247644,0.000104868,0.00009097573,0.00007312059,0.0002574954,0.00006841688,0.00008005061,0.000002437242],"category_scores_gemma":[0.000005362114,0.0001083656,0.0000655488,0.00008173211,0.00003301516,0.00003287704,0.00002262371,0.00005703908,0.000001027583],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003108073,"about_ca_system_score_gemma":0.0001381749,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000002738499,"about_ca_topic_score_gemma":0.00002109794,"domain_scores_codex":[0.9994144,0.00001068186,0.000188108,0.0001729099,0.00005487983,0.000159006],"domain_scores_gemma":[0.9997817,0.000008514221,0.00003971245,0.0000998473,0.00002569712,0.00004448085],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001845928,0.00003214569,0.0001117319,0.0003703964,0.00005124657,7.813145e-7,0.000430308,0.000456,0.9317458,0.000702358,0.01887657,0.0470381],"study_design_scores_gemma":[0.001481665,0.0003402896,0.00222303,0.0001218873,0.00008746423,0.00005189836,0.002994389,0.4169595,0.06208326,0.001228292,0.5117786,0.0006497126],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8245448,0.009163351,0.163598,0.0005460769,0.0003453537,0.0006103717,0.0003321086,0.00004156251,0.0008183926],"genre_scores_gemma":[0.9902861,0.001845788,0.006723052,0.0001690712,0.00009491441,0.00003617926,0.0005408665,0.00001706601,0.0002869235],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8696625,"threshold_uncertainty_score":0.441902,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4317743512","doi":"10.1093/nargab/lqad003","title":"Differential Expression Enrichment Tool (DEET): an interactive atlas of human differential gene expression","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Bioinformatics and Genomic Networks","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Canadian Institute for Advanced Research; Vector Institute; SickKids Foundation; University of Toronto","funders":"National Institutes of Health; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; Centre for Applied Genomics; Genome Canada","keywords":"Computational biology; Expression (computer science); Pipeline (software); Gene expression; Gene; Computer science; Differential (mechanical device); Biology; Gene expression profiling; DEET; Information retrieval; Data mining; Genetics","authors":[{"name":"Dustin Sokolowski","is_ca":true},{"name":"Jedid Ahn","is_ca":true},{"name":"Lauren Erdman","is_ca":true},{"name":"Huayun Hou","is_ca":true},{"name":"Kai Ellis","is_ca":true},{"name":"Liangxi Wang","is_ca":true},{"name":"Anna Goldenberg","is_ca":true},{"name":"Michael D. Wilson","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.009721680046971246,"gpt":0.2465778461921381,"spread":0.2368561661451669,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001578888,0.0002920839,0.0003149807,0.0001312535,0.0002266876,0.00008692497,0.0002873633,0.0002808159,0.00003959175],"category_scores_gemma":[0.00001222912,0.0002434607,0.0001258438,0.00009808125,0.0001031013,0.00003099788,0.0005537,0.0001643352,0.00001238434],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002488581,"about_ca_system_score_gemma":0.00004460683,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006351647,"about_ca_topic_score_gemma":0.000004141311,"domain_scores_codex":[0.9984198,0.00003473378,0.0007158102,0.0002493916,0.0002126412,0.000367679],"domain_scores_gemma":[0.9989069,0.00001635577,0.0003744986,0.0004679522,0.00008381408,0.0001505508],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001380665,0.0000880434,0.0002768489,0.0000987852,0.00005250027,0.000001051162,0.001123964,0.0000884308,0.9885929,0.00006934321,0.001500469,0.007969625],"study_design_scores_gemma":[0.001982508,0.001051702,0.001921365,0.0000927022,0.00005957377,0.00001806494,0.001504599,0.0349923,0.9538152,0.0003769824,0.003533858,0.000651206],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9895909,0.00005851988,0.009232264,0.00001575568,0.0003022024,0.0003372667,0.0001255134,0.00002929692,0.0003082129],"genre_scores_gemma":[0.9924927,0.0004864483,0.005056306,0.00005549724,0.0003112601,0.00001797613,0.001355726,0.00003000235,0.0001940442],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03490387,"threshold_uncertainty_score":0.9928038,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4221050649","doi":"10.1093/nargab/lqac026","title":"Construction of a chromosome-level Japanese stickleback species genome using ultra-dense linkage analysis with single-cell sperm sequencing","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Institute of Genetics; Japan Society for the Promotion of Science; Research Organization of Information and Systems","keywords":"Genome; Biology; Genetics; Linkage (software); Contig; Genetic linkage; Genotyping; Genomics; Chromosome; Sperm; Computational biology; Stickleback; Genotype; Gene; Fish <Actinopterygii>","authors":[{"name":"Kazutoshi Yoshitake","is_ca":false},{"name":"Asano Ishikawa","is_ca":false},{"name":"Ryo Yonezawa","is_ca":false},{"name":"Shigeharu Kinoshita","is_ca":false},{"name":"Jun Kitano","is_ca":false},{"name":"Shuichi Asakawa","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02544240629958201,"gpt":0.2044549693694163,"spread":0.1790125630698343,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002057961,0.0002294768,0.0003571156,0.0001673286,0.0002952752,0.00004221295,0.0001668495,0.00007874807,0.00002374278],"category_scores_gemma":[0.00001172536,0.0002124814,0.0001243691,0.0003204928,0.0002172069,0.000003747175,0.000205885,0.0001147584,7.39892e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009819934,"about_ca_system_score_gemma":0.0001401453,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003706752,"about_ca_topic_score_gemma":0.00002718053,"domain_scores_codex":[0.9987998,0.00003192962,0.0004743103,0.000237995,0.0001886337,0.0002673271],"domain_scores_gemma":[0.9991014,0.00001916467,0.0003646071,0.0003140388,0.0001195737,0.00008121086],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00007555017,0.00004897649,0.002274493,0.0001035477,0.0004929789,0.000003320767,0.002731857,0.01083696,0.9831325,0.00002091323,0.000006292774,0.0002726019],"study_design_scores_gemma":[0.00749285,0.007092036,0.01264675,0.00006528066,0.005200095,0.001745608,0.1392973,0.06492877,0.7400774,0.0001722724,0.017318,0.003963596],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.996067,0.0007831628,0.001419707,0.00001441221,0.00007332363,0.0002237356,0.0005633325,0.000003778957,0.0008514841],"genre_scores_gemma":[0.967439,0.0004893676,0.03162498,0.00007139955,0.00007777801,0.000005569031,0.0001482046,0.00002138637,0.000122309],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2430551,"threshold_uncertainty_score":0.8664742,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4413342468","doi":"10.1093/nargab/lqaf108","title":"Explicit Scale Simulation for analysis of RNA-sequencing count data with ALDEx2","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":6,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; National Institutes of Health","keywords":"Computer science; Normalization (sociology); Scale (ratio); Sample size determination; Data mining; Sample (material); Cutoff; False positive paradox; Statistics; Machine learning; Mathematics","authors":[{"name":"Gregory B. Gloor","is_ca":true},{"name":"Michelle Pistner Nixon","is_ca":false},{"name":"Justin D. Silverman","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02634878122862942,"gpt":0.2780306867069964,"spread":0.251681905478367,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002074201,0.0001167241,0.0002264612,0.0001072716,0.00009446238,0.00002587186,0.0001860705,0.00007604151,8.853622e-7],"category_scores_gemma":[0.00002515644,0.00009685291,0.00005320729,0.0001752415,0.00005091845,0.000002715255,0.0001893216,0.00002731245,1.572966e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001741159,"about_ca_system_score_gemma":0.00009754094,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001788556,"about_ca_topic_score_gemma":0.0001029488,"domain_scores_codex":[0.9993029,0.000005787622,0.0003186098,0.0001743694,0.00006208875,0.000136251],"domain_scores_gemma":[0.9992105,0.00003554372,0.0001542501,0.0004374147,0.0001325459,0.0000297204],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001538369,0.0001890446,0.04554809,0.001243365,0.01222119,7.43924e-7,0.003879792,0.2710411,0.5727265,0.002056193,0.001717099,0.08783855],"study_design_scores_gemma":[0.001495653,0.000338775,0.004150396,0.00002677651,0.00145156,0.000001562647,0.001667439,0.9449819,0.01930944,0.0001189332,0.02613562,0.0003219732],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.928109,0.000693471,0.06983703,0.00004345024,0.00004109845,0.0003027475,0.0004583682,0.000002048971,0.0005127892],"genre_scores_gemma":[0.9703379,0.0005888418,0.02841371,0.000179838,0.00002862525,0.000009487878,0.0003692454,0.000007392136,0.00006495077],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6739408,"threshold_uncertainty_score":0.3949547,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4317624614","doi":"10.1093/nargab/lqad006","title":"Model-driven experimental design workflow expands understanding of regulatory role of Nac in <i>Escherichia coli</i>","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Microbial Metabolic Engineering and Bioproduction","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Queen's University","funders":"National Institutes of Health; Ministry of Science and ICT, South Korea; Novo Nordisk Fonden; National Institute of General Medical Sciences; Danmarks Tekniske Universitet; Novo Nordisk; National Research Foundation of Korea; Ulsan National Institute of Science and Technology; National Research Foundation","keywords":"Escherichia coli; Workflow; Computer science; Computational biology; Business; Chemistry; Biochemical engineering; Process management; Biology; Engineering; Biochemistry; Gene; Database","authors":[{"name":"Joon Young Park","is_ca":false},{"name":"Sang‐Mok Lee","is_ca":false},{"name":"Ali Ebrahim","is_ca":false},{"name":"Zoe K. Scott-Nevros","is_ca":false},{"name":"Laurence Yang","is_ca":true},{"name":"Anand V. Sastry","is_ca":false},{"name":"Sang Woo Seo","is_ca":false},{"name":"Bernhard Ø. Palsson","is_ca":false},{"name":"Donghyuk Kim","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0216666012447547,"gpt":0.2228370784855388,"spread":0.2011704772407841,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000204483,0.00009631567,0.0001509663,0.00008163076,0.00002705052,0.000006863963,0.00008503821,0.0001032304,7.899211e-7],"category_scores_gemma":[0.00001003097,0.00009315649,0.00003677702,0.0001440052,0.00005534708,0.00000577861,0.00006863567,0.00004562622,5.202717e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002457876,"about_ca_system_score_gemma":0.00004417214,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00000381188,"about_ca_topic_score_gemma":0.000001433115,"domain_scores_codex":[0.9994096,0.00001018994,0.0002818981,0.00009994896,0.00006624933,0.0001321106],"domain_scores_gemma":[0.9997012,0.00000322032,0.00009492911,0.0001471998,0.00002060006,0.00003283324],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00004774571,0.00002045442,0.00008082165,0.00004687357,0.00001642189,1.046245e-7,0.0007947482,0.02102335,0.9768854,0.0001421658,0.000121869,0.0008200043],"study_design_scores_gemma":[0.0003737224,0.0001119896,0.0001057518,0.00002523217,0.000007582296,0.000002129871,0.001083482,0.1458098,0.8518782,0.0001042498,0.0003800679,0.0001177569],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9914683,0.0005673154,0.007635908,0.0000101536,0.00006204885,0.0001341337,0.00001763284,0.000008815719,0.00009571283],"genre_scores_gemma":[0.9790167,0.0007462277,0.02011109,0.00001398689,0.00003633741,0.00000374715,0.00002803905,0.00001091156,0.00003300107],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1250072,"threshold_uncertainty_score":0.3798812,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3037354611","doi":"10.1093/nargab/lqaa043","title":"Factorial study of the RNA-seq computational workflow identifies biases as technical gene signatures","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"Fonds de recherche du Québec – Nature et technologies; Fonds de Recherche du Québec - Santé; Natural Sciences and Engineering Research Council of Canada","keywords":"Modularity (biology); RNA-Seq; Pipeline (software); Computational biology; Computer science; Workflow; RNA; Modular design; Software; Gene; Biology; Genetics; Transcriptome; Gene expression; Database; Programming language","authors":[{"name":"Joël Simoneau","is_ca":true},{"name":"Ryan Gosselin","is_ca":true},{"name":"Michelle S. Scott","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01929335731755154,"gpt":0.2429575639041722,"spread":0.2236642065866206,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00009318686,0.0001520417,0.0001845115,0.00002123464,0.0001510695,0.00004210164,0.0002390946,0.0001063737,0.000003461688],"category_scores_gemma":[0.0001071416,0.0001097211,0.00007977031,0.00008685623,0.0001084573,0.000001873455,0.0003776481,0.00009553749,0.000001994357],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00000752955,"about_ca_system_score_gemma":0.00007130846,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001416424,"about_ca_topic_score_gemma":0.00001911121,"domain_scores_codex":[0.9991521,0.00002391925,0.0003619365,0.0001580695,0.0001656823,0.0001383453],"domain_scores_gemma":[0.9994617,0.00003364567,0.000167875,0.0001819525,0.00008889865,0.00006597465],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0004011201,0.000374323,0.007692194,0.0001194602,0.0006074055,0.000002807137,0.007078776,0.02265776,0.9493709,0.0002435504,0.004218216,0.007233486],"study_design_scores_gemma":[0.01856792,0.01535394,0.1614878,0.0001646931,0.00125589,0.0001512994,0.0291735,0.04517936,0.6532864,0.007400204,0.06363356,0.004345412],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9980605,0.0006344064,0.0002924911,0.0001738189,0.0002286154,0.00034512,0.00008074287,0.000003922098,0.000180432],"genre_scores_gemma":[0.9965737,0.0002425565,0.002524171,0.0003643734,0.0002351169,0.000006016764,0.00002588211,0.00001132708,0.00001683883],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2960845,"threshold_uncertainty_score":0.4474295,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4409905545","doi":"10.1093/nargab/lqaf048","title":"Discovering governing equations of biological systems through representation learning and sparse model discovery","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Gene Regulatory Network Analysis","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Autoencoder; Computer science; Identification (biology); Representation (politics); Gene regulatory network; Artificial intelligence; Machine learning; Dynamical systems theory; Biological network; Complex system; Biological data; Nonlinear system; PageRank; Systems biology; Key (lock); Theoretical computer science; Deep learning; Data mining; Computational biology; Gene; Biology; Bioinformatics; Gene expression","authors":[{"name":"Mehrshad Sadria","is_ca":true},{"name":"Vasu Swaroop","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02124096170560071,"gpt":0.2637873461398547,"spread":0.2425463844342539,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001548432,0.0001061464,0.0001739702,0.00003909052,0.0001150681,0.0000658988,0.00006129714,0.0001043251,3.988523e-7],"category_scores_gemma":[0.00006327251,0.00009267232,0.00005108383,0.00009479669,0.00008016852,0.00002062786,0.0001791538,0.00006810781,3.801366e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001642258,"about_ca_system_score_gemma":0.00005140935,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001978892,"about_ca_topic_score_gemma":0.000006792376,"domain_scores_codex":[0.9993159,0.00002459034,0.0003226799,0.0001450755,0.00006829643,0.0001234985],"domain_scores_gemma":[0.9995976,0.00002600359,0.000172355,0.0001394437,0.0000391848,0.00002538246],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001015175,0.0000504393,0.03867861,0.0003669077,0.0004461965,6.362888e-7,0.001228286,0.7352794,0.2100444,0.007768797,0.0002232499,0.005811573],"study_design_scores_gemma":[0.0003243302,0.00007481765,0.0006259998,0.00004044619,0.00007211568,0.000003843114,0.002067787,0.988667,0.005973592,0.0001944893,0.001796445,0.0001592015],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7858101,0.00147127,0.2119612,0.00002190273,0.00003817147,0.0001006096,0.00001822845,0.000004775178,0.0005737559],"genre_scores_gemma":[0.9907578,0.002779139,0.005941623,0.00002467181,0.00003181415,0.000004986414,0.0001117131,0.000005902776,0.0003423638],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2533875,"threshold_uncertainty_score":0.3779067,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4366818125","doi":"10.1093/nargab/lqad038","title":"SUP: a probabilistic framework to propagate genome sequence uncertainty, with applications","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Public Health Agency of Canada; Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Western University","keywords":"Resampling; Computer science; Propagation of uncertainty; Probabilistic logic; Sequence (biology); Representation (politics); Variance (accounting); Algorithm; Data mining; Artificial intelligence; Biology","authors":[{"name":"Devan Becker","is_ca":true},{"name":"David Champredon","is_ca":true},{"name":"Connor Chato","is_ca":true},{"name":"Gopi Gugan","is_ca":true},{"name":"Art F. Y. Poon","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01576609938421404,"gpt":0.2448559491589844,"spread":0.2290898497747704,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001845222,0.0001970669,0.0001729726,0.0000736495,0.0002105394,0.00006874185,0.0002076536,0.0001062623,0.000002311855],"category_scores_gemma":[0.0000277732,0.0001589012,0.00003751061,0.0002839564,0.0001158164,0.000001707993,0.0002002329,0.00008400682,0.00005468295],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002377794,"about_ca_system_score_gemma":0.00009932439,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007332177,"about_ca_topic_score_gemma":0.0000185949,"domain_scores_codex":[0.9990121,0.00001075714,0.000267089,0.0002612501,0.0001112671,0.0003375415],"domain_scores_gemma":[0.9992644,0.00002058086,0.00008122452,0.0003704992,0.0001110745,0.000152278],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0007446606,0.0002882732,0.006850411,0.001730748,0.001113858,0.00002403007,0.01435043,0.1448759,0.7101156,0.01745198,0.003474927,0.0989793],"study_design_scores_gemma":[0.001451663,0.002740213,0.006467366,0.0001062783,0.0001769657,0.0001658731,0.003451917,0.03006088,0.004274028,0.009587438,0.9393823,0.002135035],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9877435,0.0003003085,0.009359669,0.0004608966,0.00005232721,0.001209049,0.0002960319,0.00002560375,0.0005525929],"genre_scores_gemma":[0.9449559,0.002094392,0.05055433,0.000907774,0.0002499554,0.0003935381,0.0003085048,0.00004902147,0.0004865296],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9359074,"threshold_uncertainty_score":0.6479805,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4404333429","doi":"10.1093/nargab/lqae145","title":"Refining SARS-CoV-2 intra-host variation by leveraging large-scale sequencing data","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"SARS-CoV-2 and COVID-19 Research","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Mila - Quebec Artificial Intelligence Institute; Montreal Heart Institute","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Fonds de Recherche du Québec - Santé; Institut de Valorisation des Données; Canada Foundation for Innovation","keywords":"Workflow; Computational biology; Host (biology); DNA sequencing; Computer science; Cluster analysis; Visualization; Biology; Data mining; Genetics; Artificial intelligence; DNA; Database","authors":[{"name":"Fatima Mostefai","is_ca":true},{"name":"Jean‐Christophe Grenier","is_ca":true},{"name":"Raphaël Poujol","is_ca":true},{"name":"Julie Hussin","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06431370519734829,"gpt":0.3304385491763903,"spread":0.266124843979042,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009824208,0.0001827446,0.0002506103,0.0001902634,0.0001732329,0.0003221194,0.0002278448,0.0001448146,0.000004410714],"category_scores_gemma":[0.00008597637,0.0001579679,0.00004433131,0.0002853907,0.0000474885,0.0004451659,0.0003189487,0.0003923065,0.00008413548],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002633574,"about_ca_system_score_gemma":0.0004226823,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001302382,"about_ca_topic_score_gemma":0.00003396641,"domain_scores_codex":[0.9985276,0.0000166095,0.0004701776,0.0002766224,0.0003086245,0.0004002906],"domain_scores_gemma":[0.9992204,0.00008575957,0.00007383039,0.0005105725,0.00006815726,0.00004127293],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001125015,0.00007002256,0.001774159,0.001906391,0.0002626949,0.00007404178,0.01900078,0.000001916209,0.8972982,0.0004341234,0.01516514,0.0639],"study_design_scores_gemma":[0.0006400109,0.00009666858,0.00007720057,0.0001981788,0.00007656134,0.0001246683,0.001365328,0.5161564,0.03971944,0.00006422629,0.4412618,0.0002195742],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9633734,0.002425485,0.02694477,0.0007924984,0.0003768727,0.0003222607,0.0004453265,0.0001759984,0.005143424],"genre_scores_gemma":[0.8372538,0.0006448654,0.07744149,0.08203439,0.0010681,0.00002744358,0.001170713,0.0001744945,0.0001846808],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8575788,"threshold_uncertainty_score":0.6441746,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4411754998","doi":"10.1093/nargab/lqaf087","title":"MOLGENIS VIP: an end-to-end DNA variant interpretation pipeline for research and diagnostics configurable to support rapid implementation of new methods","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek","keywords":"Scalability; Pipeline (software); Computer science; Data science; DNA sequencing; Genome; Variety (cybernetics); Software; Protocol (science); Computational biology; Data mining; Biology; Medicine; Artificial intelligence; Genetics; Database; Gene","authors":[{"name":"W Maassen","is_ca":false},{"name":"Lennart Johansson","is_ca":false},{"name":"Bart Charbon","is_ca":false},{"name":"Dennis Hendriksen","is_ca":false},{"name":"Sander van den Hoek","is_ca":false},{"name":"Mariska Slofstra","is_ca":false},{"name":"René Mulder","is_ca":false},{"name":"Martine T. Meems-Veldhuis","is_ca":false},{"name":"Robert Sietsma","is_ca":false},{"name":"Henny H. Lemmink","is_ca":false},{"name":"Cleo C. van Diemen","is_ca":false},{"name":"Mariëlle van Gijn","is_ca":false},{"name":"Morris A. Swertz","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.04216451958384585,"gpt":0.3995709628316272,"spread":0.3574064432477814,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009796494,0.0001465238,0.0002264815,0.0001641789,0.0001424209,0.00006401438,0.0001418879,0.00009843807,0.000009881399],"category_scores_gemma":[0.0001482685,0.0001410034,0.00003733937,0.0001351715,0.00006114475,0.000004443714,0.0002073816,0.00005480216,9.726253e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002208491,"about_ca_system_score_gemma":0.00022589,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008162114,"about_ca_topic_score_gemma":0.000103635,"domain_scores_codex":[0.9988933,0.00003820696,0.0004906295,0.0002230299,0.00009135903,0.0002634795],"domain_scores_gemma":[0.9991115,0.0001073642,0.000100389,0.0002357344,0.0002883198,0.0001566723],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0002346708,0.00004189177,0.000280063,0.0001558227,0.000131582,2.93627e-7,0.003562842,0.00006749311,0.4544094,0.001030477,0.005246486,0.534839],"study_design_scores_gemma":[0.002411844,0.005629057,0.003577775,0.00004613729,0.0001724081,0.00001527891,0.009866436,0.0110947,0.7267193,0.003407957,0.2365142,0.0005448522],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7813899,0.0008893271,0.2146468,0.0006238882,0.0002465516,0.001238215,0.0004492069,0.000003680369,0.0005123971],"genre_scores_gemma":[0.7248366,0.002648905,0.2707936,0.0008577693,0.0001158703,0.00006415499,0.0002872321,0.00002134206,0.0003745793],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5342941,"threshold_uncertainty_score":0.574995,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3158876629","doi":"10.1093/nargab/lqab027","title":"On the use of direct-coupling analysis with a reduced alphabet of amino acids combined with super-secondary structure motifs for protein fold prediction","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Protein Structure and Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"BC Children's Hospital; University of British Columbia","funders":"Instituto de Salud Carlos III; Ministerio de Economía y Competitividad; Generalitat de Catalunya","keywords":"Protein secondary structure; Protein folding; Sequence (biology); Protein structure prediction; Protein structure; Structural motif; Alphabet; Computational biology; Protein sequencing; Amino acid; Chemistry; Biological system; Peptide sequence; Crystallography; Biology; Biochemistry","authors":[{"name":"Bernat Anton","is_ca":false},{"name":"Mireia Besalú","is_ca":false},{"name":"Oriol Fornés","is_ca":true},{"name":"Jaume Bonet","is_ca":false},{"name":"Alexis Molina","is_ca":false},{"name":"Rubén Molina-Fernández","is_ca":false},{"name":"Gemma De las Cuevas","is_ca":false},{"name":"Narcís Fernández‐Fuentes","is_ca":false},{"name":"Baldo Oliva","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.008394857317764562,"gpt":0.1939350341460129,"spread":0.1855401768282484,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001075253,0.0001643697,0.0002830132,0.00006229619,0.00009515691,0.00003583733,0.00009767715,0.0001397505,0.000004006395],"category_scores_gemma":[0.00005010813,0.0001008119,0.00008443888,0.0002031658,0.000100914,0.00001213945,0.00005273925,0.0001029673,3.969766e-8],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001219392,"about_ca_system_score_gemma":0.0001429076,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006572624,"about_ca_topic_score_gemma":0.00005846303,"domain_scores_codex":[0.9992337,0.00001450391,0.0003175089,0.0001638987,0.0001241142,0.0001462363],"domain_scores_gemma":[0.9991338,0.0000308739,0.0002198536,0.0003607143,0.0002180072,0.00003674621],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001884175,0.00008852341,0.00304068,0.0004943596,0.002292244,0.000001647764,0.0007743719,0.01028584,0.9771647,0.001077081,0.0000968363,0.00279948],"study_design_scores_gemma":[0.001910944,0.002913116,0.002126418,0.00008005489,0.0007003131,0.00001745115,0.0005546802,0.1526633,0.8377107,0.000260654,0.0007439028,0.000318491],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9885662,0.00009774276,0.009825756,0.00004300839,0.00002151369,0.000534096,0.0008704442,0.000004740188,0.00003653233],"genre_scores_gemma":[0.9753551,0.00005035595,0.02380625,0.00007565274,0.00002159719,0.00001751972,0.0006093847,0.00001452912,0.0000495902],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1423775,"threshold_uncertainty_score":0.411099,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4401385491","doi":"10.1093/nargab/lqae098","title":"Evaluating cell type deconvolution in FFPE breast tissue: application to benign breast disease","year":2024,"lang":"en","type":"review","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"National Cancer Institute; National Institutes of Health; Mayo Clinic","keywords":"Deconvolution; Transcriptome; Computational biology; Cell type; RNA; Computer science; Cell; Pattern recognition (psychology); Artificial intelligence; Biology; Gene expression; Gene; Algorithm; Genetics","authors":[{"name":"Yuanhang Liu","is_ca":false},{"name":"Robert A. Vierkant","is_ca":false},{"name":"Aditya Bhagwate","is_ca":false},{"name":"William Jons","is_ca":false},{"name":"Melody Stallings‐Mann","is_ca":false},{"name":"Bryan M. McCauley","is_ca":false},{"name":"Jodi M. Carter","is_ca":true},{"name":"Melissa Stephens","is_ca":false},{"name":"Michael E. Pfrender","is_ca":false},{"name":"Laurie E. Littlepage","is_ca":false},{"name":"Derek C. Radisky","is_ca":false},{"name":"Julie M. Cunningham","is_ca":false},{"name":"Amy C. Degnim","is_ca":false},{"name":"Stacey J. Winham","is_ca":false},{"name":"Chen Wang","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03001953645356214,"gpt":0.3274310347500039,"spread":0.2974114982964418,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.000332834,0.0003887708,0.000524454,0.0001744934,0.00007791628,0.0001139903,0.0002387066,0.0003466039,0.000004708066],"category_scores_gemma":[0.00001562921,0.000341922,0.0001401296,0.0002595131,0.00003776542,0.000008721546,0.0001620283,0.0002214199,0.0001066994],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001100826,"about_ca_system_score_gemma":0.0003560603,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002230582,"about_ca_topic_score_gemma":0.00003182842,"domain_scores_codex":[0.9984146,0.0000387432,0.0007103063,0.0003900776,0.0001447721,0.0003014885],"domain_scores_gemma":[0.9991089,0.00001216708,0.000187381,0.0003861639,0.00008424475,0.000221196],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004508648,0.00005631947,0.00001726304,0.007694665,0.00004197945,0.00000202673,0.0001342537,0.00007965664,0.001174539,0.00002357387,0.0003345943,0.990396],"study_design_scores_gemma":[0.0003869186,0.0002764431,0.00003897206,0.001634841,0.0006910455,0.0001245922,0.00006318749,0.01449058,0.00006993916,0.00004791445,0.9813074,0.0008681306],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.005094885,0.9904845,0.00174433,0.00004162067,0.0003565085,0.001294223,0.0007266488,0.00002328421,0.0002340099],"genre_scores_gemma":[0.002115108,0.9917395,0.003602833,0.00010055,0.0003351682,0.00006158047,0.00174401,0.00006941539,0.0002318005],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9895279,"threshold_uncertainty_score":0.9999033,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4362562485","doi":"10.1093/nargab/lqad031","title":"Known sequence features explain half of all human gene ends","year":2022,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Occupational Cancer Research Centre; University of Toronto","funders":"Canadian Institutes of Health Research; University of Toronto","keywords":"Polyadenylation; Gene; Biology; Genetics; Coding region; Computational biology; Sequence (biology); Primary transcript; Context (archaeology); RNA; Messenger RNA; Alternative splicing","authors":[{"name":"Aleksei Shkurin","is_ca":true},{"name":"Sara E. Pour","is_ca":true},{"name":"Timothy R. Hughes","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02065026306047327,"gpt":0.2755685505550601,"spread":0.2549182874945868,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003086326,0.0001083802,0.0001299612,0.000064369,0.0002159251,0.00002654307,0.0002205414,0.00006063054,0.00002002705],"category_scores_gemma":[0.00001733163,0.0001015091,0.00005668381,0.00006338835,0.00007351683,0.000005333718,0.0003777274,0.0001316187,0.000001295938],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002646016,"about_ca_system_score_gemma":0.00007171374,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006183576,"about_ca_topic_score_gemma":0.0000164512,"domain_scores_codex":[0.9992054,0.00002641445,0.0002509595,0.0001229194,0.0001636458,0.0002307383],"domain_scores_gemma":[0.9995179,0.000007778409,0.0001090108,0.0002371914,0.00004115672,0.00008691924],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00003818433,0.00002605935,0.0001373269,0.00004345999,0.00004184022,0.000004042631,0.0003483881,0.0001316828,0.9913236,0.0003374384,0.001834989,0.005733017],"study_design_scores_gemma":[0.001514517,0.002007969,0.0006181187,0.00001332733,0.00003473734,0.0003785792,0.002675666,0.007453121,0.7056831,0.0002420634,0.2787909,0.0005879198],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9974541,0.0007230897,0.0001970455,0.0001520417,0.00004229538,0.0001347822,0.0001160084,0.000005493042,0.001175119],"genre_scores_gemma":[0.9923308,0.0004709035,0.005119083,0.0003223826,0.00008258912,0.00001500169,0.0003559701,0.00001393563,0.001289283],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2856405,"threshold_uncertainty_score":0.4139419,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4405620286","doi":"10.1093/nargab/lqae176","title":"SARS-CoV-2 Illumina GeNome Assembly Line (SIGNAL), a Snakemate workflow for rapid and bulk analysis of Illumina sequencing of SARS-CoV-2 genomes","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"SARS-CoV-2 and COVID-19 Research","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Ontario Institute for Cancer Research; Canada Research Chairs; Vector Institute; University of Saskatchewan; University of Manitoba; Public Health Agency of Canada; University Health Network; Health Sciences Centre; Sunnybrook Health Science Centre; University of Toronto; Dalhousie University; Perimeter Institute; McMaster University","funders":"Genome Canada; Canadian Institutes of Health Research; Michael G. DeGroote Institute for Infectious Disease Research, McMaster University; University of Toronto; Natural Sciences and Engineering Research Council of Canada; McMaster University","keywords":"Workflow; Illumina dye sequencing; Whole genome sequencing; DNA sequencing; Genome; Personal genomics; Computer science; Computational biology; Biology; Genetics; Database; DNA; Gene","authors":[{"name":"Jalees A. Nasir","is_ca":true},{"name":"Finlay Maguire","is_ca":true},{"name":"Kendrick M. Smith","is_ca":true},{"name":"Emily M. Panousis","is_ca":true},{"name":"Sheridan J.C. Baker","is_ca":true},{"name":"Patryk Aftanas","is_ca":true},{"name":"Amogelang R. Raphenya","is_ca":true},{"name":"Brian Alcock","is_ca":true},{"name":"Hassaan Maan","is_ca":true},{"name":"Natalie Knox","is_ca":true},{"name":"Arinjay Banerjee","is_ca":true},{"name":"Karen Mossman","is_ca":true},{"name":"Bo Wang","is_ca":true},{"name":"Jared T. Simpson","is_ca":true},{"name":"Robert Kozak","is_ca":true},{"name":"Samira Mubareka","is_ca":true},{"name":"Andrew G. McArthur","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05747356832884352,"gpt":0.3294186607319,"spread":0.2719450924030565,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001254238,0.0004239451,0.001289676,0.001385159,0.0001351807,0.0001397919,0.0002521143,0.0002937357,0.000003905806],"category_scores_gemma":[0.00009795978,0.0003536975,0.0004522474,0.001354697,0.0003877228,0.0002364188,0.0002130607,0.0002941346,0.000007607644],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002323982,"about_ca_system_score_gemma":0.0005438849,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009372464,"about_ca_topic_score_gemma":0.00005812773,"domain_scores_codex":[0.9971041,0.00003267539,0.001363001,0.0004120073,0.0004814912,0.0006067154],"domain_scores_gemma":[0.9983352,0.0003112297,0.0003688562,0.0005211589,0.0003945479,0.00006900774],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0003425542,0.00008280219,0.0003618274,0.004414143,0.002338421,0.00003297995,0.006105578,0.00001611681,0.94279,0.0002255596,0.0001134891,0.04317653],"study_design_scores_gemma":[0.00136799,0.001065039,0.0001925314,0.0003199147,0.00231347,0.00009250381,0.001463431,0.3454148,0.6133795,0.0001877581,0.03374171,0.0004613398],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9833161,0.009186061,0.005133728,0.000180927,0.00009193106,0.0009953988,0.0004114539,0.00005775865,0.0006266124],"genre_scores_gemma":[0.9727031,0.001729982,0.02140032,0.003712693,0.0001713365,0.0000465996,0.0001204167,0.00008136696,0.00003419271],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3453987,"threshold_uncertainty_score":0.9998915,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W7117923635","doi":"10.1093/nargab/lqaf208","title":"G-quadruplex structures as modulators of alternative promoter usage","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"DNA and Nucleic Acid Chemistry","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Vermont Agency of Natural Resources; Institut National Du Cancer; Agence Nationale de la Recherche; CNIB","keywords":"Promoter; Gene isoform; Gene; Transcription (linguistics); Transcription factor; Regulation of gene expression; DNA; Gene expression","authors":[{"name":"Rongxin Zhang","is_ca":false},{"name":"Jean-Louis Mergny","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.004621207016402525,"gpt":0.2325125404515891,"spread":0.2278913334351865,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00008077212,0.0001312073,0.0001446098,0.00004128944,0.0000517915,0.00002290018,0.0001593809,0.0001371558,0.00001183033],"category_scores_gemma":[0.00002906378,0.0001126114,0.00005629794,0.00005479343,0.0001081494,0.000005152066,0.0001643596,0.0000665781,0.000001975267],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001147129,"about_ca_system_score_gemma":0.00006632734,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000008896021,"about_ca_topic_score_gemma":0.00000141341,"domain_scores_codex":[0.9993815,0.000006676505,0.0002685959,0.0001266099,0.00007606904,0.0001405935],"domain_scores_gemma":[0.9995459,0.000006014691,0.0001168816,0.0002268447,0.00005581268,0.00004856349],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001429629,0.00006425211,0.001902239,0.0004476704,0.0003115769,0.000002055234,0.0008663678,0.00011375,0.9538018,0.00234384,0.002157184,0.0378463],"study_design_scores_gemma":[0.0006006912,0.0001430756,0.0006473656,0.00002450165,0.00003211787,0.00001408368,0.0004169347,0.003633775,0.9667796,0.002074794,0.02542957,0.0002034674],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9908788,0.0002676934,0.0009055432,0.00005665033,0.00009131149,0.0001179176,0.00003981579,0.000005476766,0.007636847],"genre_scores_gemma":[0.9935115,0.000325697,0.005117365,0.0002700045,0.00005209451,0.000002550084,0.00005659975,0.000008415514,0.0006557709],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03764284,"threshold_uncertainty_score":0.4592159,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3217246491","doi":"10.1093/nargab/lqab105","title":"RNA-Scoop: interactive visualization of transcripts in single-cell transcriptomes","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Canada's Michael Smith Genome Sciences Centre; University of British Columbia","funders":"National Human Genome Research Institute; National Institutes of Health; Genome British Columbia; Canada's Michael Smith Genome Sciences Centre; Genome Canada","keywords":"Transcriptome; RNA; Computational biology; Visualization; Biology; Cell; Single-cell analysis; Computer science; Genetics; Gene expression; Gene; Data mining","authors":[{"name":"Maria Stephenson","is_ca":true},{"name":"Ka Ming Nip","is_ca":true},{"name":"Saber Hafezqorani","is_ca":true},{"name":"Kristina Gagalova","is_ca":true},{"name":"Chen Yang","is_ca":true},{"name":"René L. Warren","is_ca":true},{"name":"İnanç Birol","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01306987496882045,"gpt":0.228204744840605,"spread":0.2151348698717845,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001324293,0.0001808592,0.0002506641,0.00009438066,0.00004352483,0.00004052901,0.000112355,0.0001830198,0.00001041759],"category_scores_gemma":[0.00002435183,0.0001838806,0.00009658774,0.0001648239,0.00007660336,0.00001927298,0.00003197186,0.00009848466,0.000001486558],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002587031,"about_ca_system_score_gemma":0.0001116441,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001726587,"about_ca_topic_score_gemma":0.00009225534,"domain_scores_codex":[0.9989251,0.00003505922,0.0005224615,0.0001918918,0.0001134041,0.0002120392],"domain_scores_gemma":[0.999469,0.0000149776,0.0001290032,0.0001877388,0.0001293652,0.00006988895],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001205573,0.0002558457,0.001369182,0.000215655,0.00003251334,0.000003199505,0.00213214,0.00007845152,0.9887475,0.00009460849,0.00006693273,0.00688337],"study_design_scores_gemma":[0.001719977,0.000378849,0.000721767,0.00005865133,0.00003480254,0.00002123415,0.001429649,0.003641653,0.9833239,0.00005518279,0.008320272,0.0002940455],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9844031,0.001283649,0.01095187,0.00005306759,0.0002170912,0.0001618838,0.00005915186,0.000008040211,0.002862176],"genre_scores_gemma":[0.993282,0.001513225,0.004502859,0.0002699897,0.00004745847,0.000003850119,0.0002149802,0.00002255469,0.0001430412],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.008878967,"threshold_uncertainty_score":0.7498433,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4323076270","doi":"10.1093/nargab/lqad017","title":"BaM-seq and TBaM-seq, highly multiplexed and targeted RNA-seq protocols for rapid, low-cost library generation from bacterial samples","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Institute of General Medical Sciences; Natural Sciences and Engineering Research Council of Canada; Howard Hughes Medical Institute; National Institutes of Health; National Science Foundation","keywords":"RNA-Seq; Computational biology; DNA sequencing; Transcriptome; RNA; Genomic library; Deep sequencing; Biology; Computer science; Gene; Gene expression; Genetics; Genome; Base sequence","authors":[{"name":"Grace E. Johnson","is_ca":false},{"name":"Darren J. Parker","is_ca":false},{"name":"Jean‐Benoît Lalanne","is_ca":false},{"name":"Mirae Parker","is_ca":false},{"name":"Gene‐Wei Li","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0278299997355659,"gpt":0.2409403088289274,"spread":0.2131103090933615,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001828727,0.0002874359,0.0003049727,0.00007314258,0.0002848137,0.000222552,0.0001313871,0.0002338206,0.00000711131],"category_scores_gemma":[0.00005243151,0.000240008,0.00006684857,0.00007621268,0.000121142,0.00001345385,0.0003104039,0.00006904761,0.00000345139],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000009812952,"about_ca_system_score_gemma":0.00008223773,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001869863,"about_ca_topic_score_gemma":0.00002034435,"domain_scores_codex":[0.9987506,0.00002548497,0.0004722634,0.0003347277,0.00008613055,0.0003308266],"domain_scores_gemma":[0.999312,0.00005201395,0.0001831284,0.0002502432,0.00006005749,0.0001425172],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0002541475,0.00002607645,0.001376458,0.0001347614,0.0001483649,8.272267e-7,0.0006680595,0.00004440969,0.9612916,0.000104934,0.004449815,0.03150057],"study_design_scores_gemma":[0.007747516,0.001140372,0.02304511,0.00007040171,0.000141917,0.0000175233,0.0008438376,0.05181618,0.4656768,0.001212645,0.446809,0.001478738],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9918849,0.0006127473,0.0003700047,0.0003209052,0.0002182062,0.004482165,0.002029269,0.00002120169,0.00006054084],"genre_scores_gemma":[0.8414009,0.01484901,0.1243023,0.001293662,0.002801621,0.003822669,0.01076649,0.0001863659,0.000576978],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4956148,"threshold_uncertainty_score":0.9787241,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4414887972","doi":"10.1093/nargab/lqaf132","title":"VIRUS-MVP: a framework for comprehensive surveillance of viral mutations and their functional impacts","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Viral Infections and Outbreaks Research","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Public Health Agency of Canada; University of British Columbia; University of Calgary; Simon Fraser University","funders":"Canadian Institutes of Health Research; Simon Fraser University; Michael Smith Health Research BC; Genome Canada","keywords":"Genomics; Functional genomics; Annotation; Resource (disambiguation); Mutation; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Genome; Modular design","authors":[{"name":"Muhammad Zohaib Anwar","is_ca":true},{"name":"Ivan S Gill","is_ca":true},{"name":"Madeline Iseminger","is_ca":true},{"name":"Anoosha Sehar","is_ca":true},{"name":"Emma Griffiths","is_ca":true},{"name":"Damion Dooley","is_ca":true},{"name":"Jun Duan","is_ca":true},{"name":"Khushi Vora","is_ca":true},{"name":"Gary Van Domselaar","is_ca":true},{"name":"Fiona S. L. Brinkman","is_ca":true},{"name":"Paul M. K. Gordon","is_ca":true},{"name":"William Hsiao","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03253870125687595,"gpt":0.3202955777649885,"spread":0.2877568765081125,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001127779,0.00008771023,0.0002014555,0.0001326544,0.0001336027,0.00003056543,0.00002326177,0.00008440701,0.000006609346],"category_scores_gemma":[0.000071136,0.00006417348,0.00005359665,0.0001310549,0.0001176208,0.000055062,0.00004385346,0.0001144828,0.000001375523],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002990017,"about_ca_system_score_gemma":0.0001175976,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002212908,"about_ca_topic_score_gemma":0.00001227168,"domain_scores_codex":[0.9994869,0.000007582893,0.0002444732,0.00007024897,0.00005773231,0.0001330115],"domain_scores_gemma":[0.9993763,0.0002072442,0.00006598345,0.00009390806,0.0001929684,0.00006360505],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.005102556,0.001230525,0.06202413,0.01247132,0.003280894,0.000006403369,0.01453067,0.0006976476,0.178887,0.1749232,0.02152434,0.5253214],"study_design_scores_gemma":[0.01028647,0.003459458,0.3107401,0.000767854,0.0002554602,0.0002283056,0.0110139,0.507126,0.02881665,0.03831022,0.08817405,0.0008214896],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9314887,0.0007466684,0.06545635,0.0006408175,0.0001053164,0.0005043299,0.0002772144,0.0000198691,0.0007607546],"genre_scores_gemma":[0.9827673,0.0007778791,0.01579671,0.0004290304,0.00004570257,0.00001229855,0.00007724118,0.000006214497,0.00008763287],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5244998,"threshold_uncertainty_score":0.2616918,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3097162993","doi":"10.1093/nargab/lqaa086","title":"aliFreeFoldMulti: alignment-free method to predict secondary structures of multiple RNA homologs","year":2020,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Sherbrooke","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"RNA; Sequence (biology); Multiple sequence alignment; Set (abstract data type); Structural alignment; Algorithm; Computer science; Nucleic acid structure; Protein secondary structure; Sequence alignment; Nucleic acid secondary structure; Computational biology; Mathematics; Biology; Genetics; Peptide sequence","authors":[{"name":"Marc-André Bossanyi","is_ca":true},{"name":"Valentin Carpentier","is_ca":true},{"name":"Jean-Pierre Séhi Glouzon","is_ca":true},{"name":"Aïda Ouangraoua","is_ca":true},{"name":"Yoann Anselmetti","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01213632394337873,"gpt":0.229527913349563,"spread":0.2173915894061843,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002335929,0.0001859505,0.000246143,0.0000400436,0.00006798105,0.00002727714,0.0003584633,0.0001776809,0.00002202179],"category_scores_gemma":[0.000146063,0.0001619181,0.00008369228,0.00006470382,0.00004382135,0.000007305831,0.0003868291,0.00007444627,0.000004813729],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000008750783,"about_ca_system_score_gemma":0.00006583856,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001413884,"about_ca_topic_score_gemma":0.000004448235,"domain_scores_codex":[0.9990059,0.00003954532,0.0004057256,0.0002002518,0.0001352571,0.0002133388],"domain_scores_gemma":[0.9991934,0.0000269189,0.0001633382,0.0003592446,0.00005329269,0.0002037932],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001459569,0.00001429613,0.0001623749,0.0001369026,0.00009704789,0.000001202347,0.0007714853,0.0001120494,0.9376894,0.0002045143,0.002654982,0.05800983],"study_design_scores_gemma":[0.0009002009,0.0007020151,0.000215916,0.00001065869,0.00003374318,0.0000144338,0.0005062608,0.003443134,0.9538398,0.0004423784,0.03963634,0.0002550949],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8099552,0.0009870756,0.1844983,0.0009686768,0.0001748445,0.0008923212,0.0009832418,0.00002880493,0.001511551],"genre_scores_gemma":[0.5307975,0.0002730984,0.4664221,0.002132756,0.0001766151,0.0000151436,0.0001007325,0.00002801143,0.00005410049],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2819237,"threshold_uncertainty_score":0.6602829,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4383216982","doi":"10.1093/nargab/lqad065","title":"GCPBayes pipeline: a tool for exploring pleiotropy at the gene level","year":2023,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genetic Mapping and Diversity in Plants and Animals","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Cancer Institute; Ligue Contre le Cancer; Ovarian Cancer Research Fund; National Institutes of Health; Cancer Research UK; Government of Canada; Institut National de la Santé et de la Recherche Médicale; Fondation du cancer du sein du Québec; Canadian Institutes of Health Research; Gray Foundation; Genome Canada; European Commission; Breast Cancer Research Foundation","keywords":"Pleiotropy; Pipeline (software); Context (archaeology); Computer science; Set (abstract data type); Genome-wide association study; Computational biology; Gene; Genome; Data mining; Biology; Phenotype; Genetics; Single-nucleotide polymorphism","authors":[{"name":"Yazdan Asgari","is_ca":false},{"name":"Pierre‐Emmanuel Sugier","is_ca":false},{"name":"Taban Baghfalaki","is_ca":false},{"name":"Elise A. Lucotte","is_ca":false},{"name":"Mojgan Karimi","is_ca":false},{"name":"Mohammed Sedki","is_ca":false},{"name":"Amélie Ngo","is_ca":false},{"name":"Benoît Liquet","is_ca":false},{"name":"Thérèse Truong","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.06137306226799964,"gpt":0.2430030234408659,"spread":0.1816299611728662,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002381184,0.0001213036,0.0001068298,0.00003449927,0.0003478623,0.00005965881,0.0001390604,0.0000756571,0.000006066342],"category_scores_gemma":[0.00004319716,0.00008657426,0.00007273607,0.00005806001,0.00004992407,0.000004592513,0.0002393887,0.00003880497,0.00003104081],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000009303641,"about_ca_system_score_gemma":0.00003147017,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004562996,"about_ca_topic_score_gemma":0.000007933807,"domain_scores_codex":[0.9993419,0.000006997463,0.000198424,0.0001329799,0.00008115605,0.0002385468],"domain_scores_gemma":[0.9996243,0.00002569743,0.00006810124,0.0001842095,0.00004761326,0.00005012709],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0006437285,0.00007560135,0.004620864,0.0005576327,0.0003442907,0.000006825036,0.003661058,0.001274553,0.4938788,0.0009823637,0.3598231,0.1341312],"study_design_scores_gemma":[0.001292288,0.0004122677,0.002803725,0.0000199914,0.00006215376,0.00005070993,0.001927355,0.02577144,0.06501906,0.0001755177,0.9019638,0.0005016461],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9942534,0.000283251,0.003799117,0.0004172391,0.0001751085,0.000234795,0.0005446959,0.000018922,0.0002734954],"genre_scores_gemma":[0.9223756,0.01670451,0.03385774,0.002317697,0.001289864,0.0001305475,0.002369014,0.00006186706,0.02089315],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5421407,"threshold_uncertainty_score":0.3530396,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3158542203","doi":"10.1093/nargab/lqab035","title":"Wavelet Screening identifies regions highly enriched for differentially methylated loci for orofacial clefts","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Cleft Lip and Palate Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Norges Forskningsråd; Réseau de cancérologie Rossy","keywords":"DNA methylation; Computational biology; Epigenetics; Differentially methylated regions; Biology; Methylation; Genetics; Bioinformatics; Computer science; Gene; Gene expression","authors":[{"name":"William R. P. Denault","is_ca":false},{"name":"Julia Romanowska","is_ca":false},{"name":"Øystein Ariansen Haaland","is_ca":false},{"name":"Robert Lyle","is_ca":false},{"name":"Jack A. Taylor","is_ca":false},{"name":"Zongli Xu","is_ca":false},{"name":"Rolv T. Lie","is_ca":false},{"name":"Håkon K. Gjessing","is_ca":false},{"name":"Astanand Jugessur","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.02688918016157574,"gpt":0.2843459265464833,"spread":0.2574567463849076,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002790982,0.0001717088,0.0002207126,0.00005665054,0.0002995087,0.0001533481,0.000142937,0.0002310266,0.000006522585],"category_scores_gemma":[0.0001342371,0.0001614107,0.0001628923,0.00008311485,0.00006719967,0.000009920172,0.000163004,0.00008666944,0.000002183537],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001392858,"about_ca_system_score_gemma":0.0001422638,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005862633,"about_ca_topic_score_gemma":0.00004805345,"domain_scores_codex":[0.9989169,0.00002290812,0.0003679173,0.0002218456,0.0001163687,0.0003541253],"domain_scores_gemma":[0.9991648,0.0000606807,0.0001121577,0.0002468035,0.0002952819,0.0001203207],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004242891,0.0000742783,0.0004975665,0.0005381499,0.000496974,0.000003255225,0.0008451161,0.00003815248,0.9430974,0.00125808,0.009120217,0.04360658],"study_design_scores_gemma":[0.00868538,0.001172642,0.004789487,0.00006999141,0.0003588771,0.00009941668,0.002784546,0.05635982,0.4199678,0.00381467,0.500273,0.001624364],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8362511,0.0004638586,0.1612045,0.000553838,0.0002174666,0.0006353813,0.0004136362,0.00001803097,0.0002421693],"genre_scores_gemma":[0.9124071,0.003229552,0.07592978,0.0005380159,0.0006497594,0.00008153359,0.004533893,0.00006534798,0.002565052],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5231296,"threshold_uncertainty_score":0.6582137,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399154467","doi":"10.1093/nargab/lqae060","title":"G-quadruplex propensity in <i>H. neanderthalensis</i>, <i>H. sapiens</i> and Denisovans mitochondrial genomes","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Vermont Agency of Natural Resources; Institut National Du Cancer; Grantová Agentura České Republiky; Agence Nationale de la Recherche; CNIB","keywords":"Homo sapiens; Mitochondrial DNA; Genome; Genetics; Biology; DNA; Evolutionary biology; Gene; Geography","authors":[{"name":"Václav Brázda","is_ca":false},{"name":"Lucie Šislerová","is_ca":false},{"name":"Anne Cucchiarini","is_ca":false},{"name":"Jean‐Louis Mergny","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01166739839285171,"gpt":0.2164049901412345,"spread":0.2047375917483828,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002417795,0.0002085702,0.0002153693,0.00008779331,0.00009669692,0.0001759723,0.0001087987,0.0001672162,0.000007221936],"category_scores_gemma":[0.00002443803,0.0001772914,0.00005753394,0.00008866064,0.00008999521,0.00001746639,0.0001671939,0.0001107658,0.00001114891],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001714835,"about_ca_system_score_gemma":0.00008733697,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001433409,"about_ca_topic_score_gemma":0.00005656988,"domain_scores_codex":[0.9989807,0.00002804925,0.0003384548,0.0002744487,0.0001041891,0.0002741471],"domain_scores_gemma":[0.9995747,0.0000214933,0.00005325547,0.0002172329,0.00002984779,0.0001035102],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001894162,0.00004683205,0.001067197,0.000484787,0.0001134117,0.00004104088,0.001548291,0.00002461473,0.8372644,0.0009261178,0.001488603,0.1568053],"study_design_scores_gemma":[0.002541568,0.001113031,0.001251718,0.0002678131,0.0001464624,0.001021233,0.001536358,0.0180897,0.3902021,0.002320986,0.5797258,0.001783222],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9861283,0.009491279,0.002355044,0.0002738422,0.0003176286,0.0004691353,0.00005056206,0.00002483526,0.0008893515],"genre_scores_gemma":[0.9606297,0.011554,0.02610286,0.0007118172,0.0003634345,0.00002574531,0.00008626381,0.00004696719,0.0004791789],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5782372,"threshold_uncertainty_score":0.7229733,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4400738267","doi":"10.1093/nargab/lqae084","title":"<b> s </b>CIRCLE—An interactive visual exploration tool for single cell RNA-Seq data","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Bayerisches Staatsministerium für Wissenschaft und Kunst","keywords":"Metadata; Visualization; Computer science; Data visualization; Focus (optics); RNA-Seq; Software; Interactive visualization; Dimensionality reduction; Data mining; Gene; Artificial intelligence; World Wide Web; Transcriptome; Gene expression; Biology; Programming language","authors":[{"name":"Maximilian Seeger","is_ca":false},{"name":"Erich Schöls","is_ca":false},{"name":"Lars Barquist","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.03484221898688085,"gpt":0.2773704812826971,"spread":0.2425282622958163,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002247769,0.0001652887,0.0001372436,0.00005615893,0.0001094136,0.0002788364,0.0002261476,0.0001547868,0.000003591973],"category_scores_gemma":[0.00002615903,0.0001613469,0.00005522596,0.0000549259,0.00004684008,0.00008498605,0.0001300023,0.0000921676,0.00000952068],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002662644,"about_ca_system_score_gemma":0.00009395327,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009191744,"about_ca_topic_score_gemma":0.00002213885,"domain_scores_codex":[0.9990881,0.00001412376,0.0003276957,0.0002804897,0.00008309377,0.0002064741],"domain_scores_gemma":[0.9994286,0.00002485257,0.00006996319,0.000334819,0.00006685102,0.00007491617],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001721625,0.0001532723,0.00002188421,0.000301914,0.00006107894,0.000001596098,0.001142056,0.00005756763,0.8472745,0.0000646628,0.001265328,0.149484],"study_design_scores_gemma":[0.001156982,0.001859619,0.00002175564,0.00006239024,0.000123284,0.00003243606,0.001571764,0.409579,0.4442365,0.0003214273,0.1403437,0.0006911106],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8242715,0.0008951989,0.1717659,0.0001342654,0.0008189637,0.0005556175,0.0006206239,0.00005287143,0.0008850771],"genre_scores_gemma":[0.975026,0.0006501449,0.02013178,0.0003216684,0.0005121899,0.00001640093,0.003124433,0.00004360383,0.0001737461],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4095214,"threshold_uncertainty_score":0.6579535,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W7117139568","doi":"10.1093/nargab/lqaf189","title":"Life at the extremes: maximally divergent microbes with similar genomic signatures linked to extreme environments","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Guelph; Western University; University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; Alliance de recherche numérique du Canada","keywords":"Taxonomic rank; Genome; Extreme environment; Metagenomics; Genomics; Biological classification; Adaptation (eye); Phylogenomics; Trait","authors":[{"name":"Monireh Safari","is_ca":true},{"name":"Joseph Butler","is_ca":true},{"name":"Gurjit S. Randhawa","is_ca":true},{"name":"Kathleen A. Hill","is_ca":true},{"name":"Lila Kari","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01003646980641636,"gpt":0.1977671714855403,"spread":0.187730701679124,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001712329,0.0002945685,0.0002243146,0.00006887913,0.0003963387,0.00007032348,0.0003186696,0.0001438546,0.00001066228],"category_scores_gemma":[0.00003313809,0.0002001846,0.00008019758,0.0000881254,0.0001535305,0.000002050539,0.0007745752,0.0001008643,0.00001879758],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005450862,"about_ca_system_score_gemma":0.0001167897,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001245545,"about_ca_topic_score_gemma":0.00009483744,"domain_scores_codex":[0.998859,0.00002237257,0.0003591665,0.0002970092,0.0001197503,0.0003426549],"domain_scores_gemma":[0.9992071,0.00002141508,0.0001192965,0.0004476128,0.0000385503,0.0001659878],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004166916,0.0001144606,0.00565197,0.00009223555,0.0006623861,0.000002664967,0.001230407,0.001778646,0.9661888,0.0001210865,0.0149543,0.008786294],"study_design_scores_gemma":[0.001871366,0.0005639239,0.02536507,0.00003597814,0.0001812016,0.00001450042,0.001187353,0.001062816,0.02489216,0.00008615603,0.9440604,0.0006790949],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9897038,0.00572774,0.001436573,0.001235465,0.0001487245,0.0005771961,0.0001021403,0.000004658807,0.001063739],"genre_scores_gemma":[0.9709045,0.008414564,0.01017148,0.00759241,0.0001456065,0.00004714476,0.0001006434,0.00003865389,0.002585033],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9412967,"threshold_uncertainty_score":0.8163292,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3172664492","doi":"10.1093/nargab/lqab042","title":"Retracted and Replaced: Known sequence features can explain half of all human gene ends","year":2021,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA Research and Splicing","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":true,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Occupational Cancer Research Centre; University of Toronto","funders":"Canadian Institutes of Health Research","keywords":"Polyadenylation; Gene; Coding region; Genetics; Biology; Sequence (biology); Computational biology; Primary transcript; Context (archaeology); Regulatory sequence; Messenger RNA; Gene expression; Alternative splicing","authors":[{"name":"Aleksei Shkurin","is_ca":true},{"name":"Timothy R. Hughes","is_ca":true}],"retraction":{"nature":"Retraction","openalex_flagged":true},"screen_n_in":null,"score":{"opus":0.01873758104958645,"gpt":0.2740747268087642,"spread":0.2553371457591778,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002358387,0.0001234311,0.0001621713,0.00004506684,0.000102776,0.00005348956,0.00009819498,0.0001765474,0.000005538062],"category_scores_gemma":[0.00007435699,0.0001122505,0.00003945509,0.00006430062,0.00009811106,0.000007111012,0.0001662186,0.0001419671,4.902519e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001626924,"about_ca_system_score_gemma":0.0001109265,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000750508,"about_ca_topic_score_gemma":0.0001458718,"domain_scores_codex":[0.9991982,0.00002387986,0.0002709099,0.0001632866,0.0001189588,0.000224756],"domain_scores_gemma":[0.999374,0.00001248051,0.0001054853,0.0002681094,0.0001125102,0.0001274414],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00002708877,0.00001508024,0.0003289312,0.00008411638,0.00004786637,0.00001102968,0.0002869562,0.000005430581,0.9938865,0.0001411798,0.0008129434,0.004352864],"study_design_scores_gemma":[0.000700645,0.0003576901,0.001344308,0.00002915997,0.00002248659,0.0003323828,0.0006825298,0.001255314,0.9704661,0.00006841301,0.02448635,0.0002546259],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9974933,0.001145096,0.00009500977,0.0003085496,0.00002831932,0.00009169768,0.00008478454,0.000005131177,0.0007480801],"genre_scores_gemma":[0.9854097,0.002596758,0.009781806,0.0003407394,0.00009805502,0.000003681715,0.0004570195,0.00001475373,0.001297527],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0236734,"threshold_uncertainty_score":0.4577443,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4412492594","doi":"10.1093/nargab/lqaf099","title":"Modeling RNA duplex dynamics with Gibbs sampling enhances base-pair prediction accuracy and reveals structural activity profiles","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Institute for Research in Immunology and Cancer","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Google","keywords":"RNA; Base pair; Duplex (building); Nucleic acid secondary structure; Nucleic acid structure; Biological system; Statistical physics; Algorithm; Computational biology; Computer science; Physics; Biology; Genetics; DNA","authors":[{"name":"Simon Chasles","is_ca":true},{"name":"François Major","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01157170296863191,"gpt":0.2411867423526377,"spread":0.2296150393840058,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002322006,0.0001890902,0.0001831682,0.00005956993,0.000246774,0.0001081797,0.0001025075,0.0001636712,0.000002020133],"category_scores_gemma":[0.00005562122,0.0001507386,0.00003465532,0.00006498052,0.00005886053,0.00002813679,0.0001115294,0.0000954202,4.574124e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002760017,"about_ca_system_score_gemma":0.00008678288,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001786208,"about_ca_topic_score_gemma":0.00002261838,"domain_scores_codex":[0.9992129,0.00002304391,0.0002586404,0.0002089282,0.00009177162,0.0002046888],"domain_scores_gemma":[0.9995142,0.000026184,0.0001208168,0.0002077459,0.00006912417,0.00006196529],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000337152,0.00001954086,0.001103991,0.0004935352,0.0001534736,6.792883e-7,0.0003077891,0.002624362,0.8642547,0.0007531387,0.00003897495,0.1299127],"study_design_scores_gemma":[0.0005152009,0.0002700121,0.0004554108,0.000117175,0.0000594348,0.00002175807,0.0006501004,0.7684335,0.2281633,0.0006440856,0.0003688834,0.0003010519],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8907549,0.0003893873,0.1080516,0.0001194523,0.0000762624,0.0003271126,0.0001021279,0.00001741139,0.0001616753],"genre_scores_gemma":[0.9563514,0.0007798373,0.04248518,0.0001016323,0.00006398073,0.00001682118,0.00009360646,0.00001241659,0.00009509536],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7658092,"threshold_uncertainty_score":0.6146943,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4417268800","doi":"10.1093/nargab/lqaf170","title":"<tt>SAFARI</tt> : pangenome alignment of ancient DNA using purine/pyrimidine encodings","year":2025,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université Laval","funders":"National Human Genome Research Institute; Novo Nordisk Fonden; Novo Nordisk","keywords":"Spurious relationship; Ancient DNA; Extant taxon; Genome; DNA; Divergence (linguistics)","authors":[{"name":"Joshua Daniel Rubin","is_ca":false},{"name":"Jan van Waaij","is_ca":false},{"name":"Louis Kraft","is_ca":false},{"name":"Jouni Sirén","is_ca":false},{"name":"Peter Wad Sackett","is_ca":false},{"name":"Gabriel Renaud","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.01243245081278927,"gpt":0.2329977021079812,"spread":0.220565251295192,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002335499,0.0002110831,0.0002723559,0.00009881167,0.0001522583,0.00003475423,0.0001921611,0.0001286829,0.00000516249],"category_scores_gemma":[0.00002158599,0.0001943786,0.00009590864,0.0001140713,0.0001247203,0.000002301011,0.0004101328,0.00005666463,0.00000183627],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004310752,"about_ca_system_score_gemma":0.0001065433,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001853071,"about_ca_topic_score_gemma":0.000007917197,"domain_scores_codex":[0.9988889,0.00001197506,0.0005026488,0.0002063572,0.0001076229,0.0002825397],"domain_scores_gemma":[0.9993081,0.00001121116,0.0001961939,0.0003073692,0.000102746,0.00007436048],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00004828165,0.00005949494,0.001863816,0.0001881528,0.0001951251,7.852864e-7,0.0005319791,0.0002717566,0.98919,0.0006469906,0.0005150017,0.006488635],"study_design_scores_gemma":[0.002209206,0.0007897801,0.00979301,0.0001042079,0.0002987131,0.00005350216,0.002791954,0.01117336,0.6977842,0.000605974,0.273444,0.0009520366],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9923861,0.003352189,0.001240802,0.0001057831,0.00022194,0.0002657093,0.00008883524,0.000003865971,0.002334795],"genre_scores_gemma":[0.9857775,0.002753742,0.01064385,0.0005348115,0.00007473002,0.000005872672,0.00003787774,0.00001484905,0.0001567396],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2914058,"threshold_uncertainty_score":0.7926528,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4405282922","doi":"10.1093/nargab/lqae161","title":"Water-mediated ribonucleotide–amino acid pairs and higher-order structures at the RNA–protein interface: analysis of the crystal structure database and a topological classification","year":2024,"lang":"en","type":"article","venue":"NAR Genomics and Bioinformatics","topic":"RNA and protein synthesis mechanisms","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Interface (matter); Topology (electrical circuits); Amino acid; Database; Order (exchange); Ribonucleotide; Computer science; Crystal structure; RNA; Computational biology; Data mining; Biology; Chemistry; Crystallography; Mathematics; Biochemistry; Nucleotide; Combinatorics; Operating system; Gene","authors":[{"name":"Raman Jangra","is_ca":false},{"name":"John F. Trant","is_ca":true},{"name":"Purshotam Sharma","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.0112042506949473,"gpt":0.2291308502527272,"spread":0.2179265995577799,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002143352,0.0001614683,0.0001726557,0.00005459993,0.0001853148,0.00009121328,0.0001468239,0.0001696226,0.00004319288],"category_scores_gemma":[0.00003543725,0.00007728262,0.00005924172,0.0001295425,0.0002915308,0.00001397045,0.0003497309,0.0001077846,5.850803e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001393349,"about_ca_system_score_gemma":0.00002373334,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001080799,"about_ca_topic_score_gemma":0.000037791,"domain_scores_codex":[0.999204,0.00005305937,0.0002799953,0.0001957168,0.0001108647,0.0001564127],"domain_scores_gemma":[0.9994993,0.00002095493,0.00009138692,0.0002976422,0.00003839588,0.00005235771],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00004156567,0.000004520067,0.0001320047,0.00008281651,0.0002070624,4.261094e-7,0.0003514953,0.00001129865,0.9932802,0.0007557891,0.00006920066,0.00506359],"study_design_scores_gemma":[0.0002908505,0.0001876366,0.004239618,0.00002849308,0.0004454454,0.00002733617,0.00075437,0.01244866,0.956751,0.0006737652,0.02387741,0.0002753551],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9954773,0.001966245,0.001013487,0.000634925,0.00008352553,0.0002673589,0.0004740169,0.000007797324,0.00007539854],"genre_scores_gemma":[0.9975842,0.000553567,0.001350171,0.0001468083,0.00003701705,0.000006122629,0.0001950612,0.000009813472,0.0001172978],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03652918,"threshold_uncertainty_score":0.3151494,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}