{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009497161,0.001073672,0.0007606391,0.0003831752,0.000714979,0.0008389238,0.003000429,0.001559419,0.01045392],"category_scores_gemma":[0.003072535,0.0007802433,0.001291151,0.000654866,0.0004616512,0.001179377,0.001213256,0.002067212,0.002228819],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00117407,"about_ca_system_score_gemma":0.002725445,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008628036,"about_ca_topic_score_gemma":0.007192368,"domain_scores_codex":[0.9996321,0.00007078062,0.00002786986,0.00006102695,0.0001518857,0.00005617629],"domain_scores_gemma":[0.9989921,0.000475608,0.00006448113,0.0001132054,0.0002411365,0.0001134868],"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":[0.0006081184,0.0003129188,0.007851405,0.0008218999,0.0002997571,0.0003997294,0.0005186273,0.841994,0.03348447,0.02856033,0.03364049,0.05150836],"study_design_scores_gemma":[0.00007584555,0.00005036743,0.0003020893,0.00001892381,0.00002527836,0.00004099294,0.00002549704,0.9626763,0.01048342,0.002815807,0.02345364,0.0000318228],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.09118356,0.0009122098,0.8099864,0.0008239198,0.0006811034,0.0005409612,0.01140594,0.05602521,0.02844068],"genre_scores_gemma":[0.3745959,0.001242953,0.5618132,0.001058316,0.0001206077,0.002731434,0.02300446,0.01449513,0.02093805],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01045392,"threshold_uncertainty_score":0.03497183,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.09577065,0.002021484,0.002291886,0.0100603,0.003654592,0.01049046,0.006659635,0.005033288,0.005856654],"category_scores_gemma":[0.3399438,0.00154872,0.002318906,0.009425648,0.003003393,0.01042819,0.006821671,0.005804354,0.00765001],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002929872,"about_ca_system_score_gemma":0.009625876,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00202205,"about_ca_topic_score_gemma":0.001273102,"domain_scores_codex":[0.8351256,0.06949483,0.03781231,0.008328627,0.04636929,0.002869332],"domain_scores_gemma":[0.6667243,0.11421,0.03083031,0.08505952,0.09933067,0.003845249],"domain_codex":null,"domain_gemma":"reporting","domain_candidate":"reporting","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001028777,0.0005682214,0.01550214,0.00350509,0.0003297344,0.0006025633,0.002237392,0.02326021,0.02755965,0.4039766,0.1123536,0.409076],"study_design_scores_gemma":[0.0002953442,0.001126339,0.006697479,0.003284046,0.0003057652,0.002309495,0.001335119,0.114852,0.07189579,0.3195557,0.4777072,0.000635654],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.007747853,0.001262359,0.9622335,0.001818466,0.0006377232,0.001051068,0.005345006,0.01070775,0.009196264],"genre_scores_gemma":[0.0426852,0.0007970653,0.9385365,0.0005422104,0.000422539,0.002540367,0.01134832,0.001586727,0.001541133],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9042293,"threshold_uncertainty_score":0.5064896,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001155094,0.0001859325,0.0003544857,0.0003612538,0.0002518765,0.0004825626,0.0001701399,0.0003215641,0.001779436],"category_scores_gemma":[0.0004722328,0.000175691,0.000354344,0.0004505341,0.0002505317,0.0005215495,0.0001467302,0.0002657266,0.0008514845],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001779271,"about_ca_system_score_gemma":0.0002645714,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009082075,"about_ca_topic_score_gemma":0.0011783,"domain_scores_codex":[0.9998975,0.00002244736,0.000005322743,0.00004692095,0.0000127657,0.00001504778],"domain_scores_gemma":[0.9998347,0.00006373604,0.00004456164,0.00001746892,0.00002085404,0.00001870365],"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":[0.001585591,0.0001052817,0.1249076,0.001772266,0.0002214767,0.002224069,0.0007280359,0.03487237,0.6290202,0.01150957,0.006046867,0.1870067],"study_design_scores_gemma":[0.0001006075,0.001185711,0.4147647,0.000689329,0.0007375168,0.007501308,0.002577591,0.2336145,0.1912853,0.07136694,0.07595462,0.0002218568],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9459961,0.006808239,0.03682528,0.001095047,0.00005585098,0.00002689468,0.004773993,0.0006724676,0.003746128],"genre_scores_gemma":[0.9794793,0.002174416,0.01446383,0.0002528912,0.00003194587,0.00002569301,0.002346844,0.00005872625,0.001166367],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001779436,"threshold_uncertainty_score":0.005952835,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002416654,0.001282305,0.001054918,0.002177089,0.0008001827,0.001333648,0.002497032,0.001137327,0.003490157],"category_scores_gemma":[0.004075956,0.0006162969,0.002102435,0.001762988,0.0005600567,0.001114266,0.001985089,0.001734646,0.001605468],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001143328,"about_ca_system_score_gemma":0.002212358,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006455848,"about_ca_topic_score_gemma":0.01437117,"domain_scores_codex":[0.9990122,0.0003650922,0.00005876609,0.0002581052,0.0002525105,0.00005334307],"domain_scores_gemma":[0.9988198,0.0007125761,0.0000698709,0.0001154728,0.000202412,0.00007979857],"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":[0.0007660582,0.0003861034,0.00685945,0.001694166,0.0007556996,0.0007077458,0.0002956182,0.4938774,0.01201536,0.0493823,0.04052505,0.3927351],"study_design_scores_gemma":[0.00003346392,0.00003980352,0.0003138442,0.00002716257,0.0000237899,0.00005795239,0.00002311318,0.9707655,0.001242866,0.02016173,0.007297733,0.00001295528],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006546342,0.0006519383,0.9758553,0.0002906485,0.00005000878,0.0001615987,0.002865576,0.01242563,0.001153077],"genre_scores_gemma":[0.06296045,0.0004448566,0.9227157,0.000243318,0.00004729167,0.0003899354,0.01126067,0.0005621911,0.001375659],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006455848,"threshold_uncertainty_score":0.01283652,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003997192,0.002055055,0.001973313,0.001944963,0.001112438,0.002189891,0.001978203,0.0015563,0.006596443],"category_scores_gemma":[0.007493102,0.001130046,0.002152596,0.001476003,0.001016738,0.001247566,0.002132165,0.002827195,0.006164935],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006941632,"about_ca_system_score_gemma":0.001722106,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002099591,"about_ca_topic_score_gemma":0.0039325,"domain_scores_codex":[0.9983006,0.0002731935,0.00007871063,0.0007865889,0.0004615949,0.00009929282],"domain_scores_gemma":[0.997244,0.001508516,0.0003518268,0.0004888397,0.0002872935,0.0001195928],"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":[0.001493981,0.0002498979,0.02600822,0.005122888,0.002236638,0.001269787,0.001476722,0.05911351,0.4531073,0.01438846,0.11519,0.3203425],"study_design_scores_gemma":[0.0002934826,0.0003447559,0.01403776,0.0002208166,0.0005047353,0.001351089,0.0002115947,0.4896124,0.3625746,0.02779996,0.1026127,0.000436126],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.02534636,0.0009202833,0.8754715,0.0002904469,0.0002374487,0.0002338713,0.01316394,0.08225913,0.002077056],"genre_scores_gemma":[0.1039915,0.0008977132,0.8505668,0.0008600961,0.0001052495,0.001212886,0.0234054,0.01603645,0.002923799],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.006596443,"threshold_uncertainty_score":0.02206731,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01557363,0.001711442,0.001569307,0.005308238,0.000586858,0.001883337,0.001534521,0.001529415,0.001431418],"category_scores_gemma":[0.0233735,0.0002965961,0.001952807,0.002386928,0.0005091978,0.001520396,0.001356286,0.001319461,0.0006018497],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001116663,"about_ca_system_score_gemma":0.00146564,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008055055,"about_ca_topic_score_gemma":0.008830499,"domain_scores_codex":[0.9935051,0.002705821,0.0007204809,0.001282339,0.001501336,0.0002849597],"domain_scores_gemma":[0.9711603,0.0237345,0.000738184,0.001632641,0.002172128,0.0005622724],"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":[0.004104268,0.0007913118,0.3325195,0.002019563,0.005334178,0.0006775485,0.0002596207,0.2315205,0.005893688,0.003329105,0.01924325,0.3943075],"study_design_scores_gemma":[0.0003414163,0.00102746,0.08681041,0.0003172361,0.0009494914,0.001040557,0.0002317147,0.8895909,0.006709996,0.004034313,0.008788117,0.0001584499],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8602415,0.03807263,0.07182051,0.002674946,0.0007237916,0.0002844736,0.01057103,0.005167252,0.01044379],"genre_scores_gemma":[0.9193673,0.003949122,0.05324776,0.0006237761,0.0001968822,0.0001412391,0.02038139,0.0003441144,0.001748558],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01557363,"threshold_uncertainty_score":0.08236223,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001196744,0.0003018168,0.0003496819,0.0003616111,0.000421639,0.000309848,0.0002112095,0.0002405858,0.001171599],"category_scores_gemma":[0.00007856395,0.0002127417,0.0005545949,0.0002678733,0.0002226062,0.0001902667,0.0002955984,0.0004730005,0.0004168657],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005575212,"about_ca_system_score_gemma":0.0002263489,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005535294,"about_ca_topic_score_gemma":0.00676513,"domain_scores_codex":[0.9998567,0.000005743365,0.000008825296,0.0000625739,0.00003758623,0.00002863042],"domain_scores_gemma":[0.9998989,0.000008191934,0.00003281565,0.000005955822,0.00001612089,0.00003801353],"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":[0.00007627222,0.000006980564,0.00132617,0.00004374285,0.00001031687,0.00004743176,0.0000278867,0.00002843997,0.9976752,0.0000375258,0.0001126734,0.0006074484],"study_design_scores_gemma":[0.0000608155,0.0001454751,0.3403573,0.0000258301,0.0001048549,0.0005369108,0.0002855836,0.004774192,0.6454446,0.0002168439,0.007985955,0.00006169556],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9894907,0.002829563,0.003159284,0.0001893624,0.0000908399,0.00001546206,0.002689359,0.000207219,0.001328321],"genre_scores_gemma":[0.9895802,0.0005761124,0.002082842,0.0002699,0.00001966264,0.0000248528,0.003645193,0.00008357564,0.003717624],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005535294,"threshold_uncertainty_score":0.01100618,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009040083,0.0009569684,0.002057243,0.004407595,0.0009870803,0.001923978,0.001429187,0.001763168,0.001606673],"category_scores_gemma":[0.03384924,0.000627365,0.001810431,0.004769485,0.001650544,0.00219085,0.00197167,0.001973418,0.001257854],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00141268,"about_ca_system_score_gemma":0.00169541,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002149444,"about_ca_topic_score_gemma":0.002501954,"domain_scores_codex":[0.9896185,0.002733118,0.0008831752,0.002707632,0.003717058,0.0003405981],"domain_scores_gemma":[0.9780739,0.01208962,0.004058376,0.002548612,0.002784282,0.0004452937],"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":[0.001078148,0.000342554,0.06574888,0.001752892,0.002133114,0.0006590595,0.0007382433,0.3316427,0.132947,0.06540327,0.0126003,0.3849538],"study_design_scores_gemma":[0.00004318782,0.0003183054,0.03325985,0.0000997228,0.0001763959,0.0008671411,0.000112807,0.8562729,0.03380859,0.06489047,0.00990236,0.0002482921],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0196724,0.000479341,0.9771366,0.00009556951,0.00003934959,0.00008393195,0.0007210238,0.001072062,0.0006996942],"genre_scores_gemma":[0.3925576,0.0005706032,0.5984085,0.0003967217,0.0002095498,0.0006117687,0.004725908,0.0008687357,0.001650643],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009040083,"threshold_uncertainty_score":0.04780912,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0201402,0.004887388,0.004479834,0.009011571,0.002220207,0.01007706,0.01704995,0.004830769,0.0200118],"category_scores_gemma":[0.0302803,0.00387586,0.008381544,0.00801459,0.003671907,0.01410996,0.01836657,0.005782382,0.01666444],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00351689,"about_ca_system_score_gemma":0.004984847,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01677063,"about_ca_topic_score_gemma":0.01380289,"domain_scores_codex":[0.9903515,0.001871961,0.001572559,0.002382892,0.003084523,0.0007365821],"domain_scores_gemma":[0.9865991,0.005749542,0.0008241055,0.004342023,0.001400344,0.001084973],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.003137407,0.0006025703,0.005140649,0.006565049,0.001396538,0.002831526,0.002610375,0.07110447,0.02062962,0.1103043,0.5240912,0.2515863],"study_design_scores_gemma":[0.001513554,0.0003950361,0.003461482,0.00162448,0.0004042124,0.001758733,0.001057605,0.2971455,0.02243902,0.2338715,0.4352413,0.001087448],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001813847,0.0009748029,0.5910475,0.0008975827,0.0001920145,0.000653599,0.03889904,0.362146,0.00337556],"genre_scores_gemma":[0.03687442,0.001744707,0.6829357,0.001577258,0.0001888734,0.002619819,0.2218936,0.04829109,0.003874532],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0201402,"threshold_uncertainty_score":0.1065128,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001246661,0.0008896518,0.0008472149,0.006222958,0.0005914202,0.001105403,0.0008718146,0.0006089984,0.008372777],"category_scores_gemma":[0.003549899,0.0002593179,0.0009425549,0.002996022,0.0003067155,0.001242818,0.0008858124,0.0005596699,0.003463812],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004392291,"about_ca_system_score_gemma":0.0004960991,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008646972,"about_ca_topic_score_gemma":0.000907174,"domain_scores_codex":[0.9988798,0.0002917152,0.0001599802,0.000267231,0.0002921843,0.0001090061],"domain_scores_gemma":[0.9979659,0.0007853809,0.000286793,0.0003087825,0.0005451075,0.0001079882],"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":[0.002649112,0.0005647654,0.04861657,0.002986327,0.0003916032,0.0008653697,0.0004952931,0.01102564,0.5711436,0.005918914,0.01026814,0.3450747],"study_design_scores_gemma":[0.0002589661,0.001206101,0.1004906,0.0004738953,0.0004678122,0.004343539,0.000868884,0.3572037,0.4295518,0.01804806,0.08683831,0.0002483786],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.5840777,0.002904861,0.3640577,0.0003672148,0.0001697321,0.0005996197,0.0186364,0.01594919,0.01323757],"genre_scores_gemma":[0.7405499,0.000587378,0.2291354,0.00009899146,0.00004664459,0.0002428797,0.02606397,0.0006376223,0.002637304],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.008372777,"threshold_uncertainty_score":0.02800977,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001443316,0.00127531,0.00123531,0.003448678,0.001100235,0.001376656,0.0008331025,0.0009592427,0.001883478],"category_scores_gemma":[0.003880121,0.0004673222,0.001303469,0.004001426,0.0004142768,0.001358952,0.001750923,0.001157464,0.001675735],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005071544,"about_ca_system_score_gemma":0.001603105,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003383288,"about_ca_topic_score_gemma":0.006539413,"domain_scores_codex":[0.9984081,0.0002702296,0.0001291755,0.0005803916,0.000480607,0.0001315031],"domain_scores_gemma":[0.997982,0.0009169186,0.0002211355,0.0003522648,0.0003892881,0.0001383254],"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":[0.001979535,0.0008291077,0.2040492,0.01029694,0.001382128,0.00275863,0.001652083,0.03949452,0.302993,0.006267819,0.05232405,0.375973],"study_design_scores_gemma":[0.0002183788,0.0005646966,0.2628385,0.001820657,0.001050819,0.002892167,0.003280919,0.1440329,0.1044468,0.03148255,0.4470634,0.0003082017],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6041881,0.01417868,0.1084112,0.001449378,0.0003197056,0.0002650296,0.2546517,0.005008765,0.01152767],"genre_scores_gemma":[0.3670221,0.00550806,0.1543221,0.0008569309,0.000105759,0.0005464797,0.4681684,0.0006958061,0.002774435],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003448678,"threshold_uncertainty_score":0.00763303,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005286013,0.0001778891,0.000287369,0.00105876,0.000192962,0.0007932028,0.0002004206,0.0003298667,0.001444164],"category_scores_gemma":[0.001994645,0.0001097005,0.000173108,0.0006767861,0.0004431747,0.0007058091,0.0002752545,0.000308903,0.0005036691],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001598121,"about_ca_system_score_gemma":0.0001320585,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005298841,"about_ca_topic_score_gemma":0.0006583281,"domain_scores_codex":[0.9996474,0.00007981535,0.0000220404,0.0001459878,0.00006846189,0.00003620316],"domain_scores_gemma":[0.9981943,0.0009007055,0.0004329706,0.0001301051,0.000198462,0.0001434105],"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":[0.0007884388,0.00006990385,0.4523744,0.0005837358,0.0003518547,0.0002484351,0.0006275222,0.004955314,0.4929264,0.001706064,0.0006015711,0.04476646],"study_design_scores_gemma":[0.00001910049,0.0005248535,0.8508533,0.0001342297,0.0002712482,0.001225235,0.001001149,0.03459929,0.099641,0.006788059,0.004867412,0.00007523597],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9911582,0.00122839,0.005793555,0.00009524803,0.000008074476,0.000005575931,0.0004512819,0.0001010907,0.001158615],"genre_scores_gemma":[0.997896,0.0001120903,0.001451445,0.00002386714,0.000007771757,0.00000271251,0.0003447332,0.00001627141,0.0001452149],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001444164,"threshold_uncertainty_score":0.004831195,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003191692,0.0001109578,0.0001768497,0.0006037793,0.0001754974,0.0002692755,0.0002201844,0.0002323071,0.001677236],"category_scores_gemma":[0.0005658905,0.00008954859,0.0001399573,0.0003397337,0.0004462159,0.0002114631,0.0002120818,0.0001612203,0.0001339106],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001199897,"about_ca_system_score_gemma":0.0001176685,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003920967,"about_ca_topic_score_gemma":0.0006872559,"domain_scores_codex":[0.9998304,0.00003338182,0.00001065421,0.00006450435,0.00003786107,0.00002315041],"domain_scores_gemma":[0.9997447,0.00007853688,0.00007008341,0.00003818536,0.00002211118,0.00004629096],"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":[0.0005397208,0.00008734634,0.4073966,0.0000937009,0.0001360766,0.001674968,0.0009998819,0.0004230891,0.5610805,0.005548319,0.0002480405,0.02177191],"study_design_scores_gemma":[0.00004593195,0.0004535937,0.9406605,0.00003082223,0.0001350125,0.007433136,0.0006014878,0.003942043,0.03726378,0.006369584,0.003036344,0.00002784328],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9934384,0.0006739132,0.004721629,0.0001114835,0.000009199379,0.00001338689,0.0001014409,0.00003451483,0.0008960457],"genre_scores_gemma":[0.9990435,0.0001038238,0.0004550344,0.0000374155,0.000008725865,0.000006444894,0.00006595438,0.000004471926,0.0002744579],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001677236,"threshold_uncertainty_score":0.005610943,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00158003,0.0009545026,0.0008986493,0.001091405,0.000305007,0.0008017107,0.0009970336,0.001077258,0.002009379],"category_scores_gemma":[0.003022502,0.0005489266,0.0008586466,0.0006125365,0.0004102554,0.0005999831,0.001008137,0.001506229,0.001546795],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003457964,"about_ca_system_score_gemma":0.0006871733,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001328054,"about_ca_topic_score_gemma":0.002512511,"domain_scores_codex":[0.9986233,0.0003069479,0.00006940857,0.0004990466,0.0003762689,0.0001250123],"domain_scores_gemma":[0.9989012,0.0005037397,0.0001485458,0.0001477479,0.0002208121,0.00007795215],"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":[0.0003314902,0.0001080229,0.006996641,0.0002917473,0.0001649024,0.0003812417,0.0001345434,0.00769216,0.9072444,0.0015223,0.001619058,0.07351345],"study_design_scores_gemma":[0.00007635337,0.0005980338,0.01469489,0.00004236133,0.000265737,0.002031388,0.00005895699,0.2588547,0.7067597,0.003896533,0.01255835,0.0001631017],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2045575,0.001347914,0.7837214,0.0006107558,0.000168664,0.0002434645,0.002793112,0.004435686,0.00212152],"genre_scores_gemma":[0.4543739,0.0005883951,0.5369962,0.0008603106,0.0001038587,0.0003021444,0.003280989,0.0006170284,0.002877174],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.002009379,"threshold_uncertainty_score":0.008356094,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003157583,0.0005509676,0.0004737714,0.0005869477,0.0002925076,0.0003509955,0.0003812806,0.000535442,0.001274769],"category_scores_gemma":[0.001222795,0.0002253787,0.0003993469,0.0004444238,0.000131139,0.000282133,0.0002195662,0.0002865026,0.0002800398],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003538927,"about_ca_system_score_gemma":0.000678506,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005874961,"about_ca_topic_score_gemma":0.004464462,"domain_scores_codex":[0.9999175,0.00002963848,0.000003819992,0.00001960346,0.00001779633,0.00001161282],"domain_scores_gemma":[0.9996964,0.0002051523,0.00001755903,0.00001804592,0.00003489148,0.00002802101],"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":[0.0006383882,0.0003269874,0.009472709,0.0001579417,0.00007514666,0.0002067195,0.00007778403,0.8937038,0.03985631,0.002235704,0.0005871037,0.05266137],"study_design_scores_gemma":[0.00002355746,0.00004166651,0.0007976149,0.000003056907,0.000006595021,0.00002332568,0.00001711614,0.9951461,0.003304917,0.0004501256,0.0001816827,0.000004258773],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8986335,0.0002028467,0.09774811,0.00008593033,0.000007962752,0.00004869624,0.0007424041,0.001350903,0.001179623],"genre_scores_gemma":[0.9325437,0.0001614345,0.0653353,0.00001684175,0.000005110499,0.00005463362,0.001406331,0.00008660388,0.000390094],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005874961,"threshold_uncertainty_score":0.01168156,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002580454,0.0007920898,0.00078455,0.001643839,0.0004406839,0.0008617415,0.0007935132,0.0007844316,0.001696398],"category_scores_gemma":[0.00709076,0.0003657408,0.000803956,0.0009995169,0.0003731896,0.0004891838,0.001593892,0.0005951075,0.000452536],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000276111,"about_ca_system_score_gemma":0.0005288555,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001579107,"about_ca_topic_score_gemma":0.004369201,"domain_scores_codex":[0.9982945,0.0004015613,0.0001752698,0.0006299206,0.00039619,0.0001025035],"domain_scores_gemma":[0.9977379,0.001362343,0.0002869477,0.0003189912,0.0002092291,0.00008459812],"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":[0.002405964,0.0002998683,0.1763435,0.000945112,0.001624573,0.003086355,0.0006046372,0.06024189,0.3587764,0.003734919,0.01122236,0.3807146],"study_design_scores_gemma":[0.0003329696,0.0005453448,0.1713342,0.0001583286,0.0004329482,0.005390919,0.000366888,0.4801867,0.3080187,0.01005963,0.02298473,0.0001887268],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.630362,0.001484146,0.3542119,0.0003259126,0.00009000492,0.0002180267,0.005918686,0.00520983,0.002179476],"genre_scores_gemma":[0.69973,0.0004814703,0.2841868,0.000355285,0.0000373306,0.000196031,0.01268722,0.0005082811,0.001817404],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.002580454,"threshold_uncertainty_score":0.0136469,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001034651,0.001640629,0.00130338,0.003652806,0.001037596,0.001575036,0.001453969,0.00104761,0.007102691],"category_scores_gemma":[0.003268161,0.001058017,0.0009026152,0.004202473,0.0003903398,0.001698202,0.002289011,0.001593981,0.01502905],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007684511,"about_ca_system_score_gemma":0.001976519,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002467646,"about_ca_topic_score_gemma":0.003803016,"domain_scores_codex":[0.999012,0.0001354954,0.00008553643,0.0002777008,0.0003457189,0.0001435329],"domain_scores_gemma":[0.9982078,0.0004033406,0.000305601,0.0003735339,0.0003566587,0.0003531186],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.004623462,0.0004972205,0.04423932,0.005797208,0.0002258406,0.001248886,0.0007694194,0.01438696,0.09313759,0.01072445,0.581399,0.2429507],"study_design_scores_gemma":[0.0004924415,0.0006221021,0.03364914,0.0007501809,0.0001755414,0.001450753,0.0003795967,0.0478516,0.0678697,0.01517339,0.8312212,0.0003643086],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.04762596,0.004087121,0.07473957,0.0005922453,0.0002590501,0.0003693311,0.6672465,0.1975396,0.007540611],"genre_scores_gemma":[0.06198782,0.001558858,0.08714309,0.0002173768,0.00008074037,0.0006800535,0.8375592,0.008012318,0.00276061],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.007102691,"threshold_uncertainty_score":0.02376091,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006159349,0.0009147465,0.0005832815,0.0003976501,0.0004177046,0.0007308157,0.0007367883,0.0007075894,0.001610038],"category_scores_gemma":[0.001489408,0.0003383315,0.0005712594,0.000302998,0.0003785922,0.0005364026,0.0005392632,0.000755172,0.0002799117],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004254855,"about_ca_system_score_gemma":0.001405878,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001679496,"about_ca_topic_score_gemma":0.002889832,"domain_scores_codex":[0.9997995,0.0000570915,0.00001267089,0.00005398216,0.00005149786,0.00002522218],"domain_scores_gemma":[0.9995926,0.0002527029,0.00004096865,0.00002813377,0.00005682338,0.00002880155],"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":[0.0004478715,0.0005431324,0.007182652,0.0003140365,0.0001975811,0.0001900658,0.00009733297,0.8812737,0.03172577,0.00604818,0.001603993,0.0703756],"study_design_scores_gemma":[0.00003399854,0.0001412485,0.0001819925,0.000003902629,0.00002082191,0.00001879819,0.00001601375,0.9938793,0.003730593,0.001266961,0.0007020761,0.000004276307],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3788892,0.0003947451,0.6098813,0.0005823405,0.0001140188,0.0002702425,0.0005042924,0.002312283,0.007051594],"genre_scores_gemma":[0.5835348,0.0002568231,0.4124604,0.0003729218,0.00002798004,0.0003518391,0.0009582068,0.000157874,0.001879201],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001679496,"threshold_uncertainty_score":0.005386114,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004070596,0.0006381907,0.0007041519,0.0005493616,0.0003274597,0.0005404956,0.0008469913,0.001110813,0.001141123],"category_scores_gemma":[0.001042146,0.0005451859,0.0006962002,0.0004157646,0.0005428073,0.000601313,0.0003832377,0.0008643693,0.0004348709],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008279852,"about_ca_system_score_gemma":0.0007441726,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006933968,"about_ca_topic_score_gemma":0.007966621,"domain_scores_codex":[0.9998424,0.00003906182,0.000008029128,0.00006198536,0.00002879875,0.00001973551],"domain_scores_gemma":[0.9995468,0.0002988247,0.00005716178,0.00003190042,0.00004018418,0.00002509688],"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":[0.00003573526,0.00007702697,0.001181126,0.00002045807,0.00002614878,0.0000264403,0.00001335623,0.9837357,0.002869956,0.0008414817,0.0002163928,0.01095608],"study_design_scores_gemma":[0.00000187162,0.000005993452,0.00008336507,4.836805e-7,0.000001660027,0.000003098634,0.000001226191,0.9987804,0.0003738799,0.0007014558,0.00004553766,9.503117e-7],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4298896,0.0005310968,0.5648219,0.0003408784,0.00002413361,0.00007435227,0.0005536877,0.001986588,0.001777759],"genre_scores_gemma":[0.8967908,0.000216798,0.09958357,0.0001010135,0.00002612095,0.0001477064,0.001053741,0.0001018658,0.001978506],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006933968,"threshold_uncertainty_score":0.01378721,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001360654,0.0004461248,0.0006639729,0.001193546,0.0003604634,0.0007909901,0.0004559118,0.00039544,0.0004813867],"category_scores_gemma":[0.003446684,0.000187239,0.0009136298,0.00122353,0.0003888388,0.0004239365,0.0004130363,0.000607644,0.0002258208],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005869799,"about_ca_system_score_gemma":0.0008449563,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001752197,"about_ca_topic_score_gemma":0.003319958,"domain_scores_codex":[0.9993247,0.0001753608,0.00005258713,0.0002304921,0.000165457,0.00005142728],"domain_scores_gemma":[0.997285,0.001742824,0.0002789131,0.0001957095,0.0003963032,0.0001013421],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001982074,0.000584468,0.3027177,0.001308934,0.0009088418,0.0007393935,0.0002341342,0.3011838,0.2912033,0.004932446,0.004007697,0.09019719],"study_design_scores_gemma":[0.00003603898,0.0002569325,0.06234463,0.00003360041,0.0001262797,0.0002134442,0.00009918013,0.8969175,0.0315113,0.005984891,0.0024277,0.00004863673],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9022971,0.0006914685,0.08235062,0.0002837716,0.00003173317,0.00008953896,0.01250065,0.001069697,0.0006853681],"genre_scores_gemma":[0.9473526,0.0001695419,0.0333717,0.000110625,0.0000205765,0.0001196356,0.01853906,0.00009473849,0.0002215431],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001752197,"threshold_uncertainty_score":0.00719589,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002318502,0.0002450219,0.0002047385,0.0003222232,0.00008629187,0.0002619053,0.000159043,0.0001829373,0.001269695],"category_scores_gemma":[0.0002301995,0.0001380369,0.0001972296,0.0001497544,0.0002402011,0.0001303352,0.0003143507,0.0003488805,0.000280612],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001313022,"about_ca_system_score_gemma":0.0001785801,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002492905,"about_ca_topic_score_gemma":0.0003581383,"domain_scores_codex":[0.9998423,0.00003045922,0.00001178664,0.00003414077,0.00005341489,0.00002788184],"domain_scores_gemma":[0.9998487,0.00005131556,0.00005275982,0.00001156689,0.00001520793,0.00002032156],"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":[0.00004732174,0.000008543373,0.001129337,0.00001675932,0.00000357969,0.00003873372,0.000007825094,0.0001007047,0.9976826,0.00006680356,0.000005159396,0.0008926356],"study_design_scores_gemma":[0.00001413956,0.0003229494,0.02703959,0.00001157765,0.0000315615,0.0007691821,0.00006043748,0.001749851,0.9682673,0.0001789851,0.001548798,0.00000572106],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9934586,0.0004661677,0.005393956,0.00002935708,0.000003155943,0.00001263609,0.0001803112,0.00004261726,0.0004132212],"genre_scores_gemma":[0.9954875,0.0002548498,0.003132595,0.00001610647,0.000002181241,0.00001138181,0.0004184189,0.00002309452,0.0006538238],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001269695,"threshold_uncertainty_score":0.004247546,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009350859,0.002676544,0.001098151,0.001505176,0.0005833508,0.0009669993,0.00198389,0.001446804,0.004486693],"category_scores_gemma":[0.001955223,0.0009744935,0.001693942,0.000629081,0.0002470517,0.001566406,0.001185842,0.001853747,0.002879595],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001140882,"about_ca_system_score_gemma":0.001223454,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006730518,"about_ca_topic_score_gemma":0.01045761,"domain_scores_codex":[0.9995779,0.00003027136,0.00002339087,0.0001872288,0.0001150175,0.00006626606],"domain_scores_gemma":[0.9996792,0.0001162443,0.00004835326,0.00004754612,0.00008189498,0.00002679578],"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":[0.001368625,0.0003702135,0.01276135,0.001613365,0.0006624933,0.0005200371,0.0001920091,0.1495602,0.1952424,0.004507742,0.04860563,0.584596],"study_design_scores_gemma":[0.00007511042,0.0001891441,0.001834249,0.00008080929,0.0001128826,0.0001918513,0.00003633364,0.9196302,0.05984448,0.003521418,0.01441249,0.00007099495],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.08074184,0.003278532,0.7610054,0.0003616813,0.0004275969,0.000389073,0.01751286,0.1293862,0.006896824],"genre_scores_gemma":[0.2063826,0.001443092,0.7214662,0.0007880926,0.0001217208,0.0007844884,0.05409346,0.004536648,0.01038356],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006730518,"threshold_uncertainty_score":0.01500946,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004062114,0.001025594,0.001191259,0.0008272302,0.0003522183,0.0006517311,0.0009240853,0.0006406787,0.0006274377],"category_scores_gemma":[0.005392809,0.0002679967,0.001062375,0.0006812396,0.0003271324,0.0009506326,0.0006806731,0.0007893366,0.0002283473],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004944149,"about_ca_system_score_gemma":0.0007354257,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003526653,"about_ca_topic_score_gemma":0.005161383,"domain_scores_codex":[0.9990005,0.000538011,0.00004844607,0.0001682598,0.0001473705,0.00009750937],"domain_scores_gemma":[0.998007,0.001336072,0.00009683416,0.0001849861,0.0003058517,0.00006927722],"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":[0.0003373849,0.0001892885,0.008094857,0.00005578319,0.000434203,0.0001198759,0.0000706888,0.8202017,0.009328529,0.001429647,0.001068249,0.1586698],"study_design_scores_gemma":[0.000008172438,0.00006194203,0.0007720752,0.000003436945,0.00002348793,0.00001707327,0.00001051622,0.9961123,0.001461361,0.001363194,0.000160249,0.00000627933],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1174683,0.0005664103,0.8804254,0.0002291328,0.00003173326,0.00006862496,0.0001329049,0.000616456,0.0004610336],"genre_scores_gemma":[0.8121673,0.0002112556,0.1858978,0.0001723008,0.00005034441,0.0001664508,0.0005851173,0.00006226318,0.0006871865],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004062114,"threshold_uncertainty_score":0.02148277,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002445725,0.001718546,0.001726275,0.006632308,0.0009244793,0.001426273,0.001991658,0.0006563969,0.02450541],"category_scores_gemma":[0.005026546,0.0008597596,0.002052912,0.006118332,0.000457544,0.0009176845,0.00253039,0.001749454,0.00616386],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008527592,"about_ca_system_score_gemma":0.00201353,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00187549,"about_ca_topic_score_gemma":0.005017678,"domain_scores_codex":[0.9981661,0.0004754858,0.0001624952,0.0004839082,0.000561549,0.0001504846],"domain_scores_gemma":[0.9977285,0.00131202,0.0002613613,0.0003161136,0.000276185,0.0001058866],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.002376393,0.0002754666,0.03512955,0.01112834,0.00271308,0.001385449,0.001405382,0.01762634,0.1927567,0.0243115,0.3809205,0.3299712],"study_design_scores_gemma":[0.0006788248,0.0004500961,0.0655404,0.0007946498,0.001498086,0.002454117,0.0003667416,0.06834607,0.1076967,0.03692508,0.7148952,0.0003540599],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.03171165,0.007058828,0.5385583,0.0006443396,0.0006298068,0.0006337366,0.3127338,0.09017926,0.01785028],"genre_scores_gemma":[0.08794935,0.003408163,0.6937876,0.0006758939,0.0001689903,0.004612388,0.1907772,0.01142118,0.007199047],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.02450541,"threshold_uncertainty_score":0.08197874,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004153215,0.0003197647,0.0004716747,0.0008876511,0.0003897528,0.0004584433,0.0004085458,0.0004325234,0.001417362],"category_scores_gemma":[0.0003636293,0.0003094005,0.0006585308,0.0006612842,0.0001853139,0.000295754,0.000578589,0.0006155038,0.000695519],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000294521,"about_ca_system_score_gemma":0.0003873458,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002721513,"about_ca_topic_score_gemma":0.005453659,"domain_scores_codex":[0.9998075,0.00002379144,0.00001525099,0.00007890172,0.00005658331,0.00001797489],"domain_scores_gemma":[0.9997401,0.00006103646,0.00005362205,0.00004767862,0.000067594,0.00003006121],"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":[0.00006173454,0.00002199993,0.001762615,0.000161206,0.00004318796,0.0001138967,0.0001218673,0.0008941716,0.9807621,0.0004241514,0.0001691779,0.01546404],"study_design_scores_gemma":[0.00007438927,0.0005283017,0.09595405,0.00009527153,0.0004312831,0.001094684,0.0004179357,0.03412828,0.8256388,0.001561151,0.03996852,0.000107238],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6044644,0.001342832,0.384261,0.0001821674,0.0000910925,0.0003202572,0.005253435,0.001737094,0.002347776],"genre_scores_gemma":[0.502461,0.0009310207,0.4791811,0.000149726,0.00002616601,0.0002497701,0.01233748,0.0003807134,0.004283091],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.002721513,"threshold_uncertainty_score":0.005411267,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00589844,0.0008697066,0.0009471595,0.0006177568,0.0008305249,0.001648262,0.001730008,0.00141652,0.00384279],"category_scores_gemma":[0.01674736,0.0006114781,0.001343529,0.0007502807,0.001314465,0.001222045,0.00156634,0.002122942,0.0006582406],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001402272,"about_ca_system_score_gemma":0.000995912,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007877944,"about_ca_topic_score_gemma":0.00569162,"domain_scores_codex":[0.998633,0.0007783763,0.00005389105,0.0001928101,0.000255888,0.00008611724],"domain_scores_gemma":[0.9877434,0.01053922,0.0003511666,0.0005690191,0.0005676082,0.0002296767],"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":[0.0001598109,0.00005262575,0.004711994,0.00009710172,0.00006563617,0.0001259995,0.0001690476,0.9688178,0.002663513,0.01662066,0.001071782,0.005444075],"study_design_scores_gemma":[0.000006348057,0.000005004258,0.0001446038,0.000002706735,0.000002227523,0.000006070648,0.000008570298,0.9960632,0.0005084936,0.002969044,0.0002782111,0.000005482308],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1913232,0.0002579017,0.7974876,0.0007809584,0.00016741,0.0001223283,0.001131707,0.00550612,0.003222809],"genre_scores_gemma":[0.6949153,0.000169862,0.2979,0.0004899587,0.00006408303,0.000566709,0.001722201,0.001558492,0.00261336],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007877944,"threshold_uncertainty_score":0.03119433,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001607604,0.001005066,0.0008080609,0.0003683248,0.0004929044,0.001205516,0.001029467,0.0006080742,0.002701232],"category_scores_gemma":[0.001698573,0.0006215457,0.001149581,0.0003266207,0.0004155518,0.00060028,0.0007852872,0.001002181,0.0009102529],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001156652,"about_ca_system_score_gemma":0.001855736,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002307316,"about_ca_topic_score_gemma":0.003060078,"domain_scores_codex":[0.9992437,0.0001596692,0.00005105423,0.0002794961,0.0002093697,0.00005655162],"domain_scores_gemma":[0.999084,0.0003418296,0.00009218309,0.0002338579,0.00017972,0.00006835913],"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":[0.0005984387,0.0002359634,0.004320972,0.0005813622,0.0001200743,0.0001744266,0.00008458596,0.1198178,0.8443503,0.004713759,0.001823902,0.02317844],"study_design_scores_gemma":[0.0001081067,0.0003265742,0.003214466,0.00003142145,0.0000793278,0.00007850004,0.00005175662,0.5578418,0.4216676,0.004570376,0.01193685,0.00009313056],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1841459,0.0002748046,0.79174,0.0003453519,0.0001273804,0.0005308233,0.004511142,0.01457702,0.00374756],"genre_scores_gemma":[0.4952387,0.0004238756,0.4927424,0.0001925936,0.00002556037,0.001791695,0.006607594,0.001465655,0.001512034],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002701232,"threshold_uncertainty_score":0.009036481,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01171695,0.001036124,0.0007545606,0.0008890578,0.000750761,0.002220436,0.0007859928,0.0007376763,0.001258586],"category_scores_gemma":[0.02349586,0.0004297401,0.00121145,0.001237763,0.0008773283,0.001275278,0.001009884,0.001365661,0.000709714],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001323772,"about_ca_system_score_gemma":0.001852022,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001054902,"about_ca_topic_score_gemma":0.001494941,"domain_scores_codex":[0.9920194,0.002851113,0.0004623964,0.002371125,0.001899477,0.0003965622],"domain_scores_gemma":[0.9817204,0.01157733,0.001395615,0.002882905,0.002067507,0.0003562596],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001677767,0.0005172053,0.06742608,0.0009022974,0.000568228,0.0001700019,0.0005962502,0.0286587,0.7744768,0.01185124,0.001510821,0.1116447],"study_design_scores_gemma":[0.00007201637,0.000945391,0.07495191,0.000104585,0.0003949903,0.0002520207,0.0003190836,0.2796929,0.6156663,0.01789899,0.009509399,0.0001924988],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4414167,0.0005571554,0.550408,0.0004519842,0.0001484134,0.000481176,0.001192241,0.002321441,0.003022887],"genre_scores_gemma":[0.7379821,0.0002869794,0.2571118,0.0003341335,0.00003986912,0.0007720628,0.001449954,0.0009880306,0.0010351],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.988283,"threshold_uncertainty_score":0.06196588,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008577793,0.0007008698,0.0008168521,0.0006376441,0.0003559695,0.0008801303,0.0009282664,0.0009550565,0.0008272991],"category_scores_gemma":[0.003336823,0.00058524,0.0008040227,0.0004195199,0.0009834835,0.001069791,0.001062231,0.002097233,0.0002583284],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006251146,"about_ca_system_score_gemma":0.0008112793,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003684106,"about_ca_topic_score_gemma":0.004901476,"domain_scores_codex":[0.9997053,0.0001128488,0.00001352275,0.00007531653,0.00006683802,0.00002625044],"domain_scores_gemma":[0.9982018,0.001334861,0.0001714727,0.0001391898,0.00009816619,0.00005444536],"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":[0.00001550019,0.00002489485,0.0005663115,0.00003929873,0.00003095331,0.00005281297,0.00003017371,0.9701611,0.001964529,0.01212758,0.0004169255,0.01456993],"study_design_scores_gemma":[0.000001080536,0.000002492837,0.00003235091,8.441054e-7,8.630472e-7,0.000003555227,0.000001337318,0.9949818,0.0001465765,0.004768323,0.00005939998,0.000001388223],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01994736,0.0001243213,0.9788293,0.0002467793,0.00000933255,0.0000200693,0.00009696338,0.0003046046,0.0004212777],"genre_scores_gemma":[0.5687734,0.000509691,0.4271495,0.000250923,0.00009164801,0.0002283821,0.0008219013,0.0001443544,0.002030259],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003684106,"threshold_uncertainty_score":0.007325292,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009087081,0.002078097,0.001334688,0.002998,0.001541628,0.003408759,0.003365298,0.002293016,0.006778936],"category_scores_gemma":[0.01841784,0.001608005,0.003411991,0.002071275,0.002688803,0.003542113,0.004275696,0.004191592,0.002122998],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001739736,"about_ca_system_score_gemma":0.00302451,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008254163,"about_ca_topic_score_gemma":0.006766155,"domain_scores_codex":[0.9962357,0.001845999,0.0001884895,0.0006315481,0.0009478548,0.000150476],"domain_scores_gemma":[0.9914142,0.005719369,0.0007053089,0.0009208221,0.0009591536,0.0002810505],"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":[0.0001479498,0.00006847778,0.001825824,0.0002524196,0.0002217628,0.000328563,0.0002737898,0.4899486,0.002923511,0.3871568,0.007855558,0.1089966],"study_design_scores_gemma":[0.00001666128,0.00003587081,0.000175932,0.00002841482,0.00002590299,0.00008845505,0.00001871536,0.8288453,0.0005996106,0.1607141,0.009414793,0.0000362672],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0003403694,0.00009164295,0.9985133,0.0001242222,0.00003059055,0.00001428918,0.0000896753,0.0004496606,0.0003463271],"genre_scores_gemma":[0.0571858,0.0007241594,0.9354717,0.0003483394,0.0004829906,0.000334531,0.0008093662,0.0008584799,0.003784626],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009087081,"threshold_uncertainty_score":0.04805768,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002291596,0.0007453915,0.0005699839,0.002427076,0.000774901,0.001743071,0.0004582548,0.0006636629,0.001302366],"category_scores_gemma":[0.005055727,0.0003462562,0.0009277893,0.001870796,0.0003367391,0.0009033004,0.001089256,0.000974689,0.0007709794],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005305892,"about_ca_system_score_gemma":0.001154307,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003290608,"about_ca_topic_score_gemma":0.01058241,"domain_scores_codex":[0.9988398,0.0002428932,0.0001122927,0.0004016035,0.00031068,0.00009277963],"domain_scores_gemma":[0.9980338,0.0007008045,0.0003309801,0.0003020545,0.0005191008,0.0001132442],"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":[0.0008124709,0.0003278853,0.1184027,0.0006755296,0.0004259965,0.0009872953,0.002091912,0.04112023,0.607427,0.004728923,0.004109609,0.2188905],"study_design_scores_gemma":[0.00008999068,0.0004529089,0.1620858,0.0002182514,0.0003364139,0.001476355,0.001676913,0.4323013,0.3369024,0.02447711,0.03970011,0.0002823869],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5615893,0.001055611,0.4216435,0.0005618194,0.0001201312,0.0003127735,0.005927999,0.005167118,0.003621707],"genre_scores_gemma":[0.5534862,0.0004517256,0.4364987,0.000184158,0.00004286298,0.0001567093,0.007021572,0.0008231733,0.001334946],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003290608,"threshold_uncertainty_score":0.01211923,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005949493,0.002691091,0.001737502,0.003300498,0.001371124,0.003576785,0.003115993,0.001966795,0.0221151],"category_scores_gemma":[0.01128068,0.002306021,0.002313394,0.002623398,0.001097578,0.002326436,0.004942333,0.004032355,0.02062496],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001037132,"about_ca_system_score_gemma":0.00302635,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002358767,"about_ca_topic_score_gemma":0.003013035,"domain_scores_codex":[0.9972645,0.0004527934,0.0002221153,0.001154244,0.0006894731,0.0002168562],"domain_scores_gemma":[0.9952623,0.001941254,0.0006315327,0.0009701988,0.0006550088,0.00053966],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.006194124,0.0003324404,0.02501819,0.004232705,0.001591651,0.002095535,0.002102511,0.01247895,0.1346776,0.01119629,0.4956569,0.3044231],"study_design_scores_gemma":[0.001729458,0.0009344955,0.03540522,0.0009780632,0.0006010621,0.00345266,0.0006093277,0.1276073,0.1884452,0.05305776,0.5862294,0.0009500947],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01358339,0.001201311,0.4916122,0.0008569147,0.0006433602,0.0008150819,0.07874364,0.4048232,0.007720886],"genre_scores_gemma":[0.05228916,0.0006845821,0.7283022,0.001534015,0.000250436,0.001524589,0.1663368,0.043561,0.005517303],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.0221151,"threshold_uncertainty_score":0.07398242,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004517465,0.0005966861,0.0006047912,0.00095492,0.0003083798,0.0005057045,0.0006247906,0.0004352118,0.001727779],"category_scores_gemma":[0.001054433,0.0002139549,0.000674445,0.0008036979,0.0003257543,0.0004449342,0.0005734278,0.0005476935,0.0006574578],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002899327,"about_ca_system_score_gemma":0.0005612334,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001706994,"about_ca_topic_score_gemma":0.002043202,"domain_scores_codex":[0.9997944,0.00005246519,0.00001342272,0.00006836848,0.00005446082,0.00001676525],"domain_scores_gemma":[0.9996687,0.0001445847,0.00004828068,0.00006392993,0.0000395817,0.00003497293],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0006452698,0.0004021443,0.006322062,0.0003242025,0.0001899958,0.0004801086,0.0001791902,0.4603365,0.2088834,0.009814251,0.001009111,0.3114138],"study_design_scores_gemma":[0.00001274451,0.00009883912,0.0007026791,0.000006241721,0.00001480745,0.0001255741,0.00001169676,0.9850359,0.01118869,0.002121318,0.0006705833,0.00001104052],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2238335,0.0002493479,0.770476,0.00004860021,0.00001568275,0.0001005359,0.0002804192,0.002778539,0.002217362],"genre_scores_gemma":[0.4297567,0.0001815182,0.5672613,0.00005930871,0.00001194017,0.00009875574,0.000915082,0.0003365766,0.001378746],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.001727779,"threshold_uncertainty_score":0.005780041,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001633987,0.0008999299,0.001157038,0.001994302,0.0001581142,0.0009906522,0.0008745635,0.001102172,0.001495489],"category_scores_gemma":[0.001996812,0.0003203441,0.0007375171,0.001267943,0.0004598571,0.0007606488,0.0006164907,0.001232471,0.001201541],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006564886,"about_ca_system_score_gemma":0.0009064366,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001228769,"about_ca_topic_score_gemma":0.001955178,"domain_scores_codex":[0.9996537,0.00008001904,0.00003115456,0.00008581197,0.0001265097,0.00002279488],"domain_scores_gemma":[0.9992647,0.0004018714,0.00006504037,0.0000255301,0.0002094883,0.00003338656],"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":[0.00007819411,0.00002743081,0.0004772911,0.007385355,0.0001742895,0.0001396752,0.00004180682,0.001369347,0.004616138,0.002803397,0.008149654,0.9747373],"study_design_scores_gemma":[0.00005606124,0.0004099418,0.005368592,0.005309742,0.000704196,0.004268414,0.0001247231,0.004497235,0.02160604,0.008839966,0.9486786,0.0001365485],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0008690496,0.9911826,0.005620904,0.0003813032,0.0002041302,0.00001576541,0.00007411997,0.00004407474,0.001608121],"genre_scores_gemma":[0.008536041,0.9831655,0.005721737,0.0005177065,0.0002060032,0.00003537272,0.0002034353,0.00003098222,0.001583155],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.001994302,"threshold_uncertainty_score":0.008641422,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004982805,0.0006908778,0.0005179509,0.0008042292,0.000339174,0.0005418907,0.0002859298,0.0006894121,0.003590745],"category_scores_gemma":[0.001442394,0.000225928,0.0009075967,0.0006196251,0.0002784384,0.0004355085,0.0003584013,0.0004353682,0.00182117],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002369931,"about_ca_system_score_gemma":0.0002916941,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001721625,"about_ca_topic_score_gemma":0.002812275,"domain_scores_codex":[0.9996343,0.00009105226,0.00001726879,0.0001548593,0.00005417467,0.00004826737],"domain_scores_gemma":[0.9992365,0.0005084538,0.000057568,0.0001048418,0.00005625042,0.00003654369],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.002392908,0.0002740393,0.3730685,0.0008240253,0.0008913514,0.002315869,0.0004182681,0.1899446,0.0981767,0.00564182,0.008859016,0.3171931],"study_design_scores_gemma":[0.00006844312,0.0003380558,0.2055072,0.0001370888,0.0004856175,0.00388499,0.0002366909,0.7234491,0.03221598,0.02026748,0.0133434,0.0000659798],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9087055,0.002819222,0.07639153,0.0003151555,0.0000503105,0.00004626872,0.005516494,0.0009914769,0.005164179],"genre_scores_gemma":[0.9899412,0.0003650462,0.004236158,0.00007107676,0.0000205086,0.00001942801,0.004053901,0.00008543221,0.001207234],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003590745,"threshold_uncertainty_score":0.01201218,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003261157,0.001319699,0.0008916899,0.001305576,0.001103271,0.001735315,0.001207134,0.0007221901,0.01125014],"category_scores_gemma":[0.003674778,0.001247556,0.001628083,0.001157181,0.0004794053,0.0009885634,0.001854052,0.002418488,0.0136561],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005369656,"about_ca_system_score_gemma":0.001808334,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002926424,"about_ca_topic_score_gemma":0.007002568,"domain_scores_codex":[0.9980901,0.0003369141,0.0001624265,0.0006810917,0.0005820988,0.0001473807],"domain_scores_gemma":[0.9985943,0.0003637155,0.0001711727,0.0002872044,0.0004355482,0.0001480161],"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":[0.002337286,0.0002106112,0.009707727,0.001834577,0.0005387982,0.0006829679,0.001681669,0.007421254,0.4579982,0.01091376,0.2963023,0.2103708],"study_design_scores_gemma":[0.0004013343,0.0006863163,0.01277594,0.0003012369,0.0002028662,0.001018612,0.0002518706,0.07052416,0.3000699,0.009818411,0.6035073,0.0004419163],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.03080422,0.001067333,0.7941718,0.0006077948,0.0004459356,0.001300118,0.04981844,0.1039513,0.01783317],"genre_scores_gemma":[0.04198376,0.0005608923,0.8293557,0.0009090495,0.0001055919,0.002080908,0.0969687,0.01673815,0.01129738],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.01125014,"threshold_uncertainty_score":0.03763545,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002812135,0.0002091944,0.000281958,0.0003496422,0.0001413865,0.0003968441,0.0001603474,0.0002207934,0.0008722479],"category_scores_gemma":[0.0005116996,0.0002039346,0.0001979034,0.0002577163,0.0002149289,0.0002158486,0.0002404128,0.0003735773,0.0002758592],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002469702,"about_ca_system_score_gemma":0.0001757342,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002597723,"about_ca_topic_score_gemma":0.0006773507,"domain_scores_codex":[0.9997714,0.00004429136,0.00001409096,0.0000765727,0.0000599605,0.00003370364],"domain_scores_gemma":[0.9996619,0.0001159847,0.0001169559,0.00002489107,0.00003734558,0.00004298467],"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":[0.0002157401,0.00001619845,0.005923923,0.0001102121,0.00002478146,0.00007797822,0.0000397585,0.0006484278,0.9867964,0.0003094068,0.0001074613,0.005729635],"study_design_scores_gemma":[0.00003169131,0.0004680554,0.04974537,0.00002915481,0.00008526034,0.0004716247,0.00009795808,0.01257262,0.9293789,0.0009909143,0.006094726,0.00003375746],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9899399,0.002210228,0.005983541,0.00006794166,0.00001983432,0.00002286861,0.0006292872,0.0002148574,0.0009116617],"genre_scores_gemma":[0.9942122,0.0005760815,0.003894605,0.00006891278,0.0000104577,0.00002194203,0.0007227344,0.00002967346,0.0004632338],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0008722479,"threshold_uncertainty_score":0.002918005,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001112688,0.001235352,0.0007973045,0.001801823,0.0009192143,0.001564573,0.001568837,0.0009279342,0.0194229],"category_scores_gemma":[0.002131416,0.0005347704,0.001026539,0.001361473,0.0004982039,0.001498893,0.00221046,0.002062299,0.002606861],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005335886,"about_ca_system_score_gemma":0.0009946839,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003045668,"about_ca_topic_score_gemma":0.004678515,"domain_scores_codex":[0.9995909,0.00006379012,0.00003302083,0.0000922114,0.0001580724,0.00006189848],"domain_scores_gemma":[0.9988239,0.0006699109,0.00006866314,0.0001177075,0.0001683188,0.0001513618],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002327893,0.000260565,0.007996872,0.003325416,0.0005516602,0.002033794,0.002986112,0.02205985,0.3037578,0.01585978,0.3843224,0.2545178],"study_design_scores_gemma":[0.0006768632,0.0002170823,0.01589712,0.0006213018,0.0001706282,0.002013463,0.00114471,0.4010266,0.2247116,0.06676086,0.2861866,0.0005732107],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.02794247,0.0006908601,0.5363787,0.0009405218,0.0004456374,0.0002035868,0.0400125,0.3870187,0.006367055],"genre_scores_gemma":[0.2622591,0.001444174,0.631652,0.001115423,0.0002667517,0.001740191,0.04744452,0.04704072,0.007037144],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.0194229,"threshold_uncertainty_score":0.06497604,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002536088,0.001664912,0.001405914,0.001789634,0.001141566,0.001729859,0.00188673,0.001240807,0.005066273],"category_scores_gemma":[0.002272081,0.001273826,0.001240953,0.001704873,0.0007971188,0.001336384,0.001637334,0.002826435,0.005618275],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007779725,"about_ca_system_score_gemma":0.001511816,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006317373,"about_ca_topic_score_gemma":0.002161594,"domain_scores_codex":[0.9969868,0.0007174875,0.0002632061,0.0006557885,0.001158284,0.0002183444],"domain_scores_gemma":[0.9990746,0.0002725147,0.0001832058,0.000189465,0.0001951863,0.00008506879],"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":[0.0001582222,0.00007311557,0.0004113847,0.0008721653,0.00006517475,0.00009683635,0.00008908111,0.0007487473,0.9581376,0.002355138,0.00399663,0.03299578],"study_design_scores_gemma":[0.00003752661,0.0002005002,0.001780829,0.0001023998,0.00006665095,0.0005156941,0.00005756492,0.007809398,0.900325,0.002367475,0.0866412,0.00009566886],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03095503,0.006768635,0.9343514,0.0005263647,0.0007432292,0.001246358,0.01023767,0.009131054,0.006040239],"genre_scores_gemma":[0.04635984,0.004844915,0.9191062,0.0008165882,0.000192404,0.003330036,0.01688957,0.001255214,0.007205178],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005066273,"threshold_uncertainty_score":0.0169484,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005868885,0.002579668,0.001843663,0.003869913,0.001330699,0.004800529,0.003941725,0.002486418,0.01284671],"category_scores_gemma":[0.007498861,0.001809,0.003622846,0.002005452,0.00133307,0.004059921,0.006721331,0.003744483,0.008313132],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001658501,"about_ca_system_score_gemma":0.003928344,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008106047,"about_ca_topic_score_gemma":0.009989901,"domain_scores_codex":[0.9982134,0.0004289163,0.0002529267,0.0004593005,0.0004674056,0.000177975],"domain_scores_gemma":[0.9979349,0.0008498899,0.0001940032,0.0004622869,0.000258678,0.0003002415],"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":[0.002409315,0.000604372,0.01826398,0.007435571,0.00146656,0.003168261,0.00353933,0.08295561,0.05981154,0.2019079,0.2816013,0.3368362],"study_design_scores_gemma":[0.0003701961,0.0002197984,0.004504823,0.001101344,0.0002477197,0.001232863,0.0004421967,0.252959,0.0168873,0.2298905,0.4916661,0.0004781476],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002944281,0.0009325063,0.8256953,0.0006412515,0.0002178111,0.0005174485,0.01717948,0.1473864,0.004485555],"genre_scores_gemma":[0.04936736,0.002557166,0.8594089,0.0009169186,0.0001848245,0.001887289,0.05847544,0.02251422,0.004687863],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01284671,"threshold_uncertainty_score":0.0429765,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009075155,0.001490805,0.0009948134,0.001501002,0.0009472589,0.0007201133,0.001574354,0.0013041,0.008307171],"category_scores_gemma":[0.001816869,0.0005767362,0.001150411,0.0008783859,0.0003010856,0.001283652,0.0008633754,0.001176391,0.003526048],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005794159,"about_ca_system_score_gemma":0.001313845,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003274743,"about_ca_topic_score_gemma":0.005765484,"domain_scores_codex":[0.9995766,0.00007797228,0.0000301429,0.0001241909,0.0001511423,0.00004001557],"domain_scores_gemma":[0.9995217,0.00020598,0.00005544505,0.00007198762,0.0001038161,0.00004112937],"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":[0.0009390151,0.0003065287,0.006838586,0.0007168034,0.0003968899,0.0004316972,0.0002365731,0.1104945,0.05535132,0.00916458,0.04877533,0.7663482],"study_design_scores_gemma":[0.00007430316,0.0001177884,0.00125233,0.00003252138,0.00004208507,0.000305758,0.00003784076,0.9508452,0.02602488,0.005923558,0.01529376,0.00004986129],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01590445,0.0006611955,0.9381005,0.000139317,0.0001007107,0.0001509749,0.001132629,0.04166089,0.002149362],"genre_scores_gemma":[0.1014804,0.0003200964,0.8848027,0.0002406046,0.00004277581,0.0002529601,0.004801989,0.0036139,0.004444552],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008307171,"threshold_uncertainty_score":0.02779019,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003380366,0.003477603,0.002103804,0.003767779,0.001351751,0.002071918,0.002827182,0.00137624,0.07607163],"category_scores_gemma":[0.007893314,0.00185996,0.003929852,0.002139788,0.0006681141,0.002280156,0.003303849,0.003267267,0.02436588],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006891398,"about_ca_system_score_gemma":0.002320842,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003968278,"about_ca_topic_score_gemma":0.005462916,"domain_scores_codex":[0.9990521,0.0001941815,0.00008252297,0.0003569605,0.0002176367,0.00009667715],"domain_scores_gemma":[0.9974781,0.001647578,0.0001575505,0.0003582521,0.0002357836,0.0001227925],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002098874,0.000320368,0.01322924,0.005959467,0.002445476,0.001290045,0.0016026,0.009106901,0.05036825,0.01256988,0.6909103,0.2100985],"study_design_scores_gemma":[0.002155089,0.0004849523,0.029253,0.0009791473,0.001209873,0.002173028,0.0004633952,0.1198276,0.06345431,0.07552043,0.7035041,0.0009750359],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"software","genre_gemma":"software","genre_scores_codex":[0.004817883,0.000633258,0.3908231,0.0003985132,0.0003970453,0.000318157,0.08597822,0.5136858,0.002947935],"genre_scores_gemma":[0.04246612,0.001025955,0.6877502,0.001291942,0.0001838523,0.003182756,0.1284624,0.1284273,0.007209549],"genre_candidate":"software","genre_consensus":"software","teacher_disagreement_score":0.07607163,"threshold_uncertainty_score":0.2544849,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009176614,0.0003523627,0.0005032799,0.0004247611,0.0003062589,0.0004121857,0.0006485534,0.0004939295,0.001674333],"category_scores_gemma":[0.003097165,0.0002184586,0.0007161102,0.0003921578,0.000313438,0.0003164662,0.0005360382,0.0003775346,0.0001977804],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000251348,"about_ca_system_score_gemma":0.0007153766,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007030213,"about_ca_topic_score_gemma":0.006621133,"domain_scores_codex":[0.9998581,0.00004932541,0.000007092093,0.00003177704,0.00002394593,0.00002958476],"domain_scores_gemma":[0.9993567,0.0004266947,0.00005517376,0.00006194774,0.00005451698,0.00004494768],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0006597892,0.0001721969,0.0895246,0.0001670998,0.0003540878,0.0006182864,0.0001280919,0.835922,0.02335906,0.003860675,0.002978121,0.0422559],"study_design_scores_gemma":[0.00004771912,0.00004328374,0.006737152,0.000005184801,0.0000310519,0.00003846726,0.00003055064,0.9893217,0.002159296,0.001246935,0.0003303272,0.000008344579],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9445405,0.0001688799,0.05279151,0.0002548698,0.00003828646,0.00003031436,0.0005660933,0.0005183876,0.001091133],"genre_scores_gemma":[0.9732562,0.00007953163,0.02539463,0.00007384946,0.00001119533,0.0000284004,0.000719052,0.00005498949,0.0003821547],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.007030213,"threshold_uncertainty_score":0.0139786,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001218106,0.0001914373,0.0002404323,0.0008763253,0.0003014966,0.0004021536,0.0001554186,0.0002642159,0.001802909],"category_scores_gemma":[0.0003850625,0.00009686218,0.0002708371,0.0005905384,0.0001768807,0.0001322832,0.0002498163,0.0001995966,0.0004001182],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000287463,"about_ca_system_score_gemma":0.0001595659,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002506444,"about_ca_topic_score_gemma":0.004041844,"domain_scores_codex":[0.9998916,0.000010666,0.00000946619,0.00005262097,0.00002151933,0.00001403157],"domain_scores_gemma":[0.9997413,0.00005834453,0.00009326541,0.00001546861,0.00005435306,0.00003730683],"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":[0.00061033,0.00003819271,0.06595866,0.0004405454,0.0001820886,0.0003049268,0.0009540259,0.0009162374,0.9142074,0.000744421,0.0003376136,0.01530563],"study_design_scores_gemma":[0.00002909618,0.0003467615,0.8841994,0.00007702391,0.0002366836,0.001716176,0.0009289564,0.003646801,0.09158349,0.001069862,0.01611756,0.00004817635],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9953256,0.0005514724,0.0007996291,0.0000270979,0.000006896308,0.00001056322,0.001959225,0.00003195097,0.001287466],"genre_scores_gemma":[0.993321,0.0001545432,0.001156198,0.00002414543,0.000004737074,0.00001085485,0.004548552,0.00001288919,0.000766909],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.002506444,"threshold_uncertainty_score":0.006031394,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001937003,0.00176606,0.0009112592,0.00337845,0.0009349982,0.002477203,0.002092637,0.001330084,0.1390212],"category_scores_gemma":[0.004079141,0.0008927227,0.001448279,0.001568333,0.0006583927,0.002555955,0.003041713,0.002027427,0.04005928],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005835897,"about_ca_system_score_gemma":0.001075456,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003516199,"about_ca_topic_score_gemma":0.005724124,"domain_scores_codex":[0.9993997,0.0001041821,0.0000545548,0.0001163466,0.0002344703,0.00009078121],"domain_scores_gemma":[0.9979141,0.001199782,0.0001327547,0.0002015895,0.0003144458,0.0002373561],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0005601547,0.0001141298,0.002061608,0.00147259,0.0001362825,0.0007535397,0.0008348135,0.001563196,0.03785883,0.004838933,0.8088124,0.1409935],"study_design_scores_gemma":[0.0004566222,0.0001901493,0.007349227,0.0006113119,0.00008407331,0.001293016,0.0003497964,0.04921483,0.07626272,0.01670535,0.8470396,0.0004432808],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"software","genre_gemma":"software","genre_scores_codex":[0.008054894,0.0005355995,0.2760209,0.001148193,0.0006439615,0.0003396296,0.06361067,0.6314406,0.01820547],"genre_scores_gemma":[0.06870565,0.001051179,0.6460989,0.002947122,0.0003788912,0.002173455,0.09857206,0.1455009,0.03457175],"genre_candidate":"software","genre_consensus":"software","teacher_disagreement_score":0.1390212,"threshold_uncertainty_score":0.4650723,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003979146,0.0003520164,0.0005996256,0.0006440238,0.0004188388,0.001111237,0.0003115539,0.0004910476,0.000807004],"category_scores_gemma":[0.001054782,0.0002405803,0.0004044244,0.0009104902,0.0004445403,0.0005871481,0.0009810751,0.000603374,0.0003964527],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001537259,"about_ca_system_score_gemma":0.00021691,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003262202,"about_ca_topic_score_gemma":0.0007114479,"domain_scores_codex":[0.999681,0.00005779967,0.00002228926,0.0001383872,0.00006084786,0.00003968371],"domain_scores_gemma":[0.9994975,0.0001349092,0.0002138176,0.00004982456,0.00003247691,0.00007136263],"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":[0.00159194,0.0001672135,0.3703555,0.0006553519,0.0003374853,0.001930004,0.0006873998,0.01102063,0.5110425,0.00250807,0.001044939,0.0986589],"study_design_scores_gemma":[0.00005783237,0.0006153357,0.7325054,0.000202527,0.0004068183,0.005427054,0.001695154,0.1211963,0.1150351,0.01400608,0.008687134,0.0001651926],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9794465,0.0007022096,0.01816801,0.0001049096,0.00000860814,0.0000184176,0.0006333978,0.0002185789,0.0006994245],"genre_scores_gemma":[0.9761156,0.000296653,0.02188415,0.0000899962,0.00001235915,0.00001874181,0.00141589,0.00003504494,0.0001315485],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.001111237,"threshold_uncertainty_score":0.002699733,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.000936649,0.0007824397,0.000642267,0.0006436819,0.0004672474,0.0008781717,0.0007113121,0.00104785,0.006050928],"category_scores_gemma":[0.004699748,0.0002541508,0.000937424,0.0008346739,0.0004962417,0.001297735,0.0006118728,0.001088989,0.003968939],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003772737,"about_ca_system_score_gemma":0.0004517677,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002484944,"about_ca_topic_score_gemma":0.003142513,"domain_scores_codex":[0.999448,0.000158513,0.00003159572,0.0002118454,0.00009450812,0.00005556259],"domain_scores_gemma":[0.9980388,0.001053709,0.000167484,0.0004521159,0.0002040261,0.00008381277],"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":[0.003625298,0.0003116336,0.2171141,0.0008381447,0.0005319662,0.003811791,0.0008924771,0.1591797,0.04946089,0.008695281,0.0160218,0.5395169],"study_design_scores_gemma":[0.00007863447,0.000562443,0.08863071,0.0001967249,0.0003294196,0.004250268,0.0003576774,0.8258367,0.02069002,0.03118386,0.02777651,0.0001069681],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"other","genre_scores_codex":[0.8212448,0.003912374,0.1535054,0.001342562,0.0004832135,0.00008953122,0.005243917,0.002613322,0.01156482],"genre_scores_gemma":[0.974943,0.0006251943,0.01360948,0.000316024,0.0001002371,0.00003594198,0.004988681,0.0003563905,0.00502514],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.9989522,"threshold_uncertainty_score":0.02024233,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000871249,0.0004429239,0.000446914,0.0004657182,0.0002649629,0.0005314595,0.0005544114,0.0005818873,0.0007460759],"category_scores_gemma":[0.002509146,0.0002077333,0.0003254907,0.0002802965,0.0003557398,0.0006637581,0.0003226564,0.0004621551,0.0001872894],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005406147,"about_ca_system_score_gemma":0.0006816848,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005882429,"about_ca_topic_score_gemma":0.005808964,"domain_scores_codex":[0.9998052,0.00006527856,0.000007926395,0.00005313379,0.00004599205,0.00002245083],"domain_scores_gemma":[0.9988796,0.0007895301,0.00006498202,0.0001042849,0.00009961577,0.00006190466],"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":[0.0001098433,0.00005359347,0.004181675,0.00003170316,0.00002500809,0.00002888715,0.00003410652,0.975093,0.00526122,0.002129564,0.0002859307,0.01276547],"study_design_scores_gemma":[0.000002227305,0.000007978613,0.0001269096,7.425697e-7,0.000001342027,0.000002691551,0.000002127491,0.9983865,0.0009343079,0.000487767,0.00004527432,0.000002130721],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.776251,0.0002682504,0.2195803,0.0001426051,0.00002943954,0.00003661871,0.0002786895,0.001641109,0.001772086],"genre_scores_gemma":[0.9663628,0.00005988004,0.03283706,0.00003556121,0.000008130495,0.00002467664,0.0002716936,0.00008429496,0.0003158826],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.005882429,"threshold_uncertainty_score":0.0116964,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001253193,0.001996079,0.0009642512,0.002031899,0.0008420749,0.001487321,0.002650388,0.001839964,0.03229816],"category_scores_gemma":[0.005649575,0.0009987943,0.001140215,0.002271382,0.0005959438,0.001832351,0.001862959,0.002158121,0.01388533],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006565809,"about_ca_system_score_gemma":0.000794663,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001827279,"about_ca_topic_score_gemma":0.005717763,"domain_scores_codex":[0.9993368,0.00008449935,0.00008218453,0.0002651849,0.0001867537,0.0000444693],"domain_scores_gemma":[0.9992089,0.0002647307,0.0001425932,0.0001504871,0.0001733496,0.00006008368],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00141893,0.0001837429,0.004741402,0.002801321,0.0004795405,0.000962782,0.001166257,0.0115117,0.1591296,0.01927501,0.465488,0.3328417],"study_design_scores_gemma":[0.0004835766,0.0005159275,0.006785372,0.0002598778,0.0001454333,0.001339863,0.0002031989,0.2098121,0.234262,0.02368261,0.5221419,0.0003681367],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.01502293,0.0006943512,0.5582939,0.0004710072,0.0006873802,0.0003064692,0.02826047,0.3895429,0.006720557],"genre_scores_gemma":[0.05219991,0.0005156575,0.812243,0.0006303482,0.0001496915,0.0006112364,0.0717519,0.05399459,0.007903533],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.03229816,"threshold_uncertainty_score":0.1080481,"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001005787,0.0006799725,0.0007749093,0.003912697,0.0005591726,0.00105269,0.000674932,0.0004028697,0.00233791],"category_scores_gemma":[0.0025138,0.0002035977,0.0008279904,0.004993764,0.0003516222,0.00095175,0.0008756287,0.000594148,0.0008657768],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004623398,"about_ca_system_score_gemma":0.0006967059,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003015203,"about_ca_topic_score_gemma":0.006520015,"domain_scores_codex":[0.9993801,0.0001278687,0.00005785931,0.0001981029,0.0001566553,0.00007938172],"domain_scores_gemma":[0.9989464,0.0003886889,0.0002370526,0.0001762055,0.000126609,0.0001251128],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.003377027,0.00192504,0.4243667,0.006094557,0.00150894,0.001482514,0.001140916,0.08792114,0.1390902,0.01989651,0.1301198,0.1830766],"study_design_scores_gemma":[0.0003642909,0.0006370865,0.2195266,0.0002362589,0.0004891109,0.0009619779,0.0009184519,0.6588827,0.035404,0.01978601,0.06262767,0.0001658947],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9225457,0.002003068,0.008782905,0.0002178364,0.00002505913,0.00009577242,0.06159173,0.002704247,0.002033564],"genre_scores_gemma":[0.7301756,0.001190219,0.02840734,0.00007074957,0.00001895885,0.0001521088,0.2390143,0.000342539,0.0006281975],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.003912697,"threshold_uncertainty_score":0.007821083,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}