{"id":"W3177472573","doi":"10.1101/2021.06.20.449169","title":"High-accuracy alignment ensembles enable unbiased assessments of sequence homology and phylogeny","year":2021,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":162,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"University of British Columbia","keywords":"Phylogenetic tree; Phylogenetics; Inference; Polyphyly; Computer science; Multiple sequence alignment; Tree (set theory); Hidden Markov model; Computational biology; Biology; Artificial intelligence; Evolutionary biology; Sequence alignment; Mathematics; Clade; Genetics; Combinatorics; Gene","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004852024,0.001012256,0.001227543,0.001498189,0.001125736,0.002019004,0.001021781,0.001505549,0.001733894],"category_scores_gemma":[0.01799474,0.0007563957,0.000716945,0.001335066,0.0009131225,0.003123494,0.002618222,0.002883482,0.001154115],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005208501,"about_ca_system_score_gemma":0.0005947681,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00055747,"about_ca_topic_score_gemma":0.0007501063,"domain_scores_codex":[0.9964178,0.001634524,0.0001440171,0.0005945105,0.000993156,0.0002161378],"domain_scores_gemma":[0.9888145,0.005487756,0.001156358,0.002549634,0.001492817,0.0004989289],"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":[0.0006823107,0.0001938436,0.02360933,0.0003454865,0.0005607904,0.0003499216,0.0005220772,0.541878,0.1593064,0.06146989,0.006091019,0.2049909],"study_design_scores_gemma":[0.00001237666,0.00005132802,0.002356505,0.00002970709,0.00003862423,0.00008790469,0.00003519679,0.935631,0.02365245,0.03530599,0.002761453,0.00003747997],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.09656768,0.0005181149,0.8991399,0.0002453578,0.00009168089,0.00003113454,0.0003126163,0.0014956,0.00159785],"genre_scores_gemma":[0.6054863,0.0004107165,0.3908184,0.00016424,0.0001524611,0.0001098559,0.001162526,0.0006967658,0.0009987537],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004852024,"threshold_uncertainty_score":0.02566028,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01917017094318144,"score_gpt":0.2490745188194095,"score_spread":0.229904347876228,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}