{"id":"W4405969769","doi":"10.1101/2024.12.31.630607","title":"Empathi: Embedding-based Phage Protein Annotation Tool by Hierarchical Assignment","year":2025,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Bacteriophages and microbial interactions","field":"Environmental Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université Laval","funders":"","keywords":"Computational biology; Genome; Biology; Annotation; Homology (biology); Protein sequencing; Protein function prediction; Bacteriophage; Genetics; Peptide sequence; Protein function; Amino acid; Gene; Escherichia coli","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001254863,0.001845765,0.0007318768,0.003907107,0.0005985136,0.001349509,0.001495558,0.001488936,0.00714426],"category_scores_gemma":[0.003472622,0.000487505,0.001358461,0.001760019,0.0003725533,0.001937666,0.002127842,0.001632362,0.00537162],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005749514,"about_ca_system_score_gemma":0.0006904102,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001644197,"about_ca_topic_score_gemma":0.001795088,"domain_scores_codex":[0.9989921,0.0001713685,0.00007646903,0.0003717327,0.0002945363,0.00009376484],"domain_scores_gemma":[0.9987167,0.0005606423,0.0001438596,0.0002424144,0.0002227296,0.0001136951],"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":[0.002502776,0.000886012,0.02775963,0.002455832,0.0005865634,0.0009415934,0.0005492812,0.03178284,0.07253266,0.006459058,0.218184,0.6353599],"study_design_scores_gemma":[0.0002413195,0.0004539841,0.009681604,0.0001908492,0.000125707,0.0006971357,0.0002351955,0.8665452,0.04470844,0.01644852,0.06050809,0.0001640072],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.09608347,0.002023558,0.5188527,0.0006823285,0.0004843418,0.000443509,0.0517229,0.3250051,0.00470215],"genre_scores_gemma":[0.3074809,0.000678786,0.5585297,0.0005047899,0.0001983692,0.0008213703,0.1194655,0.006154,0.006166614],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.00714426,"threshold_uncertainty_score":0.02389997,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007548414655204148,"score_gpt":0.2284023541719888,"score_spread":0.2208539395167847,"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."}}