{"id":"W4412807337","doi":"10.3389/fphar.2025.1617142","title":"PolyLLM: polypharmacy side effect prediction via LLM-based SMILES encodings","year":2025,"lang":"en","type":"article","venue":"Frontiers in Pharmacology","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Windsor","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Polypharmacy; Drug; Computer science; Side effect (computer science); Representation (politics); Adverse effect; Pharmacology; Machine learning; Medicine; Artificial intelligence","routes":{"ca_aff":true,"ca_fund":true,"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":[0.0004415614,0.001056263,0.0006941834,0.001066415,0.0002176089,0.0007622518,0.0009060726,0.0008218654,0.006944901],"category_scores_gemma":[0.002307912,0.000225091,0.0009325378,0.0006806636,0.0002481896,0.001262335,0.0007746536,0.0009247397,0.001915859],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008273504,"about_ca_system_score_gemma":0.000881078,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00322413,"about_ca_topic_score_gemma":0.005511636,"domain_scores_codex":[0.999759,0.00006206258,0.00002349994,0.0000480455,0.00007363952,0.00003376042],"domain_scores_gemma":[0.9993697,0.0002990377,0.000095862,0.00008825202,0.0001146395,0.00003257188],"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":[0.001541966,0.0005146519,0.01055091,0.0006432529,0.0002113345,0.0006758684,0.00008748985,0.3983504,0.01135555,0.01388886,0.02987852,0.5323012],"study_design_scores_gemma":[0.0000289463,0.0000987696,0.0004440269,0.00002266784,0.00002157552,0.00006601774,0.00001049946,0.9861668,0.003606641,0.00681113,0.002709315,0.00001370598],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1648242,0.003153494,0.7669923,0.001873466,0.0004402288,0.0003543084,0.01602194,0.03355677,0.01278333],"genre_scores_gemma":[0.7647402,0.001070999,0.2103128,0.0007251832,0.0001365725,0.0004522484,0.01551162,0.0006768063,0.006373478],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006944901,"threshold_uncertainty_score":0.023233,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.006346681836912317,"score_gpt":0.3037333860147088,"score_spread":0.2973867041777964,"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."}}