{"id":"W4223937330","doi":"10.3390/molecules27082513","title":"A Consensus Compound/Bioactivity Dataset for Data-Driven Drug Design and Chemogenomics","year":2022,"lang":"en","type":"article","venue":"Molecules","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"European Commission; European Federation of Pharmaceutical Industries and Associations; Diamond Light Source; Innovative Medicines Initiative; McGill University","keywords":"chEMBL; PubChem; Drug discovery; Computer science; Chemical space; Computational biology; Combinatorial chemistry; Chemistry; Bioinformatics; Biology","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008652178,0.0001196959,0.0001532002,0.00006418316,0.0003250448,0.0001435745,0.001139723,0.00001294343,0.000004213422],"category_scores_gemma":[0.00008600775,0.000134717,0.00002419919,0.0001459265,0.00007046462,0.0001665503,0.00231697,0.0001030427,0.00000223899],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005747469,"about_ca_system_score_gemma":0.0001655618,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004179799,"about_ca_topic_score_gemma":0.000005800051,"domain_scores_codex":[0.9985091,0.0003898307,0.0001541083,0.0005493169,0.0002085197,0.0001891278],"domain_scores_gemma":[0.9978801,0.001165796,0.00008942647,0.0007740107,0.00002733151,0.00006335836],"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":[0.0003017206,0.0005595715,0.0001564869,0.00007817541,0.0002424891,0.0001790282,0.00118129,0.6522875,0.0103852,0.06095843,0.2120135,0.06165656],"study_design_scores_gemma":[0.0003794922,0.00003945992,0.0001324155,0.000001626238,0.00001190498,0.00007473803,0.00002665066,0.9597485,0.0009374451,0.008623977,0.02985687,0.000166915],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03295198,0.0001352468,0.9616835,0.001531866,0.0001537457,0.0003981164,0.003077792,0.00004589861,0.00002190216],"genre_scores_gemma":[0.1426117,0.000003150633,0.8547646,0.0006266797,0.00002424529,0.00006116425,0.001877497,0.00001373053,0.0000172626],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.307461,"threshold_uncertainty_score":0.5493601,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07759341046365603,"score_gpt":0.321248517982168,"score_spread":0.2436551075185119,"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."}}