{"id":"W4413257724","doi":"10.1016/j.cell.2025.07.033","title":"A generative deep learning approach to de novo antibiotic design","year":2025,"lang":"en","type":"article","venue":"Cell","topic":"Computational Drug Discovery Methods","field":"Computer Science","cited_by":63,"is_retracted":false,"has_abstract":false,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"Audacious Project; National Science Foundation Graduate Research Fellowship Program; National Institute of Allergy and Infectious Diseases; National Institute of General Medical Sciences; Defense Threat Reduction Agency; Knut och Alice Wallenbergs Stiftelse; National Institutes of Health; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; Flu Lab; Broad Institute; Siebel Scholars Foundation; Oracle; Division of Intramural Research, National Institute of Allergy and Infectious Diseases; Wyss Foundation; James S. McDonnell Foundation","keywords":"Biology; In silico; Antibiotics; Neisseria gonorrhoeae; Computational biology; Synthetic biology; Staphylococcus aureus; Generative Design; Antibacterial activity; Antimicrobial; Antibiotic resistance; Generative grammar; Generative model; Antimicrobial peptides; Microbiology; Artificial intelligence; Bacteria; Genetics; Computer science; Gene; Materials science","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005198974,0.00009646091,0.0001091652,0.0001358938,0.000114072,0.0001495777,0.0005047141,0.00003351337,0.000002241086],"category_scores_gemma":[0.0001009232,0.00009760746,0.00003801184,0.0006280279,0.00001640827,0.0001276728,0.0002507165,0.0001283718,0.00004839108],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006419588,"about_ca_system_score_gemma":0.0001676374,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004849656,"about_ca_topic_score_gemma":4.937746e-7,"domain_scores_codex":[0.9988242,0.0003808791,0.0001157271,0.0003360003,0.0001257229,0.0002174796],"domain_scores_gemma":[0.9993265,0.0003086078,0.00002856396,0.0002149276,0.0000542549,0.00006719236],"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":[0.000004182581,0.00007341975,0.0001130757,0.00001713562,0.000007894858,0.000005187818,0.001272431,0.9678521,0.005551656,0.01500831,0.0002566582,0.009837963],"study_design_scores_gemma":[0.0001458298,0.00003420032,0.001606087,0.000009977025,0.000004484035,0.000004512742,0.00003876023,0.9560584,0.03692696,0.004374937,0.0006801253,0.0001157154],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01484294,0.0001178599,0.9588838,0.0002153844,0.0001558775,0.0001613488,1.654211e-7,0.00008160053,0.02554105],"genre_scores_gemma":[0.3690277,0.000001556385,0.6294178,0.0004913525,0.00002160614,0.000004956701,7.874633e-7,0.000004038308,0.001030105],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3541848,"threshold_uncertainty_score":0.3980317,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03029241096829063,"score_gpt":0.2923361754246098,"score_spread":0.2620437644563192,"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."}}