{"id":"W4416774276","doi":"10.1039/d5sc05225d","title":"SynTwins: a retrosynthesis-guided framework for synthesizable molecular analog generation","year":2025,"lang":"en","type":"article","venue":"Chemical Science","topic":"Machine Learning in Materials Science","field":"Materials Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; Canadian Institute for Advanced Research","funders":"National Research Foundation of Korea; Seoul National University; National Research Foundation","keywords":"Bottleneck; Virtual screening; Similarity (geometry); Range (aeronautics); Benchmarking; Generative grammar","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.00259922,0.0002185927,0.0003035782,0.0001645656,0.0005199708,0.0005438638,0.001604822,0.0001427796,0.0003607179],"category_scores_gemma":[0.01129125,0.0001904601,0.00008475166,0.001207027,0.001041114,0.0004048676,0.0003126976,0.0001436561,0.00009335478],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002289446,"about_ca_system_score_gemma":0.0003382754,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002961457,"about_ca_topic_score_gemma":7.921995e-7,"domain_scores_codex":[0.9970672,0.00007490101,0.0004248427,0.001031735,0.0006435879,0.0007576987],"domain_scores_gemma":[0.9982562,0.0004481946,0.0001581962,0.000715708,0.000249858,0.0001718408],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00001055395,0.0000307463,0.00008825563,0.00003404549,0.000001584822,0.000001462323,0.00004096563,0.0004584533,0.9529556,0.0453952,0.0005298603,0.0004532834],"study_design_scores_gemma":[0.0000895894,0.00002018801,0.00008398113,0.00009642402,0.00001200802,0.000003950123,0.00001473571,0.0242271,0.9530028,0.02151418,0.000721208,0.0002138187],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7139415,0.00006360815,0.2815756,0.00107605,0.0007963449,0.0003341293,0.0000106477,0.0001694378,0.00203274],"genre_scores_gemma":[0.8045321,0.000002863749,0.1943053,0.000786452,0.0001201144,0.0001061311,0.000002498498,0.0000116627,0.0001329951],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.09059055,"threshold_uncertainty_score":0.9970371,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01987913973713504,"score_gpt":0.3221345168819746,"score_spread":0.3022553771448396,"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."}}