{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T20:10:01Z","timestamp":1789071001468,"version":"build-2803163510"},"reference-count":36,"publisher":"Oxford University Press (OUP)","issue":"5","license":[{"start":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T00:00:00Z","timestamp":1788998400000},"content-version":"vor","delay-in-days":9,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,9,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>As a critical step in drug discovery, molecule optimization aims to improve specific properties of lead compounds. Inspired by isosteres (i.e. molecular substructures with similar reactive electron shells), existing AI-based methods have been proposed for molecule optimization. However, they often rely on available matched molecular pairs (MMPs), thus facing two essential challenges. First, the limited diversity of available MMPs leads to biased training and a lack of novelty in the optimized molecules, in both single-objective (SO) and multi-objective (MO) optimization. Second, the low number of available MMPs leads to limited generalization in MO optimization. To overcome the limitations of MMPs, we propose DFRL-Mol, a novel dual-stage reinforcement learning framework. It consists of three main modules: the Structure\u2013Property Analyst (SPA), the Decorator (DCR), and Curriculum-guided Dual-Stage Policy Optimization (CDSPO). SPA and DCR are two LLM-based models, which are responsible for locating the substructures to be optimized and inferring their appropriate isosteric replacement, respectively. CDSPO leverages curriculum-guided reinforcement learning to facilitate collaboration between the two models. DFRL-Mol ensures the generation of diverse molecular representations while effectively overcoming the issues of bias and data sparsity associated with MMPs. The comparisons with state-of-the-art methods demonstrate the enhanced performance of DFRL-Mol in multiple scenarios of molecule optimization, including SO optimization, substructure-constrained (SC) SO optimization, MO optimization, and MO molecule generation. More importantly, detailed analysis of experimental results demonstrates that the models\u2019 behavior is consistent with chemical intuition.<\/jats:p>","DOI":"10.1093\/bib\/bbag493","type":"journal-article","created":{"date-parts":[[2026,8,22]],"date-time":"2026-08-22T11:42:01Z","timestamp":1787398921000},"source":"Crossref","is-referenced-by-count":0,"title":["DFRL-Mol: a dual-stage framework of reinforcement learning for multi-scenario molecule optimization"],"prefix":"10.1093","volume":"27","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-4574-9886","authenticated-orcid":false,"given":"Zhenyi","family":"Wu","sequence":"first","affiliation":[{"name":"School of Life Science and Technology, Northwestern Polytechnical University , 127 West Youyi Road, Beilin District, Xi\u2019an, Shaanxi 710072 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pengcheng","family":"Zhao","sequence":"additional","affiliation":[{"name":"School of Computer Science, Northwestern Polytechnical University , 127 West Youyi Road, Beilin District, Xi\u2019an, Shaanxi 710072 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jianing","family":"Li","sequence":"additional","affiliation":[{"name":"School of Computer Science, Northwestern Polytechnical University , 127 West Youyi Road, Beilin District, Xi\u2019an, Shaanxi 710072 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qiong","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Life Science and Technology, Northwestern Polytechnical University , 127 West Youyi Road, Beilin District, Xi\u2019an, Shaanxi 710072 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qin","family":"Zhang","sequence":"additional","affiliation":[{"name":"School of Life Science and Technology, Northwestern Polytechnical University , 127 West Youyi Road, Beilin District, Xi\u2019an, Shaanxi 710072 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5314-8127","authenticated-orcid":false,"given":"Hui","family":"Yu","sequence":"additional","affiliation":[{"name":"School of Computer Science, Northwestern Polytechnical University , 127 West Youyi Road, Beilin District, Xi\u2019an, Shaanxi 710072 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhe","family":"Yu","sequence":"additional","affiliation":[{"name":"Department of Orthopedic Surgery, Orthopedics Oncology Institute of Chinese PLA, Tangdu Hospital, Air Force Military Medical University , 1 Xinsi Road, Baqiao District, Xi\u2019an 710038 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2303-273X","authenticated-orcid":false,"given":"Jian-Yu","family":"Shi","sequence":"additional","affiliation":[{"name":"School of Life Science and Technology, Northwestern Polytechnical University , 127 West Youyi Road, Beilin District, Xi\u2019an, Shaanxi 710072 ,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2026,9,10]]},"reference":[{"key":"2026091015383531200_ref1","doi-asserted-by":"publisher","first-page":"3147","DOI":"10.1021\/cr950066q","article-title":"Bioisosterism: a rational approach in drug design","volume":"96","author":"Patani","year":"1996","journal-title":"Chem Rev"},{"key":"2026091015383531200_ref2","journal-title":"8th International Conference on Learning Representations (ICLR 2020)"},{"key":"2026091015383531200_ref3","doi-asserted-by":"publisher","first-page":"1572","DOI":"10.1021\/acscentsci.9b00576","article-title":"Molecular transformer: a model for uncertainty-calibrated chemical reaction prediction","volume":"5","author":"Schwaller","year":"2019","journal-title":"ACS Cent Sci"},{"key":"2026091015383531200_ref4","volume-title":"Proceedings of the 37th International Conference on Machine Learning"},{"key":"2026091015383531200_ref5","doi-asserted-by":"publisher","first-page":"724","DOI":"10.1016\/j.drudis.2013.03.003","article-title":"Matched molecular pair analysis in drug discovery","volume":"18","author":"Dossetter","year":"2013","journal-title":"Drug Discov Today"},{"key":"2026091015383531200_ref6","doi-asserted-by":"publisher","volume-title":"Artificial Intelligence in Drug Design. Methods in Molecular Biology, vol 2390","DOI":"10.1007\/978-1-0716-1787-8_23"},{"key":"2026091015383531200_ref7","doi-asserted-by":"publisher","first-page":"828","DOI":"10.1039\/C9ME00039A","article-title":"Deep learning for molecular design\u2014a review of the state of the art","volume":"4","author":"Elton","year":"2019","journal-title":"Molecular Systems Design & Engineering"},{"key":"2026091015383531200_ref8","doi-asserted-by":"publisher","volume-title":"Applications of Evolutionary Computation. Lecture Notes in Computer Science, vol 12694","DOI":"10.1007\/978-3-030-72699-7_6"},{"key":"2026091015383531200_ref9","doi-asserted-by":"publisher","first-page":"e2200215","DOI":"10.1002\/minf.202200215","article-title":"Fragment-based deep molecular generation using hierarchical chemical graph representation and multi-resolution graph variational autoencoder","volume":"42","author":"Gao","year":"2023","journal-title":"Mol Inform"},{"key":"2026091015383531200_ref10","doi-asserted-by":"publisher","first-page":"11","DOI":"10.1186\/s13321-023-00681-4","article-title":"Generative model based on junction tree variational autoencoder for HOMO value prediction and molecular optimization","volume":"15","author":"Kondratyev","year":"2023","journal-title":"J Cheminform"},{"key":"2026091015383531200_ref11","volume-title":"Proceedings of the 35th International Conference on Machine Learning."},{"key":"2026091015383531200_ref12","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1805.11973"},{"key":"2026091015383531200_ref13","doi-asserted-by":"publisher","first-page":"2","DOI":"10.1186\/s13321-019-0404-1","article-title":"Mol-CycleGAN: a generative model for molecular optimization","volume":"12","author":"Maziarka","year":"2020","journal-title":"J Cheminform"},{"key":"2026091015383531200_ref14","doi-asserted-by":"publisher","volume-title":"Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD '20)","DOI":"10.1145\/3394486.3403104"},{"key":"2026091015383531200_ref15","volume":"31","journal-title":"Adv Neural Inf Proces Syst"},{"key":"2026091015383531200_ref16","doi-asserted-by":"publisher","first-page":"48","DOI":"10.1186\/s13321-017-0235-x","article-title":"Molecular de-novo design through deep reinforcement learning","volume":"9","author":"Olivecrona","year":"2017","journal-title":"J Cheminform"},{"key":"2026091015383531200_ref17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2409.02231"},{"key":"2026091015383531200_ref18","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2502.12845"},{"key":"2026091015383531200_ref19","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2211.09085"},{"key":"2026091015383531200_ref20","doi-asserted-by":"publisher","first-page":"1256","DOI":"10.1038\/s42256-022-00580-7","article-title":"Large-scale chemical language representations capture molecular structure and properties","volume":"4","author":"Ross","year":"2022","journal-title":"Nature Machine Intelligence"},{"key":"2026091015383531200_ref21","doi-asserted-by":"publisher","first-page":"2514","DOI":"10.1039\/D4SC03921A","article-title":"A review of large language models and autonomous agents in chemistry","volume":"16","author":"Ramos","year":"2025","journal-title":"Chem Sci"},{"key":"2026091015383531200_ref22"},{"key":"2026091015383531200_ref23","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2402.03300"},{"key":"2026091015383531200_ref24","doi-asserted-by":"publisher","first-page":"16921","DOI":"10.1021\/acs.jmedchem.5c01641","article-title":"A data-driven perspective on bioisostere evaluation: mapping the benzene bioisostere landscape with BioSTAR","volume":"68","author":"Hern\u00e1ndez-Llad\u00f3","year":"2025","journal-title":"J Med Chem"},{"key":"2026091015383531200_ref25","doi-asserted-by":"publisher","volume-title":"Proceedings of the 26th Annual International Conference on Machine Learning (ICML 2009).","DOI":"10.1145\/1553374.1553380"},{"key":"2026091015383531200_ref26","doi-asserted-by":"publisher","first-page":"4350","DOI":"10.1021\/jm020155c","article-title":"Do structurally similar molecules have similar biological activity?","volume":"45","author":"Martin","year":"2002","journal-title":"J Med Chem"},{"key":"2026091015383531200_ref27","doi-asserted-by":"publisher","first-page":"1040","DOI":"10.1038\/s42256-021-00410-2","article-title":"A deep generative model for molecule optimization via one fragment modification","volume":"3","author":"Chen","year":"2021","journal-title":"Nature machine intelligence"},{"key":"2026091015383531200_ref28","doi-asserted-by":"publisher","DOI":"10.1101\/2023.04.27.538051"},{"key":"2026091015383531200_ref29","doi-asserted-by":"publisher","first-page":"1359","DOI":"10.1038\/s42256-024-00916-5","article-title":"Leveraging language model for advanced multiproperty molecular optimization via prompt engineering","volume":"6","author":"Wu","year":"2024","journal-title":"Nature Machine Intelligence"},{"key":"2026091015383531200_ref30","journal-title":"The Thirteenth International Conference on Learning Representations (ICLR 2025)"},{"key":"2026091015383531200_ref31","volume-title":"Proceedings of the 37th International Conference on Machine Learning."},{"key":"2026091015383531200_ref32","journal-title":"The Tenth International Conference on Learning Representations (ICLR 2022)."},{"key":"2026091015383531200_ref33","doi-asserted-by":"publisher","first-page":"2887","DOI":"10.1021\/jm9602928","article-title":"The properties of known drugs. 1. Molecular frameworks","volume":"39","author":"Bemis","year":"1996","journal-title":"J Med Chem"},{"key":"2026091015383531200_ref34","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1038\/s44386-025-00017-2","article-title":"\"recent advances in molecular representation methods and their applications in scaffold hopping,\" npj","volume":"2","author":"Wang","year":"2025","journal-title":"Drug Discovery"},{"key":"2026091015383531200_ref35","doi-asserted-by":"publisher","volume-title":"Methods and Principles in Medicinal Chemistry","DOI":"10.1002\/9783527675142"},{"key":"2026091015383531200_ref36","volume-title":"Proceedings of the 38th International Conference on Machine Learning"}],"container-title":["Briefings in Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/27\/5\/bbag493\/71022330\/bbag493.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/27\/5\/bbag493\/71022330\/bbag493.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T19:38:44Z","timestamp":1789069124000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bib\/article\/doi\/10.1093\/bib\/bbag493\/8790176"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,1]]},"references-count":36,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2026,9,1]]}},"URL":"https:\/\/doi.org\/10.1093\/bib\/bbag493","relation":{},"ISSN":["1467-5463","1477-4054"],"issn-type":[{"value":"1467-5463","type":"print"},{"value":"1477-4054","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,9]]},"published":{"date-parts":[[2026,9,1]]},"article-number":"bbag493"}}