{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,21]],"date-time":"2026-05-21T04:11:04Z","timestamp":1779336664005,"version":"3.51.4"},"reference-count":46,"publisher":"Oxford University Press (OUP)","issue":"5","license":[{"start":{"date-parts":[[2026,4,5]],"date-time":"2026-04-05T00:00:00Z","timestamp":1775347200000},"content-version":"vor","delay-in-days":1,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,5,3]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Motivation<\/jats:title>\n                    <jats:p>De novo molecular design remains a fundamental challenge in drug discovery, requiring simultaneous optimization of multiple conflicting objectives such as drug-likeness, synthetic accessibility, and novelty while maintaining chemical validity. We present HybridMolGen, a novel unified framework that synergistically combines three complementary deep learning paradigms: (1) diffusion probabilistic models that generate high-quality, chemically valid molecular samples through gradual noise removal, (2) SE(3)-equivariant graph neural networks that enforce geometric and topological constraints ensuring structural validity and molecular diversity, and (3) property-conditioned transformers that enable fine-grained control over multiple objectives through multi-layer cross-attention modulation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>These components operate within a multi-objective reinforcement learning paradigm that discovers optimal property tradeoffs without manual weight tuning. Extensive benchmarking on MOSES, GuacaMol, and ZINC-250k datasets demonstrates state-of-the-art performance: 99.7% validity, 94.3% novelty, average QED score of 0.753, and 4.9% improvement in GuacaMol overall scores. Critically, HybridMolGen discovers 1.57\u00d7 more molecules satisfying all target criteria simultaneously (91.3% versus 58.3% for CPRL) and generates 2.23\u00d7 more Pareto-efficient solutions compared to traditional scalarization, demonstrating genuine architectural synergy beyond simple component aggregation. Comprehensive ablation studies confirm that the three-way integration outperforms even the best two-component combination by 6.5%, positioning HybridMolGen as a powerful tool for accelerating drug discovery pipelines.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Availability and implementation<\/jats:title>\n                    <jats:p>Implementation code is available as Supplementary Material, available as supplementary data at Bioinformatics online.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1093\/bioinformatics\/btag170","type":"journal-article","created":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T11:47:45Z","timestamp":1775130465000},"source":"Crossref","is-referenced-by-count":0,"title":["HybridMolGen: a unified framework for goal-directed molecular generation via multi-objective reinforcement learning"],"prefix":"10.1093","volume":"42","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8571-0949","authenticated-orcid":false,"given":"Masoud","family":"Amiri","sequence":"first","affiliation":[{"name":"Department of Biomedical Engineering, School of Medicine, Kermanshah University of Medical Sciences , Kermanshah, 6715847141,","place":["Iran"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zahra","family":"Nasirinia","sequence":"additional","affiliation":[{"name":"Department of Biomedical Engineering, School of Medicine, Kermanshah University of Medical Sciences , Kermanshah, 6715847141,","place":["Iran"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2026,4,4]]},"reference":[{"key":"2026052023581330600_btag170-B1","doi-asserted-by":"crossref","first-page":"2064","DOI":"10.1021\/acs.jcim.1c00600","article-title":"MolGPT: molecular generation using a transformer-decoder model","volume":"62","author":"Bagal","year":"2022","journal-title":"J Chem Inf Model"},{"key":"2026052023581330600_btag170-B2","first-page":"41","article-title":"Curriculum learning","author":"Bengio","year":"2009","journal-title":"Proc ICML"},{"key":"2026052023581330600_btag170-B3","doi-asserted-by":"crossref","first-page":"90","DOI":"10.1038\/nchem.1243","article-title":"Quantifying the chemical beauty of drugs","volume":"4","author":"Bickerton","year":"2012","journal-title":"Nat Chem"},{"key":"2026052023581330600_btag170-B4","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1002\/(SICI)1098-1128(199601)16:1<3::AID-MED1>3.0.CO;2-6","article-title":"The art and practice of structure-based drug design","volume":"16","author":"Bohacek","year":"1996","journal-title":"Med Res Rev"},{"key":"2026052023581330600_btag170-B5","doi-asserted-by":"crossref","first-page":"1096","DOI":"10.1021\/acs.jcim.8b00839","article-title":"GuacaMol: benchmarking models for de novo molecular design","volume":"59","author":"Brown","year":"2019","journal-title":"J Chem Inf Model"},{"key":"2026052023581330600_btag170-B77"},{"key":"2026052023581330600_btag170-B6","doi-asserted-by":"crossref","first-page":"182","DOI":"10.1109\/4235.996017","article-title":"A fast and elitist multiobjective genetic algorithm: NSGA-II","volume":"6","author":"Deb","year":"2002","journal-title":"IEEE Trans Evol Computat"},{"key":"2026052023581330600_btag170-B7"},{"key":"2026052023581330600_btag170-B8","author":"De Cao","year":"2018"},{"key":"2026052023581330600_btag170-B9","doi-asserted-by":"crossref","first-page":"20","DOI":"10.1016\/j.jhealeco.2016.01.012","article-title":"Innovation in the pharmaceutical industry: new estimates of R&D costs","volume":"47","author":"DiMasi","year":"2016","journal-title":"J Health Econ"},{"key":"2026052023581330600_btag170-B10","doi-asserted-by":"crossref","first-page":"8","DOI":"10.1186\/1758-2946-1-8","article-title":"Estimation of synthetic accessibility score of drug-like molecules","volume":"1","author":"Ertl","year":"2009","journal-title":"J Cheminform"},{"key":"2026052023581330600_btag170-B11","first-page":"3371","article-title":"Multi-objective molecule generation using interpretable substructures","volume":"119","author":"Fu","year":"2020","journal-title":"Proc ICML"},{"key":"2026052023581330600_btag170-B12","doi-asserted-by":"crossref","first-page":"268","DOI":"10.1021\/acscentsci.7b00572","article-title":"Automatic chemical design using a data-driven continuous representation of molecules","volume":"4","author":"G\u00f3mez-Bombarelli","year":"2018","journal-title":"ACS Cent Sci"},{"key":"2026052023581330600_btag170-B13","first-page":"2672","article-title":"Generative adversarial nets","volume":"27","author":"Goodfellow","year":"2014","journal-title":"NeurIPS"},{"key":"2026052023581330600_btag170-B14"},{"key":"2026052023581330600_btag170-B16","doi-asserted-by":"crossref","first-page":"26","DOI":"10.1007\/s10458-022-09552-y","article-title":"A practical guide to multi-objective reinforcement learning and planning","volume":"36","author":"Hayes","year":"2022","journal-title":"Auton Agents Multi-Agent Syst"},{"key":"2026052023581330600_btag170-B17","first-page":"6840","article-title":"Denoising diffusion probabilistic models","volume":"33","author":"Ho","year":"2020","journal-title":"NeurIPS"},{"key":"2026052023581330600_btag170-B18","first-page":"8867","article-title":"Equivariant diffusion for molecule generation in 3D","volume":"162","author":"Hoogeboom","year":"2022","journal-title":"Proc ICML"},{"key":"2026052023581330600_btag170-B20","first-page":"2323","article-title":"Junction tree variational autoencoder for molecular graph generation","volume":"80","author":"Jin","year":"2018","journal-title":"Proc ICML"},{"key":"2026052023581330600_btag170-B21","author":"Kingma","year":"2014"},{"key":"2026052023581330600_btag170-B22","first-page":"5702","article-title":"QADD: quality-aware drug design using reinforcement learning","volume":"63","author":"Krishnan","year":"2023","journal-title":"J Chem Inf Model"},{"key":"2026052023581330600_btag170-B23","doi-asserted-by":"crossref","first-page":"436","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"2026052023581330600_btag170-B24","author":"Li","year":"2023"},{"key":"2026052023581330600_btag170-B25","first-page":"1149","article-title":"Multi-modal molecular pre-training for drug discovery","volume":"5","author":"Liu","year":"2023","journal-title":"Nat Mach Intell"},{"key":"2026052023581330600_btag170-B26","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4615-5563-6","volume-title":"Nonlinear Multiobjective Optimization","author":"Miettinen","year":"1998"},{"key":"2026052023581330600_btag170-B27","doi-asserted-by":"crossref","first-page":"150","DOI":"10.1038\/s42004-024-01233-z","article-title":"Geometry-complete diffusion for 3D molecule generation and optimization","volume":"7","author":"Morehead","year":"2024","journal-title":"Commun Chem"},{"key":"2026052023581330600_btag170-B28","doi-asserted-by":"crossref","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":"2026052023581330600_btag170-B29","doi-asserted-by":"crossref","first-page":"203","DOI":"10.1038\/nrd3078","article-title":"How to improve R&D productivity: the pharmaceutical industry\u2019s grand challenge","volume":"9","author":"Paul","year":"2010","journal-title":"Nat Rev Drug Discov"},{"key":"2026052023581330600_btag170-B30","doi-asserted-by":"crossref","first-page":"675","DOI":"10.1007\/s10822-013-9672-4","article-title":"Estimation of the size of drug-like chemical space based on GDB-17 data","volume":"27","author":"Polishchuk","year":"2013","journal-title":"J Comput Aided Mol Des"},{"key":"2026052023581330600_btag170-B31","doi-asserted-by":"crossref","first-page":"565644","DOI":"10.3389\/fphar.2020.565644","article-title":"Molecular sets (MOSES): a benchmarking platform for molecular generation models","volume":"11","author":"Polykovskiy","year":"2020","journal-title":"Front Pharmacol"},{"key":"2026052023581330600_btag170-B32","doi-asserted-by":"crossref","first-page":"67","DOI":"10.1613\/jair.3987","article-title":"A survey of multi-objective sequential decision-making","volume":"48","author":"Roijers","year":"2013","journal-title":"JAIR"},{"key":"2026052023581330600_btag170-B33","first-page":"9323","article-title":"E(n) equivariant graph neural networks","volume":"139","author":"Satorras","year":"2021","journal-title":"Proc ICML"},{"key":"2026052023581330600_btag170-B34","first-page":"1027","article-title":"Structure-based drug design with equivariant diffusion models","volume":"3","author":"Schneuing","year":"2023","journal-title":"Nat Comput Sci"},{"key":"2026052023581330600_btag170-B35","author":"Schulman","year":"2017"},{"key":"2026052023581330600_btag170-B36","doi-asserted-by":"crossref","first-page":"120","DOI":"10.1021\/acscentsci.7b00512","article-title":"Generating focused molecule libraries for drug discovery with recurrent neural networks","volume":"4","author":"Segler","year":"2018","journal-title":"ACS Cent Sci"},{"key":"2026052023581330600_btag170-B37","author":"Shi","year":"2020"},{"key":"2026052023581330600_btag170-B38","author":"Song","year":"2021"},{"key":"2026052023581330600_btag170-B39","doi-asserted-by":"crossref","first-page":"2324","DOI":"10.1021\/acs.jcim.5b00559","article-title":"ZINC 15\u2014ligand discovery for everyone","volume":"55","author":"Sterling","year":"2015","journal-title":"J Chem Inf Model"},{"key":"2026052023581330600_btag170-B40","first-page":"3483","article-title":"Multi-objective reinforcement learning using sets of Pareto dominating policies","volume":"15","author":"Van Moffaert","year":"2014","journal-title":"J Mach Learn Res"},{"key":"2026052023581330600_btag170-B41","first-page":"5998","article-title":"Attention is all you need","volume":"30","author":"Vaswani","year":"2017","journal-title":"NeurIPS"},{"key":"2026052023581330600_btag170-B45","first-page":"106363","article-title":"Multi-objective molecular generation via clustered Pareto-based reinforcement learning","volume":"177","year":"2024","journal-title":"Neural Netw"},{"key":"2026052023581330600_btag170-B42","doi-asserted-by":"crossref","first-page":"29","DOI":"10.1109\/TEVC.2005.851275","article-title":"A faster algorithm for calculating hypervolume","volume":"10","author":"While","year":"2006","journal-title":"IEEE Trans Evol Computat"},{"key":"2026052023581330600_btag170-B43","first-page":"38592","article-title":"Geometric latent diffusion models for 3D molecule generation","volume":"202","author":"Xu","year":"2023","journal-title":"Proc ICML"},{"key":"2026052023581330600_btag170-B44","first-page":"31","article-title":"Graph convolutional policy network for goal-directed molecular graph generation","volume":"2018","author":"You","year":"2018","journal-title":"NeurIPS"},{"key":"2026052023581330600_btag170-B46","author":"Zhou","year":"2023"},{"key":"2026052023581330600_btag170-B47","doi-asserted-by":"crossref","DOI":"10.1109\/TEVC.2003.810758","article-title":"Performance assessment of multiobjective optimizers: an analysis and review","volume":"7","author":"Zitzler","year":"2003","journal-title":"IEEE Trans Evol Computat"}],"container-title":["Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bioinformatics\/advance-article-pdf\/doi\/10.1093\/bioinformatics\/btag170\/67787779\/btag170.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/5\/btag170\/67787779\/btag170.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article-pdf\/42\/5\/btag170\/67787779\/btag170.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,21]],"date-time":"2026-05-21T03:58:56Z","timestamp":1779335936000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bioinformatics\/article\/doi\/10.1093\/bioinformatics\/btag170\/8586882"}},"subtitle":[],"editor":[{"given":"Daisuke","family":"Kihara","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2026,4,4]]},"references-count":46,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2026,5,3]]}},"URL":"https:\/\/doi.org\/10.1093\/bioinformatics\/btag170","relation":{},"ISSN":["1367-4803","1367-4811"],"issn-type":[{"value":"1367-4803","type":"print"},{"value":"1367-4811","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,5]]},"published":{"date-parts":[[2026,4,4]]},"article-number":"btag170"}}