{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,1]],"date-time":"2026-08-01T18:31:25Z","timestamp":1785609085311,"version":"3.56.0"},"reference-count":39,"publisher":"Oxford University Press (OUP)","issue":"3","license":[{"start":{"date-parts":[[2026,5,11]],"date-time":"2026-05-11T00:00:00Z","timestamp":1778457600000},"content-version":"vor","delay-in-days":10,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,5,4]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Multi-target compounds, or polypharmacological agents, hold significant potential for complex diseases like cancer, where single-target therapies are often insufficient. A lack of high-quality bioactivity data limits progress in this field, especially for compounds interacting with multiple proteins simultaneously. This study introduces MT-ConBiFormer-GPT, a deep generative model designed explicitly for low-data, multi-target molecular generation, focusing on the critical PI3K\u2013AKT\u2013mTOR cancer signaling pathway. The framework integrates a variational autoencoder with a BiFormer encoder to capture long-range dependencies in SMILES strings, reducing the quadratic computational complexity associated with standard transformers and mitigating semantic discontinuities. It employs a SMILES-GPT decoder for progressive molecule generation and follows a three-phase training pipeline: unsupervised pre-training, supervised contrastive learning, and curriculum-based fine-tuning. The framework\u2019s efficacy was evaluated through a rigorous, multi-stage assessment. First, the framework was evaluated through benchmarking against state-of-the-art models, with a specialized head-to-head variant, MT-ConBiFormer-GPT_H2H, demonstrating superior performance, thereby validating its generalizability from oncology to neuropsychiatry. An internal ablation study further revealed that the full MT-ConBiFormer-GPT significantly outperformed its baseline, MT-BiFormer-GPT, in both dual- and triplet-target generation tasks, highlighting the advantages of the contrastive learning stage. Additionally, the foundational Base-BiFormer-GPT architecture, a model lacking both the contrastive and curriculum learning stages, highlighted its intrinsic robustness by achieving competitive outcomes in a distinct omics-driven design task. Docking simulations and mechanistic analyses show that the generated molecules, including high-fidelity and scaffold-hopping candidates, display more favorable binding modes than reference inhibitors. This study presents a flexible and computationally efficient framework for multi-target drug discovery in data-limited settings.<\/jats:p>","DOI":"10.1093\/bib\/bbag079","type":"journal-article","created":{"date-parts":[[2026,5,8]],"date-time":"2026-05-08T11:42:54Z","timestamp":1778240574000},"source":"Crossref","is-referenced-by-count":2,"title":["MT-ConBiFormer-GPT: multi-target molecular generation for low-data drug discovery via a contrastive BiFormer-GPT architecture and curriculum learning with cross-domain generalization"],"prefix":"10.1093","volume":"27","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-1220-4374","authenticated-orcid":false,"given":"Romina","family":"Norouzi","sequence":"first","affiliation":[{"name":"Department of Bioinformatics, Kish International Campus, University of Tehran , 16th Azar St, Kish, 1417935840 ,","place":["Iran"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2135-8864","authenticated-orcid":false,"given":"Karim","family":"Abbasi","sequence":"additional","affiliation":[{"name":"Mosaheb Institute for Mathematical Research, Kharazmi University , South Mofateh St, Tehran, 1571914911 ,","place":["Iran"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7031-4609","authenticated-orcid":false,"given":"Parvin","family":"Razzaghi","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Information Technology, Institute for Advanced Studies in Basic Sciences (IASBS) , Prof Yousef Sobouti Blvd, Zanjan, 4513766731 ,","place":["Iran"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5468-4258","authenticated-orcid":false,"given":"Sajjad","family":"Gharaghani","sequence":"additional","affiliation":[{"name":"Laboratory of Bioinformatics and Drug Design (LBD), Institute of Biochemistry and Biophysics, University of Tehran , Enghelab Square, Tehran, 131451365 ,","place":["Iran"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2026,5,11]]},"reference":[{"key":"2026051115515731300_ref1","doi-asserted-by":"publisher","first-page":"7129","DOI":"10.1073\/pnas.1820799116","article-title":"From single drug targets to synergistic network pharmacology in ischemic stroke","volume":"116","author":"Casas","year":"2019","journal-title":"Proc Natl Acad Sci U S A"},{"key":"2026051115515731300_ref2","doi-asserted-by":"publisher","first-page":"100095","DOI":"10.1016\/j.pharmr.2025.100095","article-title":"Computational drug design in the artificial intelligence era: A systematic review of molecular representations, generative architectures, and performance assessment","volume":"78","author":"Abbasi","year":"2026","journal-title":"Pharmacol Rev"},{"key":"2026051115515731300_ref3","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbab344","article-title":"Molecular design in drug discovery: A comprehensive review of deep generative models","volume":"22","author":"Cheng","year":"2021","journal-title":"Brief Bioinform"},{"key":"2026051115515731300_ref4","doi-asserted-by":"publisher","first-page":"2174","DOI":"10.1021\/acs.jcim.3c01496","article-title":"Deep generative models in de novo drug molecule generation","volume":"64","author":"Pang","year":"2023","journal-title":"J Chem Inf Model"},{"key":"2026051115515731300_ref5","doi-asserted-by":"publisher","first-page":"425","DOI":"10.1038\/s41392-021-00828-5","article-title":"Targeting PI3K\/Akt signal transduction for cancer therapy","volume":"6","author":"He","year":"2021","journal-title":"Signal Transduct Target Ther"},{"key":"2026051115515731300_ref6","doi-asserted-by":"publisher","first-page":"138","DOI":"10.1186\/s12943-023-01827-6","article-title":"PI3K\/AKT\/mTOR signaling transduction pathway and targeted therapies in cancer","volume":"22","author":"Glaviano","year":"2023","journal-title":"Mol Cancer"},{"key":"2026051115515731300_ref7","doi-asserted-by":"publisher","first-page":"1670","DOI":"10.1038\/s12276-022-00864-3","article-title":"Molecular targeted therapy for anticancer treatment","volume":"54","author":"Min","year":"2022","journal-title":"Exp Mol Med"},{"key":"2026051115515731300_ref8","doi-asserted-by":"publisher","first-page":"420","DOI":"10.1021\/acs.jmedchem.8b00760","article-title":"Polypharmacology by design: A medicinal chemist\u2019s perspective on multitargeting compounds","volume":"62","author":"Proschak","year":"2019","journal-title":"J Med Chem"},{"key":"2026051115515731300_ref9","doi-asserted-by":"publisher","first-page":"1934","DOI":"10.1093\/bioinformatics\/btt321","article-title":"Insights into polypharmacology from drug-domain associations","volume":"29","author":"Moya-Garcia","year":"2013","journal-title":"Bioinformatics"},{"key":"2026051115515731300_ref10","doi-asserted-by":"publisher","first-page":"7874","DOI":"10.1021\/jm5006463","article-title":"Polypharmacology: Challenges and opportunities in drug discovery","volume":"57","author":"Anighoro","year":"2014","journal-title":"J Med Chem"},{"key":"2026051115515731300_ref11","doi-asserted-by":"crossref","first-page":"110442","DOI":"10.1016\/j.compbiomed.2025.110442","article-title":"DFT_ANPD: A dual-feature two-sided attention network for anticancer natural products detection","volume":"194","author":"Norouzi","year":"2025","journal-title":"Comput Biol Med"},{"key":"2026051115515731300_ref12","doi-asserted-by":"publisher","first-page":"3408","DOI":"10.1038\/s41467-024-47613-w","article-title":"Prospective de novo drug design with deep interactome learning","volume":"15","author":"Atz","year":"2024","journal-title":"Nat Commun"},{"key":"2026051115515731300_ref13","article-title":"SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules","volume":"28","author":"Wolman","year":"1988","journal-title":"J Chem Inf Comput Sci"},{"key":"2026051115515731300_ref14","doi-asserted-by":"publisher","first-page":"114","DOI":"10.1038\/s41467-022-35692-6","article-title":"Leveraging molecular structure and bioactivity with chemical language models for de novo drug design","volume":"14","author":"Moret","year":"2023","journal-title":"Nat Commun"},{"key":"2026051115515731300_ref15","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbac270","article-title":"Deep generative model for therapeutic targets using transcriptomic disease-associated data-USP7 case study","volume":"23","author":"Pereira","year":"2022","journal-title":"Brief Bioinform"},{"key":"2026051115515731300_ref16","doi-asserted-by":"publisher","first-page":"1882","DOI":"10.1021\/acs.jcim.2c01301","article-title":"Gex2SGen: Designing drug-like molecules from desired gene expression signatures","volume":"63","author":"Das","year":"2023","journal-title":"J Chem Inf Model"},{"key":"2026051115515731300_ref17","doi-asserted-by":"publisher","first-page":"42","DOI":"10.1186\/s13321-023-00711-1","article-title":"MolFilterGAN: A progressively augmented generative adversarial network for triaging AI-designed molecules","volume":"15","author":"Liu","year":"2023","journal-title":"J Chem"},{"key":"2026051115515731300_ref18","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbae525","article-title":"GexMolGen: Cross-modal generation of hit-like molecules via large language model encoding of gene expression signatures","volume":"25","author":"Cheng","year":"2024","journal-title":"Brief Bioinform"},{"key":"2026051115515731300_ref19","doi-asserted-by":"publisher","first-page":"10374","DOI":"10.1021\/acs.jmedchem.4c00838","article-title":"Systematic investigation of dual-target-directed ligands","volume":"67","author":"Lembo","year":"2024","journal-title":"J Med Chem"},{"key":"2026051115515731300_ref20","doi-asserted-by":"publisher","first-page":"7946","DOI":"10.1038\/s41467-024-52060-8","article-title":"Automated design of multi-target ligands by generative deep learning","volume":"15","author":"Isigkeit","year":"2024","journal-title":"Nat Commun"},{"key":"2026051115515731300_ref21","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbab333","article-title":"De novo generation of dual-target ligands using adversarial training and reinforcement learning","volume":"22","author":"Lu","year":"2021","journal-title":"Brief Bioinform"},{"key":"2026051115515731300_ref22","doi-asserted-by":"publisher","first-page":"e1012229","DOI":"10.1371\/journal.pcbi.1012229","article-title":"MTMol-GPT: De novo multi-target molecular generation with transformer-based generative adversarial imitation learning","volume":"20","author":"Ai","year":"2024","journal-title":"PLoS Comput Biol"},{"key":"2026051115515731300_ref23","volume-title":"Proceedings of the 37th International Conference on Machine Learning","author":"Jin","year":"2020"},{"key":"2026051115515731300_ref24","doi-asserted-by":"publisher","first-page":"33","DOI":"10.1186\/s13321-018-0287-6","article-title":"Multi-objective de novo drug design with conditional graph generative model","volume":"10","author":"Li","year":"2018","journal-title":"J Chem"},{"key":"2026051115515731300_ref25","doi-asserted-by":"publisher","first-page":"10366","DOI":"10.1039\/d4sc00094c","article-title":"Structure-aware dual-target drug design through collaborative learning of pharmacophore combination and molecular simulation","volume":"15","author":"Chen","year":"2024","journal-title":"Chem Sci"},{"key":"2026051115515731300_ref26","article-title":"Reprogramming pretrained target-specific diffusion models for dual-target drug design","author":"Zhou","year":"2024"},{"key":"2026051115515731300_ref27","doi-asserted-by":"publisher","first-page":"93","DOI":"10.1186\/s13321-025-01035-y","article-title":"A 3D generation framework using diffusion model and reinforcement learning to generate multi-target compounds with desired properties","volume":"17","author":"Yuan","year":"2025","journal-title":"J Chem"},{"key":"2026051115515731300_ref28","doi-asserted-by":"crossref","first-page":"171","DOI":"10.1038\/s42256-020-0160-y","article-title":"Generative molecular design in low data regimes","volume":"2","author":"Moret","year":"2020","journal-title":"Nat Mach Intell"},{"key":"2026051115515731300_ref29","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbaf185","article-title":"Self-awareness of retrosynthesis via chemically inspired contrastive learning for reinforced molecule generation","volume":"26","author":"Zhang","year":"2025","journal-title":"Brief Bioinform"},{"key":"2026051115515731300_ref30","doi-asserted-by":"publisher","first-page":"20","DOI":"10.1186\/s13321-024-00812-5","article-title":"Reinvent 4: Modern AI-driven generative molecule design","volume":"16","author":"Loeffler","year":"2024","journal-title":"J Chem"},{"key":"2026051115515731300_ref31","doi-asserted-by":"publisher","first-page":"26","DOI":"10.1038\/s42004-025-01423-3","article-title":"fragSMILES as a chemical string notation for advanced fragment and chirality representation","volume":"8","author":"Mastrolorito","year":"2025","journal-title":"Commun Chem"},{"key":"2026051115515731300_ref32","doi-asserted-by":"publisher","DOI":"10.1093\/bib\/bbad422","article-title":"From intuition to AI: Evolution of small molecule representations in drug discovery","volume":"25","author":"McGibbon","year":"2023","journal-title":"Brief Bioinform"},{"key":"2026051115515731300_ref33","first-page":"10323","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Zhu","year":"2023"},{"key":"2026051115515731300_ref34","doi-asserted-by":"publisher","first-page":"6176","DOI":"10.1038\/s41467-024-50469-9","article-title":"Chemical language modeling with structured state space sequence models","volume":"15","author":"\u00d6z\u00e7elik","year":"2024","journal-title":"Nat Commun"},{"key":"2026051115515731300_ref35","doi-asserted-by":"publisher","first-page":"3636","DOI":"10.1038\/s41467-024-47120-y","article-title":"De novo generation of multi-target compounds using deep generative chemistry","volume":"15","author":"Munson","year":"2024","journal-title":"Nat Commun"},{"key":"2026051115515731300_ref36","doi-asserted-by":"crossref","DOI":"10.1093\/nar\/28.1.235","article-title":"The protein data bank","volume":"28","author":"Berman","year":"2000","journal-title":"Nucleic Acid Research"},{"key":"2026051115515731300_ref37","doi-asserted-by":"publisher","first-page":"455","DOI":"10.1002\/jcc.21334","article-title":"AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading","volume":"31","author":"Trott","year":"2010","journal-title":"J Comput Chem"},{"key":"2026051115515731300_ref38","doi-asserted-by":"publisher","first-page":"W530","DOI":"10.1093\/nar\/gkab294","article-title":"PLIP 2021: Expanding the scope of the protein-ligand interaction profiler to DNA and RNA","volume":"49","author":"Adasme","year":"2021","journal-title":"Nucleic Acids Res"},{"key":"2026051115515731300_ref39","volume-title":"Discovery Studio Visualizer, v21.1.0.20298","author":"Biovia","year":"2021"}],"container-title":["Briefings in Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/27\/3\/bbag079\/68268598\/bbag079.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/27\/3\/bbag079\/68268598\/bbag079.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,11]],"date-time":"2026-05-11T19:52:02Z","timestamp":1778529122000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bib\/article\/doi\/10.1093\/bib\/bbag079\/8675953"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5]]},"references-count":39,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2026,5,4]]}},"URL":"https:\/\/doi.org\/10.1093\/bib\/bbag079","relation":{},"ISSN":["1467-5463","1477-4054"],"issn-type":[{"value":"1467-5463","type":"print"},{"value":"1477-4054","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,5]]},"published":{"date-parts":[[2026,5]]},"article-number":"bbag079"}}