{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,8]],"date-time":"2026-05-08T03:53:36Z","timestamp":1778212416348,"version":"3.51.4"},"reference-count":82,"publisher":"Oxford University Press (OUP)","issue":"6","license":[{"start":{"date-parts":[[2025,12,8]],"date-time":"2025-12-08T00:00:00Z","timestamp":1765152000000},"content-version":"vor","delay-in-days":37,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,11,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Multimodal molecular representation learning, which jointly models molecular graphs and their textual descriptions, enhances predictive accuracy and interpretability by enabling more robust and reliable predictions of drug toxicity, bioactivity, and physicochemical properties through the integration of structural and semantic information. However, existing multimodal methods suffer from two key limitations: (i) they typically perform cross-modal interaction only at the final encoder layer, thus overlooking hierarchical semantic dependencies; (ii) they lack a unified prototype space for robust alignment between modalities. To address these limitations, we propose ProtoMol, a prototype-guided multimodal framework that enables fine-grained integration and consistent semantic alignment between molecular graphs and textual descriptions. ProtoMol incorporates dual-branch hierarchical encoders, utilizing Graph Neural Networks to process structured molecular graphs and Transformers to encode unstructured texts, resulting in comprehensive layer-wise representations. Then, ProtoMol introduces a layer-wise bidirectional cross-modal attention mechanism that progressively aligns semantic features across layers. Furthermore, a shared prototype space with learnable, class-specific anchors is constructed to guide both modalities toward coherent and discriminative representations. Extensive experiments on multiple benchmark datasets demonstrate that ProtoMol consistently outperforms state-of-the-art baselines across a variety of molecular property prediction tasks. Our source code is available at: https:\/\/github.com\/zky04\/Protomol.<\/jats:p>","DOI":"10.1093\/bib\/bbaf629","type":"journal-article","created":{"date-parts":[[2025,11,17]],"date-time":"2025-11-17T12:32:26Z","timestamp":1763382746000},"source":"Crossref","is-referenced-by-count":5,"title":["ProtoMol: enhancing molecular property prediction via prototype-guided multimodal learning"],"prefix":"10.1093","volume":"26","author":[{"given":"Yingxu","family":"Wang","sequence":"first","affiliation":[{"name":"Department of Machine Learning , Mohamed bin Zayed University of Artificial Intelligence, AI Diyafah Street, 7909 Abu Dhabi,","place":["United Arab Emirates"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kunyu","family":"Zhang","sequence":"additional","affiliation":[{"name":"International College , Zhengzhou University, Daxue North Road, 450000 Henan,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jiaxin","family":"Huang","sequence":"additional","affiliation":[{"name":"Department of Machine Learning , Mohamed bin Zayed University of Artificial Intelligence, AI Diyafah Street, 7909 Abu Dhabi,","place":["United Arab Emirates"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Nan","family":"Yin","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering , Hong Kong University of Science and Technology, 999077, Hong Kong,","place":["China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Siwei","family":"Liu","sequence":"additional","affiliation":[{"name":"School of Natural and Computing Science , University of Aberdeen, 32 Elphinstone Road, AB24 3EU Scotland,","place":["United Kingdom"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Eran","family":"Segal","sequence":"additional","affiliation":[{"name":"Department of Machine Learning , Mohamed bin Zayed University of Artificial Intelligence, AI Diyafah Street, 7909 Abu Dhabi,","place":["United Arab Emirates"]},{"name":"Department of Molecular Cell Biology , Weizmann Institute of Science, 7610001, Rehovot,","place":["Israel"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2025,12,8]]},"reference":[{"key":"2025120812170545400_ref1","first-page":"513","article-title":"MoleculeNet: a benchmark for molecular machine learning","volume":"9","author":"Zhenqin","year":"2017","journal-title":"Chem Sci"},{"key":"2025120812170545400_ref2","doi-asserted-by":"crossref","first-page":"bbad422","DOI":"10.1093\/bib\/bbad422","article-title":"From intuition to AI: evolution of small molecule representations in drug discovery","volume":"25","author":"McGibbon","year":"2024","journal-title":"Brief Bioinform"},{"key":"2025120812170545400_ref3","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1016\/S0169-409X(96)00423-1","article-title":"Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings","volume":"23","author":"Lipinski","year":"1997","journal-title":"Adv Drug Deliv Rev"},{"key":"2025120812170545400_ref4","doi-asserted-by":"publisher","first-page":"21234","DOI":"10.1609\/aaai.v39i20.35422","article-title":"Bridging molecular graphs and large language models","volume":"39","author":"Wang","year":"2025","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"2025120812170545400_ref5"},{"key":"2025120812170545400_ref6","doi-asserted-by":"publisher","first-page":"btae574","DOI":"10.1093\/bioinformatics\/btae574","article-title":"Chain-aware graph neural networks for molecular property prediction","volume":"40","author":"Wang","year":"2024","journal-title":"Bioinformatics"},{"key":"2025120812170545400_ref7","journal-title":"Proceedings of the 29th International Conference on Neural Information Processing Systems-Volume 2"},{"key":"2025120812170545400_ref8","doi-asserted-by":"publisher","first-page":"31","DOI":"10.1021\/ci00057a005","article-title":"Smiles, a chemical language and information system. 1. Introduction to methodology and encoding rules","volume":"28","author":"Weininger","year":"1988","journal-title":"J Chem Inf Comput Sci"},{"key":"2025120812170545400_ref9","doi-asserted-by":"publisher","first-page":"129","DOI":"10.1186\/s13321-024-00916-y","article-title":"A systematic review of deep learning chemical language models in recent era","volume":"16","author":"Flores-Hernandez","year":"2024","journal-title":"J Chem"},{"key":"2025120812170545400_ref10"},{"key":"2025120812170545400_ref11","volume-title":"The Eleventh International Conference on Learning Representations"},{"key":"2025120812170545400_ref12","doi-asserted-by":"publisher","first-page":"558","DOI":"10.1038\/s42256-024-00837-3","article-title":"Predicting equilibrium distributions for molecular systems with deep learning","volume":"6","author":"Zheng","year":"2024","journal-title":"Nat Mach Intell"},{"key":"2025120812170545400_ref13","doi-asserted-by":"publisher","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 Central Sci"},{"key":"2025120812170545400_ref14"},{"key":"2025120812170545400_ref15","doi-asserted-by":"publisher","first-page":"225","DOI":"10.1186\/s12859-024-05847-x","article-title":"Can large language models understand molecules?","volume":"25","author":"Sadeghi","year":"2024","journal-title":"BMC Bioinform"},{"key":"2025120812170545400_ref16","doi-asserted-by":"publisher","first-page":"1447","DOI":"10.1038\/s42256-023-00759-6","article-title":"Multi-modal molecule structure\u2013text model for text-based retrieval and editing","volume":"5","author":"Liu","year":"2023","journal-title":"Nat Mach Intell"},{"key":"2025120812170545400_ref17","first-page":"12559","article-title":"Self-supervised graph transformer on large-scale molecular data","volume":"33","author":"Rong","year":"2020","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref18","volume-title":"Pacific-Asia Conference on Knowledge Discovery and Data Mining"},{"key":"2025120812170545400_ref19","doi-asserted-by":"publisher","first-page":"179","DOI":"10.1207\/s15516709cog1402_1","article-title":"Finding structure in time","volume":"14","author":"Elman","year":"1990","journal-title":"Cognit Sci"},{"key":"2025120812170545400_ref20","doi-asserted-by":"publisher","first-page":"2278","DOI":"10.1109\/5.726791","article-title":"Gradient-based learning applied to document recognition","volume":"86","author":"LeCun","year":"2002","journal-title":"Proc IEEE"},{"key":"2025120812170545400_ref21"},{"key":"2025120812170545400_ref22","doi-asserted-by":"publisher","first-page":"1197","DOI":"10.1038\/s41467-024-45102-8","article-title":"Difficulty in chirality recognition for transformer architectures learning chemical structures from string representations","volume":"15","author":"Yoshikai","year":"2024","journal-title":"Nat Commun"},{"key":"2025120812170545400_ref23","first-page":"59662","article-title":"What can large language models do in chemistry? A comprehensive benchmark on eight tasks","volume":"36","author":"Guo","year":"2023","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref24","doi-asserted-by":"publisher","volume":"15","journal-title":"J Cheminform","DOI":"10.1186\/s13321-023-00725-9"},{"key":"2025120812170545400_ref25","doi-asserted-by":"crossref","first-page":"btaf061","DOI":"10.1093\/bioinformatics\/btaf061","article-title":"MolFCL: predicting molecular properties through chemistry-guided contrastive and prompt learning","volume":"41","author":"Tang","year":"2025","journal-title":"Bioinformatics"},{"key":"2025120812170545400_ref26","doi-asserted-by":"publisher","first-page":"742","DOI":"10.1021\/ci100050t","article-title":"Extended-connectivity fingerprints","volume":"50","author":"Rogers","year":"2010","journal-title":"J Chem Inf Model"},{"key":"2025120812170545400_ref27","article-title":"Chemical language models have problems with chemistry: a case study on molecule captioning task","volume":"2024","author":"Ganeeva","year":"2024","journal-title":"The Second Tiny Papers Track at ICLR"},{"key":"2025120812170545400_ref28","doi-asserted-by":"crossref","first-page":"045024","DOI":"10.1088\/2632-2153\/aba947","article-title":"Self-referencing embedded strings (selfies): a 100% robust molecular string representation","volume":"1","author":"Krenn","year":"2020","journal-title":"Mach Learn Sci Technol"},{"key":"2025120812170545400_ref29","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btae304"},{"key":"2025120812170545400_ref30","doi-asserted-by":"publisher","first-page":"754","DOI":"10.1038\/s42256-024-00856-0","article-title":"Molecular set representation learning","volume":"6","author":"Boulougouri","year":"2024","journal-title":"Nat Mach Intell"},{"key":"2025120812170545400_ref31","volume-title":"Machine learning meets quantum physics"},{"key":"2025120812170545400_ref32","doi-asserted-by":"publisher","first-page":"241722","DOI":"10.1063\/1.5019779","article-title":"SchNet\u2013a deep learning architecture for molecules and materials","volume":"148","author":"Sch\u00fctt","year":"2018","journal-title":"J Chem Phys"},{"key":"2025120812170545400_ref33","doi-asserted-by":"publisher","first-page":"btae342","DOI":"10.1093\/bioinformatics\/btae342","article-title":"Representations of lipid nanoparticles using large language models for transfection efficiency prediction","volume":"40","author":"Moayedpour","year":"2024","journal-title":"Bioinformatics"},{"key":"2025120812170545400_ref34","doi-asserted-by":"publisher","first-page":"btad258","DOI":"10.1093\/bioinformatics\/btad258","article-title":"Molecular property prediction by contrastive learning with attention-guided positive sample selection","volume":"39","author":"Wang","year":"2023","journal-title":"Bioinformatics"},{"key":"2025120812170545400_ref35","doi-asserted-by":"publisher","first-page":"bbae565","DOI":"10.1093\/bib\/bbae565","article-title":"Prototype-based contrastive substructure identification for molecular property prediction","volume":"25","author":"He","year":"2024","journal-title":"Brief Bioinform"},{"key":"2025120812170545400_ref36","doi-asserted-by":"crossref","first-page":"btae118","DOI":"10.1093\/bioinformatics\/btae118","article-title":"MolFeSCue: enhancing molecular property prediction in data-limited and imbalanced contexts using few-shot and contrastive learning","volume":"40","author":"Zhang","year":"2024","journal-title":"Bioinformatics"},{"key":"2025120812170545400_ref37"},{"key":"2025120812170545400_ref38","doi-asserted-by":"publisher","first-page":"bbae394","DOI":"10.1093\/bib\/bbae394","article-title":"Attribute-guided prototype network for few-shot molecular property prediction","volume":"25","author":"Hou","year":"2024","journal-title":"Brief Bioinform"},{"key":"2025120812170545400_ref39","first-page":"76737","article-title":"Interpretable prototype-based graph information bottleneck","volume":"36","author":"Seo","year":"2023","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref40"},{"key":"2025120812170545400_ref41","doi-asserted-by":"publisher","first-page":"bbaf251","DOI":"10.1093\/bib\/bbaf251","article-title":"Integrated multimodal hierarchical fusion and meta-learning for enhanced molecular property prediction","volume":"26","author":"Han","year":"2025","journal-title":"Brief Bioinform"},{"key":"2025120812170545400_ref42"},{"key":"2025120812170545400_ref43","doi-asserted-by":"crossref","first-page":"bbae298","DOI":"10.1093\/bib\/bbae298","article-title":"MvMRL: a multi-view molecular representation learning method for molecular property prediction","volume":"25","author":"Zhang","year":"2024","journal-title":"Brief Bioinform"},{"key":"2025120812170545400_ref44","doi-asserted-by":"publisher","first-page":"107","DOI":"10.1021\/c160017a018","article-title":"The generation of a unique machine description for chemical structures-a technique developed at chemical abstracts service","volume":"5","author":"Morgan","year":"1965","journal-title":"J Chem Doc"},{"key":"2025120812170545400_ref45","doi-asserted-by":"crossref","first-page":"56","DOI":"10.1186\/s13321-024-00846-9","article-title":"MolPROP: molecular property prediction with multimodal language and graph fusion","volume":"16","author":"Rollins","year":"2024","journal-title":"J Cheminf"},{"key":"2025120812170545400_ref46"},{"key":"2025120812170545400_ref47"},{"key":"2025120812170545400_ref48","first-page":"22099","article-title":"Connecting multi-modal contrastive representations","volume":"36","author":"Wang","year":"2023","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref49","journal-title":"Proceedings of the 27th International Conference on Neural Information Processing Systems-Volume 2"},{"key":"2025120812170545400_ref50","doi-asserted-by":"publisher","first-page":"67","DOI":"10.1186\/s13321-023-00733-9","article-title":"Explaining compound activity predictions with a substructure-aware loss for graph neural networks","volume":"15","author":"Amara","year":"2023","journal-title":"J Chem"},{"key":"2025120812170545400_ref51","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":"2025120812170545400_ref52","doi-asserted-by":"publisher","first-page":"26183","DOI":"10.1609\/aaai.v39i24.34815","article-title":"Math-puma: progressive upward multimodal alignment to enhance mathematical reasoning","volume":"39","author":"Zhuang","year":"2025","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"2025120812170545400_ref53","first-page":"32111","article-title":"Dual prototype evolving for test-time generalization of vision-language models","volume":"37","author":"Zhang","year":"2024","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref54","article-title":"Prototypical contrastive learning of unsupervised representations","author":"Li","year":"2020"},{"key":"2025120812170545400_ref55"},{"key":"2025120812170545400_ref56"},{"key":"2025120812170545400_ref57"},{"key":"2025120812170545400_ref58"},{"key":"2025120812170545400_ref59","first-page":"100","article-title":"Algorithm as 136: a k-means clustering algorithm","volume":"28","author":"Hartigan","year":"1979","journal-title":"J R Stat Soc Ser C Appl Stat"},{"key":"2025120812170545400_ref60"},{"key":"2025120812170545400_ref61","volume-title":"Proceedings of the 10th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics"},{"key":"2025120812170545400_ref62"},{"key":"2025120812170545400_ref63"},{"key":"2025120812170545400_ref64"},{"key":"2025120812170545400_ref65"},{"key":"2025120812170545400_ref66"},{"key":"2025120812170545400_ref67"},{"key":"2025120812170545400_ref68"},{"key":"2025120812170545400_ref69","first-page":"15870","article-title":"Motif-based graph self-supervised learning for molecular property prediction","volume":"34","author":"Zhang","year":"2021","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref70","first-page":"17584","article-title":"Fragment-based pretraining and finetuning on molecular graphs","volume":"36","author":"Luong","year":"2023","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref71","first-page":"25854","article-title":"Rethinking tokenizer and decoder in masked graph modeling for molecules","volume":"36","author":"Liu","year":"2023","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025120812170545400_ref72"},{"key":"2025120812170545400_ref73","volume-title":"The Eleventh International Conference on Learning Representations"},{"key":"2025120812170545400_ref74"},{"key":"2025120812170545400_ref75","doi-asserted-by":"crossref","first-page":"1666","DOI":"10.1016\/j.csbj.2024.04.030","article-title":"Multimodal fused deep learning for drug property prediction: integrating chemical language and molecular graph","volume":"23","author":"Xiaohua","year":"2024","journal-title":"Comput Struct Biotechnol J"},{"key":"2025120812170545400_ref76","doi-asserted-by":"publisher","first-page":"3044","DOI":"10.1109\/TCBB.2023.3253862","article-title":"Molecular joint representation learning via multi-modal information of smiles and graphs","volume":"20","author":"Tianyu","year":"2023","journal-title":"IEEE\/ACM Trans Comput Biol Bioinform"},{"key":"2025120812170545400_ref77","doi-asserted-by":"publisher","first-page":"111463","DOI":"10.1016\/j.patcog.2025.111463","article-title":"MDFCL: multimodal data fusion-based graph contrastive learning framework for molecular property prediction","volume":"163","author":"Gong","year":"2025","journal-title":"Pattern Recognit"},{"key":"2025120812170545400_ref78","doi-asserted-by":"publisher","first-page":"1327","DOI":"10.3233\/IDA-163129","article-title":"Davies bouldin index based hierarchical initialization k-means","volume":"21","author":"Xiao","year":"2017","journal-title":"Intell Data Anal"},{"key":"2025120812170545400_ref79"},{"key":"2025120812170545400_ref80"},{"key":"2025120812170545400_ref81"},{"key":"2025120812170545400_ref82","volume-title":"Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies"}],"container-title":["Briefings in Bioinformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/26\/6\/bbaf629\/65802342\/bbaf629.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bib\/article-pdf\/26\/6\/bbaf629\/65802342\/bbaf629.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,8]],"date-time":"2025-12-08T17:17:17Z","timestamp":1765214237000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bib\/article\/doi\/10.1093\/bib\/bbaf629\/8374029"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,1]]},"references-count":82,"journal-issue":{"issue":"6","published-print":{"date-parts":[[2025,11,1]]}},"URL":"https:\/\/doi.org\/10.1093\/bib\/bbaf629","relation":{},"ISSN":["1467-5463","1477-4054"],"issn-type":[{"value":"1467-5463","type":"print"},{"value":"1477-4054","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2025,11]]},"published":{"date-parts":[[2025,11,1]]},"article-number":"bbaf629"}}