ProtoMol: enhancing molecular property prediction via prototype-guided multimodal learning
Briefings in Bioinformatics · 2025
DOI: 10.1093/bib/bbaf629 ↗Kurz gefasst
Fachartikel in der Zeitschrift Briefings in Bioinformatics, 2025. Verfasst von Yingxu Wang und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Machine Learning, Mohamed bin Zayed University of Artificial Intelligence, AI Diyafah Street, 7909 Abu Dhabi, United Arab Emirates. Hauptthemen laut MEDLINE: Neural Networks, Computer, Machine Learning, Software, Computational Biology. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 82 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
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.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Machine Learning, Mohamed bin Zayed University of Artificial Intelligence, AI Diyafah Street, 7909 Abu Dhabi, United Arab Emirates.
- International College, Zhengzhou University, Daxue North Road, 450000 Henan, China.
- Department of Computer Science and Engineering, Hong Kong University of Science and Technology, 999077, Hong Kong, China.
- School of Natural and Computing Science, University of Aberdeen, 32 Elphinstone Road, AB24 3EU Scotland, United Kingdom.
- Department of Molecular Cell Biology, Weizmann Institute of Science, 7610001, Rehovot, Israel.
Zitieren
APA
Wang, Y., Zhang, K., Huang, J., Yin, N., Liu, S., & Segal, E. (2025). ProtoMol: enhancing molecular property prediction via prototype-guided multimodal learning. Briefings in Bioinformatics. https://doi.org/10.1093/bib/bbaf629
Vancouver
Wang Y, Zhang K, Huang J, Yin N, Liu S, Segal E. ProtoMol: enhancing molecular property prediction via prototype-guided multimodal learning. Briefings in Bioinformatics. 2025. doi:10.1093/bib/bbaf629
BibTeX
@article{Wang2025,
author = {Wang, Yingxu and Zhang, Kunyu and Huang, Jiaxin and Yin, Nan and Liu, Siwei and Segal, Eran},
title = {ProtoMol: enhancing molecular property prediction via prototype-guided multimodal learning},
journal = {Briefings in Bioinformatics},
year = {2025},
doi = {10.1093/bib/bbaf629}
}
EUVIMED