A tri-modal contrastive learning framework for protein representation learning
Zhang, Li Guo, Han Schaffer, Leah Ko, Young Su Singh, Digvijay Rahmani, Hamid Grotjahn, Danielle Villa, Elizabeth Gilson, Michael Wang, Wei
Zhang, Li
Guo, Han
Schaffer, Leah
Ko, Young Su
Singh, Digvijay
Rahmani, Hamid
Grotjahn, Danielle
Villa, Elizabeth
Gilson, Michael
Wang, Wei
Supervisor
Department
Machine Learning
Embargo End Date
Type
Journal article
Date
License
Language
English
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
Protein foundation models, particularly protein language models, have shown strong success in learning meaningful protein representations using transformer architectures pretrained on large-scale datasets through self-supervised learning. These representations have proven effective for downstream tasks such as predicting protein functions and properties. However, most existing models focus solely on amino acid sequences, overlooking other informative modalities such as 3D structures and literature text. While some recent efforts incorporate multiple modalities, they often suffer from limitations in modality coverage or training strategy. To address this gap, we propose a multimodal pretraining framework that integrates three complementary modalities-protein sequences, structures, and literature text. Our method uses the sequence modality as an anchor and aligns the other two modalities to it via contrastive learning, enabling the model to capture richer and more holistic protein representations. Across a diverse set of downstream tasks, ProteinAligner outperforms state-of-the-art foundation models in predicting protein functions and properties.
Citation
L. Zhang, H. Guo, L. Schaffer, Y.S. Ko, D. Singh, H. Rahmani , et al., "A tri-modal contrastive learning framework for protein representation learning," Cell Reports Methods, pp. 101407-101407, 2026, https://doi.org/10.1016/j.crmeth.2026.101407.
Source
Cell Reports Methods
Conference
Keywords
31 Biological Sciences, 3102 Bioinformatics and Computational Biology, 46 Information and Computing Sciences
Subjects
Source
Publisher
Elsevier