The DOI matches its Crossref record; 1 of the 1 references it cites were found in a registry or an index.
Learning the language of protein-protein interactions
Open accessPreprint — published as 10.1038/s41467-025-67971-3
Abstract
Protein Language Models (PLMs) trained on large databases of protein sequences have proven effective in modeling protein biology across a wide range of applications. However, while PLMs excel at capturing individual protein properties, they face challenges in natively representing protein–protein interactions (PPIs), which are crucial to understanding cellular processes and disease mechanisms. Here, we introduce MINT, a PLM specifically designed to model sets of interacting proteins in a contextual and scalable manner. Using unsupervised training on a large curated PPI dataset derived from the STRING database, MINT outperforms existing PLMs in diverse tasks relating to protein-protein interactions, including binding affinity prediction and estimation of mutational effects. Beyond these core capabilities, it excels at modeling interactions in complex protein assemblies and surpasses specialized models in antibody-antigen modeling and T cell receptor–epitope binding prediction. MINT’s predictions of mutational impacts on oncogenic PPIs align with experimental studies, and it provides reliable estimates for the potential for cross-neutralization of antibodies against SARS-CoV-2 variants of concern. These findings position MINT as a powerful tool for elucidating complex protein interactions, with significant implications for biomedical research and therapeutic discovery.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Highly accurate protein structure prediction with AlphaFold
Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03819-2 · Cited by 43,764
Flagged first, then by citations
Research works printed but not linked: 67
- 2019, Learning protein sequence embeddings using information from structure, arXiv preprint
- 10.1016/j.cels.2021.05.017 · DOI · Paper page
- 10.1073/pnas.2016239118 · DOI · Paper page
- 2021, Language models enable zero-shot prediction of the effects of mutations on protein function, Advances in neural information processing systems, 34, 29287
- 10.1038/s41586-023-06617-0 · DOI
- 10.1126/science.abd7331 · DOI · Paper page
- Wu, R. et al. High-resolution de novo structure prediction from primary sequence. BioRxiv 2022–07 (2022).
- 10.1073/pnas.2220778120 · DOI
- 10.1126/science.ade2574 · DOI · Paper page
- 10.1016/j.cels.2021.08.010 · DOI
- 2022, Peer: a comprehensive and multi-task benchmark for protein sequence understanding, Advances in Neural Information Processing Systems, 35, 35156
- 10.1038/s41598-022-13227-9 · DOI
- 10.1093/bioinformatics/btad663 · DOI
- 10.1093/bioinformatics/btac258 · DOI
- 10.1073/pnas.2405840121 · DOI · Paper page
- 2024, Large scale paired antibody language models, arXiv preprint
- Singh, R. et al. Learning the language of antibody hypervariability. bioRxiv 2023–04 (2023). · DOI
- 10.1093/NAR/GKAC1000 · DOI · Paper page
- Vaswani, A. Attention is all you need. Advances in Neural Information Processing Systems (2017).
- 10.1093/bib/bbae076 · DOI · Paper page
- 10.1093/bioinformatics/bty635 · DOI · Paper page
- Chungyoun, M. , Ruffolo, J. A. & Gray, J. J. Flab: Benchmarking deep learning methods for antibody fitness prediction. bioRxiv 2024–01 (2024). · DOI
- 10.1073/pnas.2418918121 · DOI
- 10.1093/bioinformatics/btac820 · DOI
- 2018, Bert: Pre-training of deep bidirectional transformers for language understanding, arXiv preprint
- 10.1021/acs.accounts.6b00491 · DOI · Paper page
- Omelchenko, A. A. et al. Sliding window interaction grammar (swing): a generalized interaction language model for peptide and protein interactions. bioRxiv (2024). · DOI
- Rao, R. M. et al. Msa transformer. In International Conference on Machine Learning, 8844–8856 (PMLR, 2021).
- 2020, Progen: Language modeling for protein generation, arXiv preprint
- 2021, Prottrans: towards cracking the language of life’s code through self-supervised learning, IEEE Transactions on Pattern Analysis and Machine Intelligence, 44, 7112
- Su, J. et al. Saprot: Protein language modeling with structure-aware vocabulary. bioRxiv (2023). · DOI
- Greenfield, E. A. Antibodies. A laboratory manual, second edition Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.[Google Scholar] (2014).
- 10.1371/journal.pone.0008187 · DOI
- Shanehsazzadeh, A. et al. Unlocking de novo antibody design with generative artificial intelligence. bioRxiv 2023–01 (2023).
- 10.1371/journal.pcbi.1007207 · DOI
- 10.1073/pnas.1613231114 · DOI · Paper page
- Desautels, T. , Zemla, A. , Lau, E. , Franco, M. & Faissol, D. Rapid in silico design of antibodies targeting sars-cov-2 using machine learning and supercomputing. BioRxiv 2020–04 (2020). · DOI
- 10.1073/pnas.0701000104 · DOI
- 10.1093/bioinformatics/btz184 · DOI
- 10.1016/B978-0-12-394292-0.00006-0 · DOI
- 10.1021/acs.jpcb.7b11367 · DOI
- 10.1146/annurev-immunol-082119-124838 · DOI
- Velez-Arce, A. et al. Tdc-2: Multimodal foundation for therapeutic science. bioRxiv 2024–06 (2024).
- 10.1038/s42256-023-00634-4 · DOI
- 10.1093/nar/gky1006 · DOI
- 10.1093/nar/gkx760 · DOI
- 10.1093/bioinformatics/btx286 · DOI
- 10.1093/bioinformatics/btad475 · DOI
- 10.1038/s42003-021-02610-3 · DOI
- 10.1038/s42256-023-00619-3 · DOI · Paper page
- 10.1093/bib/bbad086 · DOI
- 10.1093/bioinformatics/btab294 · DOI
- 10.1038/s42256-021-00383-2 · DOI · Paper page
- 10.1093/nar/gkx971 · DOI
- 10.1038/s41588-020-00774-y · DOI
- 10.1038/s41586-021-04043-8 · DOI · Paper page
- 10.1126/science.adg7492 · DOI · Paper page
- 10.1016/j.cell.2015.04.013 · DOI · Paper page
- 10.1038/s41467-019-11959-3 · DOI
- 10.1038/s41598-017-01913-y · DOI
- 10.1038/s41392-021-00710-4 · DOI
- 10.1093/bioinformatics/btaa739 · DOI · Paper page
- 10.1038/s41467-023-37926-7 · DOI
- 10.1038/s41586-021-04060-7 · DOI
- Hayes, T. et al. Simulating 500 million years of evolution with a language model. Science eads0018 (2025). · DOI · Paper page
- 10.1038/nbt.3988 · DOI · Paper page
- Jin, W. , Sarkizova, S. , Chen, X. , Hacohen, N. & Uhler, C. Unsupervised protein-ligand binding energy prediction via neural euler’s rotation equation. Advances in Neural Information Processing Systems 36 (2024).
Metadata
- Authors
- Ullanat, V., Jing, B., Sledzieski, S., & Berger, B.
- Affiliations
- MIT Computer Science and Artificial Intelligence Laboratory (US)
- Flatiron Institute (US)
- Corresponding author
- Berger, Bonnie
- Venue
- Nature Communications
- Publisher
- openRxiv
- Published
- 2025-03-10
- Volume
- 17
- Issue
- 1
- Pages
- 1199-1199
- ISSN
- 2041-1723 (ISSN-L)
- PMC
- PMC12859026
- OpenAlex
- W7118534871
- Cited by
- 26
- References
- 1
- Citations by year
- 2026: 22 · 2025: 4
- FWCI
- 45.47
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by-nc-nd
- creativecommons.org/licenses/by/4.0/ (vor)
- Locations
- Nature Communications (publishedVersion, cc-by-nc-nd)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-nd)
- PubMed (publishedVersion)
- PubMed Central (publishedVersion, cc-by-nc-nd)
- Europe PMC (PubMed Central) (submittedVersion, cc-by)
- PubMed Central (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- National Science Foundation (2141064)
- U.S. Department of Energy (DESC0022158)
- National Institutes of Health (2141064)
- National Institute of General Medical Sciences (R35GM141861)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
- Keywords
- protein-protein interaction, protein language models, Mentha, binding affinity prediction, mutation effects, protein assembly, SARS-CoV-2 variants, unsupervised learning, STRING database
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
