Protein language models trained on multiple sequence alignments learn phylogenetic relationships
Open accessPublished version of a preprint
Abstract
Self-supervised neural language models with attention have recently been applied to biological sequence data, advancing structure, function and mutational effect prediction. Some protein language models, including MSA Transformer and AlphaFold’s EvoFormer, take multiple sequence alignments (MSAs) of evolutionarily related proteins as inputs. Simple combinations of MSA Transformer’s row attentions have led to state-of-the-art unsupervised structural contact prediction. We demonstrate that similarly simple, and universal, combinations of MSA Transformer’s column attentions strongly correlate with Hamming distances between sequences in MSAs. Therefore, MSA-based language models encode detailed phylogenetic relationships. We further show that these models can separate coevolutionary signals encoding functional and structural constraints from phylogenetic correlations reflecting historical contingency. To assess this, we generate synthetic MSAs, either without or with phylogeny, from Potts models trained on natural MSAs. We find that unsupervised contact prediction is substantially more resilient to phylogenetic noise when using MSA Transformer versus inferred Potts models.
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
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.
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 · see citing works
Rives, A., Meier, J., Sercu, T., Goyal, S., Lin, Z., Liu, J., Guo, D., Ott, M., Zitnick, C. L., Ma, J., et al. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2016239118 · Cited by 3,286 · see citing works
Rives, A., Meier, J., Sercu, T., Goyal, S., Lin, Z., Liu, J., Guo, D., Ott, M., Zitnick, C. L., Ma, J., et al. (2019) · doi:10.1101/622803 · Cited by 3,280 · cited 1× · see citing works
“To these obtained embeddings, the model adds two kinds of learned [6] scalar positional encodings [53], designed to allow the model to distinguish between (a) different aligned positions (columns), and (b) between different sequence positions (rows).”
MethodConvolutional Sequence to Sequence Learning
Gehring, J., Auli, M., Grangier, D., Yarats, D., & Dauphin, Y. N. (2017) · arXiv · doi:10.48550/arxiv.1705.03122 · Cited by 2,850 · cited 1× · see citing works
“To these obtained embeddings, the model adds two kinds of learned [6] scalar positional encodings [50], designed to allow the model to distinguish between (a) different aligned positions (columns), and (b) between different sequence positions (rows).”
MethodA primer in BERTology: What we know about how BERT works
Rogers, A., Kovaleva, O., & Rumshisky, A. (2020) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00349 · Cited by 1,552 · see citing works
Morcos, F., Pagnani, A., Lunt, B., Bertolino, A., Marks, D. S., Sander, C., Zecchina, R., Onuchic, J. N., Hwa, T., & Weigt, M. (2011) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1111471108 · Cited by 1,434 · see citing works
Protein 3D Structure Computed from Evolutionary Sequence Variation
Marks, D. S., Colwell, L. J., Sheridan, R., Hopf, T. A., Pagnani, A., Zecchina, R., & Sander, C. (2011) · PLoS ONE · doi:10.1371/journal.pone.0028766 · Cited by 1,144 · see citing works
Rethinking Attention with Performers
Choromanski, K., Likhosherstov, V., Dohan, D., Song, X., Gane, A., Sarlos, T., Hawkins, P., Davis, J., Mohiuddin, A., Kaiser, L., et al. (2020) · arXiv · doi:10.48550/arxiv.2009.14794 · Cited by 695 · see citing works
Single-sequence protein structure prediction using language models from deep learning
Chowdhury, R., Bouatta, N., Biswas, S., Rochereau, C., Church, G. M., Sorger, P. K., & AlQuraishi, M. (2021) · doi:10.1101/2021.08.02.454840 · Cited by 455 · see citing works
Single-sequence protein structure prediction using a language model and deep learning
Chowdhury, R., Bouatta, N., Biswas, S., Floristean, C., Kharkar, A., Roy, K., Rochereau, C., Ahdritz, G., Zhang, J., Church, G. M., et al. (2022) · Nature Biotechnology · doi:10.1038/s41587-022-01432-w · Cited by 455 · see citing works
Axial Attention in Multidimensional Transformers
Ho, J., Kalchbrenner, N., Weissenborn, D., & Salimans, T. (2019) · arXiv · doi:10.48550/arxiv.1912.12180 · Cited by 370 · see citing works
Transformer protein language models are unsupervised structure learners
Rao, R., Meier, J., Sercu, T., Ovchinnikov, S., & Rives, A. (2020) · doi:10.1101/2020.12.15.422761 · Cited by 341 · see citing works
An evolution-based model for designing chorismate mutase enzymes
Russ, W. P., Figliuzzi, M., Stocker, C., Barrat-Charlaix, P., Socolich, M., Kast, P., Hilvert, D., Monasson, R., Cocco, S., Weigt, M., et al. (2020) · Science · doi:10.1126/science.aba3304 · Cited by 313 · see citing works
Correlated mutations in models of protein sequences: phylogenetic and structural effects
Lapedes, A. S., Giraud, B., Liu, L., & Stormo, G. D. (1999) · Institute of Mathematical Statistics Lecture Notes - Monograph Series · doi:10.1214/lnms/1215455556 · Cited by 124 · see citing works
Flagged first, then by citations
- Authors
- Lupo, U., Sgarbossa, D., & Bitbol, A.-F.
- Affiliations
- SIB Swiss Institute of Bioinformatics (CH)
- Corresponding author
- Lupo, Umberto
- Bitbol, Anne-Florence
- Venue
- Nature Communications
- Published
- 2022-10-22
- Volume
- 13
- Issue
- 1
- Pages
- 6298-6298
- ISSN
- 2041-1723 (ISSN-L)
- PMC
- PMC9588007
- OpenAlex
- W4307068738
- Cited by
- 67
- References
- 14
- Citations by year
- 2026: 19 · 2025: 21 · 2024: 16 · 2023: 9 · 2022: 2
- FWCI
- 3.92
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- Nature Communications (publishedVersion, cc-by)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Infoscience (Ecole Polytechnique Fédérale de Lausanne) (submittedVersion, cc-by-nc-nd)
- arXiv (Cornell University) (cc-by)
- Funders
- European Commission (851173, HORIZON2020)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Machine Learning in Bioinformatics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
- Keywords
- protein language models, multiple sequence alignment, phylogenetic relationships, Hamming distance, coevolution, Potts model, protein contact prediction, historical contingency
- MeSH terms
- Algorithms, Language, Phylogeny, Proteins, Sequence Alignment
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
