Biophysics-based protein language models for protein engineering
Open accessThe DOI matches its Crossref record; 34 of the 34 references it cites were found in a registry or an index.
Abstract
Protein language models trained on evolutionary data have emerged as powerful tools for predictive problems involving protein sequence, structure and function. However, these models overlook decades of research into biophysical factors governing protein function. We propose mutational effect transfer learning (METL), a protein language model framework that unites advanced machine learning and biophysical modeling. Using the METL framework, we pretrain transformer-based neural networks on biophysical simulation data to capture fundamental relationships between protein sequence, structure and energetics. We fine-tune METL on experimental sequence–function data to harness these biophysical signals and apply them when predicting protein properties like thermostability, catalytic activity and fluorescence. METL excels in challenging protein engineering tasks like generalizing from small training sets and position extrapolation, although existing methods that train on evolutionary signals remain powerful for many types of experimental assays. We demonstrate METL’s ability to design functional green fluorescent protein variants when trained on only 64 examples, showcasing the potential of biophysics-based protein language models for protein engineering.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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
NCBI GEO: archive for functional genomics data sets—update
Barrett, T., Wilhite, S. E., Ledoux, P., Evangelista, C., Kim, I. F., Tomashevsky, M., Marshall, K. A., Phillippy, K. H., Sherman, P. M., Holko, M., et al. (2012) · Nucleic Acids Research · doi:10.1093/nar/gks1193 · Cited by 10,418
Transformers: State-of-the-Art Natural Language Processing
Wolf, T., Debut, L., Sanh, V., Chaumond, J., Delangue, C., Moi, A., Cistac, P., Rault, T., Louf, R., Funtowicz, M., et al. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations · doi:10.18653/v1/2020.emnlp-demos.6 · Cited by 8,741
Varadi, M., Anyango, S., Deshpande, M., Nair, S., Natassia, C., Yordanova, G., Yuan, D., Stroe, O., Wood, G., Laydon, A., et al. (2022) · Nucleic Acids Research · doi:10.1093/nar/gkab1061 · Cited by 7,850
Evolutionary-scale prediction of atomic level protein structure with a language model
Lin, Z., Akin, H., Rao, R., Hie, B., Zhu, Z., Lu, W., Smetanin, N., Verkuil, R., Kabeli, O., Shmueli, Y., et al. (2022) · doi:10.1101/2022.07.20.500902 · Cited by 5,990
MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets
Steinegger, M., & Söding, J. (2017) · Nature Biotechnology · doi:10.1038/nbt.3988 · Cited by 4,840 · cited 1×
“6f452) 125 and structure-based clustering with Foldseek (v9.”
TM-align: a protein structure alignment algorithm based on the TM-score
Zhang, Y. (2005) · Nucleic Acids Research · doi:10.1093/nar/gki524 · Cited by 3,774
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
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 14×
“METL-Global represents an initial step toward a universal biophysics-based protein representation, providing comparable or better performance than a similarly sized ESM-2 model when fine-tuning with small training sets (Fig.”
InfluentialFast and accurate protein structure search with Foldseek
van Kempen, M., Kim, S. S., Tumescheit, C., Mirdita, M., Lee, J., Gilchrist, C. L. M., Söding, J., & Steinegger, M. (2022) · doi:10.1101/2022.02.07.479398 · Cited by 2,598
Fast and accurate protein structure search with Foldseek
van Kempen, M., Kim, S. S., Tumescheit, C., Mirdita, M., Lee, J., Gilchrist, C. L. M., Söding, J., & Steinegger, M. (2024) · Nature Biotechnology · doi:10.1038/s41587-023-01773-0 · Cited by 2,596 · cited 1×
“427df8a) 126 .”
Self-Attention with Relative Position Representations
Shaw, P., Uszkoreit, J., & Vaswani, A. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers) · doi:10.18653/v1/n18-2074 · Cited by 2,505 · cited 5×
“We implemented relative position embeddings as described by Shaw et al. 13 .”
InfluentialProtTrans: Toward Understanding the Language of Life Through Self-Supervised Learning
Elnaggar, A., Heinzinger, M., Dallago, C., Rehawi, G., Wang, Y., Jones, L., Gibbs, T., Feher, T., Angerer, C., Steinegger, M., et al. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2021.3095381 · Cited by 2,329
The Rosetta All-Atom Energy Function for Macromolecular Modeling and Design
Alford, R. F., Leaver-Fay, A., Jeliazkov, J. R., O’Meara, M. J., DiMaio, F. P., Park, H., Shapovalov, M. V., Renfrew, P. D., Mulligan, V. K., Kappel, K., et al. (2017) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.7b00125 · Cited by 1,532
Unified rational protein engineering with sequence-based deep representation learning
Alley, E. C., Khimulya, G., Biswas, S., AlQuraishi, M., & Church, G. M. (2019) · Nature Methods · doi:10.1038/s41592-019-0598-1 · Cited by 1,131 · cited 3×
“The pretrained METL-Local model successfully identifies known interacting positions in GB1’s dynamic β1-β2 loop region, with pairwise combinations of positions 7, 9 and 11 all ranking in the top 10% of predicted positional epistasis.”
Large language models generate functional protein sequences across diverse families
Madani, A., Krause, B., Greene, E. R., Subramanian, S., Mohr, B. P., Holton, J. M., Olmos, J. L., Xiong, C., Sun, Z. Z., Socher, R., et al. (2023) · Nature Biotechnology · doi:10.1038/s41587-022-01618-2 · Cited by 1,078 · cited 2×
“Predictive models can guide searches over the sequence–function landscape to enhance natural proteins or design new proteins 6,34,35 .”
The frontier of simulation-based inference
Cranmer, K., Brehmer, J., & Louppe, G. (2020) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1912789117 · Cited by 1,032
Simulating 500 million years of evolution with a language model
Hayes, T., Rao, R., Akin, H., Sofroniew, N. J., Oktay, D., Lin, Z., Verkuil, R., Tran, V. Q., Deaton, J., Wiggert, M., et al. (2025) · Science · doi:10.1126/science.ads0018 · Cited by 937
Disease variant prediction with deep generative models of evolutionary data
Frazer, J., Notin, P., Dias, M., Gomez, A., Min, J. K., Brock, K., Gal, Y., & Marks, D. S. (2021) · Nature · doi:10.1038/s41586-021-04043-8 · Cited by 791 · cited 18×
“In addition to using EVE as a stand-alone zero-shot method, we incorporated the EVE score into a supervised learning model.”
InfluentialMutation effects predicted from sequence co-variation
Hopf, T. A., Ingraham, J. B., Poelwijk, F. J., Schärfe, C. P. I., Springer, M., Sander, C., & Marks, D. S. (2017) · Nature Biotechnology · doi:10.1038/nbt.3769 · Cited by 775 · cited 1×
“We queried the UniRef100 database with search parameters consistent with those in EVMutation 147 : a position sequence filter of 70%, a sequence fragment filter of 50%, 100% removal of similar sequences and 80% down-weighting of similar sequences.”
Local fitness landscape of the green fluorescent protein
Sarkisyan, K. S., Bolotin, D. A., Meer, M. V., Usmanova, D. R., Mishin, A. S., Sharonov, G. V., Ivankov, D. N., Bozhanova, N. G., Baranov, M. S., Soylemez, O., et al. (2016) · Nature · doi:10.1038/nature17995 · Cited by 591 · cited 1×
“We had the genes for the 20 GFP METL designs and the 20 random baselines synthesized and cloned into an expression vector as a fusion protein with the fluorescent protein mKate2, emulating the conditions used to collect the training data 36 .”
Global analysis of protein folding using massively parallel design, synthesis, and testing
Rocklin, G. J., Chidyausiku, T. M., Goreshnik, I., Ford, A., Houliston, S., Lemak, A., Carter, L., Ravichandran, R., Mulligan, V. K., Chevalier, A., et al. (2017) · Science · doi:10.1126/science.aan0693 · Cited by 501 · cited 1×
“13) 11 to compute full-atom energy terms (ref2015 score function), centroid-atom energy terms (score3 score function) and custom filter terms based on Rocklin et al. 129 .”
Learning the protein language: Evolution, structure, and function
Bepler, T., & Berger, B. (2021) · Cell Systems · doi:10.1016/j.cels.2021.05.017 · Cited by 483 · cited 2×
“These models learn meaningful, low-dimensional representations that capture the semantic organization of protein space and have broad utility in protein engineering 2 .”
Low-N protein engineering with data-efficient deep learning
Biswas, S., Khimulya, G., Alley, E. C., Esvelt, K. M., & Church, G. M. (2021) · Nature Methods · doi:10.1038/s41592-021-01100-y · Cited by 422 · cited 2×
“PLMs are an important component in many existing methods for low-N protein engineering 3,18,99–107 .”
High-resolution de novo structure prediction from primary sequence
Wu, R., Ding, F., Wang, R., Shen, R., Zhang, X., Luo, S., Su, C., Wu, Z., Xie, Q., Berger, B., et al. (2022) · doi:10.1101/2022.07.21.500999 · Cited by 392 · cited 1×
“Although sequence-based PLMs can learn about protein structure from evolutionary statistics 19,57–59 , many recent PLMs directly incorporate structural information 60–64 , and we envision METL-Global would continue to use this prior knowledge when it is available.”
A statistical framework for analyzing deep mutational scanning data
Rubin, A. F., Gelman, H., Lucas, N., Bajjalieh, S. M., Papenfuss, A. T., Speed, T. P., & Fowler, D. M. (2017) · Genome Biology · doi:10.1186/s13059-017-1272-5 · Cited by 262
Transfer learning to leverage larger datasets for improved prediction of protein stability changes
Dieckhaus, H., Brocidiacono, M., Randolph, N. Z., & Kuhlman, B. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2314853121 · Cited by 205
Mighell, T. L., Evans-Dutson, S., & O’Roak, B. J. (2018) · The American Journal of Human Genetics · doi:10.1016/j.ajhg.2018.03.018 · Cited by 204
Activity-enhancing mutations in an E3 ubiquitin ligase identified by high-throughput mutagenesis
Starita, L. M., Pruneda, J. N., Lo, R. S., Fowler, D. M., Kim, H. J., Hiatt, J. B., Shendure, J., Brzovic, P. S., Fields, S., & Klevit, R. E. (2013) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1303309110 · Cited by 183
Rapid protein stability prediction using deep learning representations
Blaabjerg, L. M., Kassem, M. M., Good, L. L., Jonsson, N., Cagiada, M., Johansson, K. E., Boomsma, W., Stein, A., & Lindorff-Larsen, K. (2023) · eLife · doi:10.7554/elife.82593 · Cited by 177
Protein language models learn evolutionary statistics of interacting sequence motifs
Zhang, Z., Wayment-Steele, H. K., Brixi, G., Wang, H., Kern, D., & Ovchinnikov, S. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2406285121 · Cited by 140
Convolutions are competitive with transformers for protein sequence pretraining
Yang, K. K., Fusi, N., & Lu, A. X. (2024) · Cell Systems · doi:10.1016/j.cels.2024.01.008 · Cited by 137
Protein language models are biased by unequal sequence sampling across the tree of life
Ding, F., & Steinhardt, J. (2024) · doi:10.1101/2024.03.07.584001 · Cited by 57 · cited 1×
“However, the precise selective pressures for each protein are different and largely unknown, and evolutionary patterns can be confounded by historical events, phylogenetic biases and unequal sequence sampling 46 .”
Matreyek, K. A., Stephany, J. J., Ahler, E., & Fowler, D. M. (2021) · Genome Medicine · doi:10.1186/s13073-021-00984-x · Cited by 48
An end-to-end framework for the prediction of protein structure and fitness from single sequence
Chen, Y., Xu, Y., Liu, D., Xing, Y., & Gong, H. (2024) · Nature Communications · doi:10.1038/s41467-024-51776-x · Cited by 45
Cited by
Democratizing protein language model training, sharing and collaboration
Su, J., Li, Z., Tao, T., Han, C., He, Y., Dai, F., Yuan, Q., Gao, Y., Si, T., Zhang, X., et al. (2026) · Nature Biotechnology · doi:10.1038/s41587-025-02859-7 · Cited by 35
Self-citationErtelt, M., Moretti, R., Meiler, J., & Schoeder, C. T. (2025) · Science Advances · doi:10.1126/sciadv.adr7338 · Cited by 15
Investigating the determinants of performance in machine learning for protein fitness prediction
Sandhu, M., Mater, A. C., Matthews, D. S., Spence, M. A., Lenskiy, A. A., & Jackson, C. (2020) · doi:10.1101/2020.09.30.319780 · Cited by 8
Designing diverse and high-performance proteins with a large language model in the loop
Gomez-Uribe, C. A., Gado, J., & Islamov, M. (2025) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1013119 · Cited by 6
Investigating the determinants of performance in machine learning for protein fitness prediction
(2025) · doi:10.1002/pro.70235 · Cited by 6
Metadata
- Authors
- Gelman, S., Johnson, B., Freschlin, C. R., Sharma, A., D’Costa, S., Peters, J., Gitter, A., & Romero, P. A.
- Venue
- Nature Methods
- Published
- 2025-09
- Volume
- 22
- Issue
- 9
- Pages
- 1868-1879
- ISSN
- 1548-7091 (ISSN-L)
- 1548-7105
- OpenAlex
- W4414131808
- Cited by
- 69
- References
- 34
- FWCI
- 16.10
- Type
- journal article
- Language
- en
- Open access
- hybrid
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.