Neural network extrapolation to distant regions of the protein fitness landscape
Open accessThe DOI matches its Crossref record; 12 of the 12 references it cites were found in a registry or an index.
Abstract
Machine learning (ML) has transformed protein engineering by constructing models of the underlying sequence-function landscape to accelerate the discovery of new biomolecules. ML-guided protein design requires models, trained on local sequence-function information, to accurately predict distant fitness peaks. In this work, we evaluate neural networks’ capacity to extrapolate beyond their training data. We perform model-guided design using a panel of neural network architectures trained on protein G (GB1)-Immunoglobulin G (IgG) binding data and experimentally test thousands of GB1 designs to systematically evaluate the models’ extrapolation. We find each model architecture infers markedly different landscapes from the same data, which give rise to unique design preferences. We find simpler models excel in local extrapolation to design high fitness proteins, while more sophisticated convolutional models can venture deep into sequence space to design proteins that fold but are no longer functional. We also find that implementing a simple ensemble of convolutional neural networks enables robust design of high-performing variants in the local landscape. Our findings highlight how each architecture’s inductive biases prime them to learn different aspects of the protein fitness landscape and how a simple ensembling approach makes protein engineering more robust. Machine learning accelerates protein engineering by predicting sequence-function relationships. Here, authors evaluate neural network architectures’ ability to extrapolate beyond training data, finding simpler models excel in local design while convolutional models explore deeper sequence spaces.
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
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
ColabFold - Making protein folding accessible to all
Mirdita, M., Schütze, K., Moriwaki, Y., Heo, L., Ovchinnikov, S., & Steinegger, M. (2021) · doi:10.1101/2021.08.15.456425 · Cited by 9,391
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
Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
Lakshminarayanan, B., Pritzel, A., & Blundell, C. (2016) · arXiv · doi:10.48550/arxiv.1612.01474 · Cited by 3,666 · cited 1×
“Model confidence can be estimated by inputting a single sequence into an ensemble of models and evaluating the agreement in prediction across the models 29 .”
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
Watson, J. L., Juergens, D., Bennett, N. R., Trippe, B. L., Yim, J., Eisenach, H. E., Ahern, W., Borst, A. J., Ragotte, R. J., Milles, L. F., et al. (2022) · doi:10.1101/2022.12.09.519842 · Cited by 2,369
De novo design of protein structure and function with RFdiffusion
Watson, J. L., Juergens, D., Bennett, N. R., Trippe, B. L., Yim, J., Eisenach, H. E., Ahern, W., Borst, A. J., Ragotte, R. J., Milles, L. F., et al. (2023) · Nature · doi:10.1038/s41586-023-06415-8 · Cited by 2,325
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
Quantifying the local adaptive landscape of a nascent bacterial community
Ascensao, J. A., Wetmore, K. M., Good, B. H., Arkin, A. P., & Hallatschek, O. (2023) · Nature Communications · doi:10.1038/s41467-022-35677-5 · Cited by 447
Research works printed but not linked: 34
- Romero, P. A. & Arnold, F. H. Exploring protein fitness landscapes by directed evolution. Nat. Rev. Mol. Cell Biol. 10, 866–876 (2009). · DOI · Paper page
- Freschlin, C. R., Fahlberg, S. A. & Romero, P. A. Machine learning to navigate fitness landscapes for protein engineering. Curr. Opin. Biotechnol. 75, 102713 (2022). · DOI · Paper page
- Wittmann, B. J., Johnston, K. E., Wu, Z. & Arnold, F. H. Advances in machine learning for directed evolution. Curr. Opin. Struct. Biol. 69, 11–18 (2021). · DOI · Paper page
- Hawkins-Hooker, A. et al. Generating functional protein variants with variational autoencoders. PLoS Comput Biol. 17, 1–23 (2021). · DOI · Paper page
- Bryant, D. H. et al. Deep diversification of an AAV capsid protein by machine learning. Nat. Biotechnol. 39, 691–696 (2021). · DOI · Paper page
- Greenhalgh, J. C., Fahlberg, S. A., Pfleger, B. F. & Romero, P. A. Machine learning-guided acyl-ACP reductase engineering for improved in vivo fatty alcohol production. Nat. Commun. 12, 1–10 (2021). · DOI · Paper page
- Wittmann, B. J., Yue, Y. & Arnold, F. H. Informed training set design enables efficient machine learning-assisted directed protein evolution. Cell Syst. 12, 1026–1045.e7 (2021). · DOI · Paper page
- Gelman, S., Fahlberg, S. A., Heinzelman, P., Romero, P. A. & Gitter, A. Neural networks to learn protein sequence-function relationships from deep mutational scanning data. Proc. Natl. Acad. Sci. USA 118, e2104878118 (2021). · DOI · Paper page
- Yang, K. K., Wu, Z., Bedbrook, C. N. & Arnold, F. H. Learned protein embeddings for machine learning. Bioinformatics 34, 2642–2648 (2018). · DOI · Paper page
- Hsu, C., Nisonoff, H., Fannjiang, C. & Listgarten, J. Learning protein fitness models from evolutionary and assay-labeled data. Nat. Biotechnol. 40, 1114–1122 (2022). · DOI · Paper page
- Xu, Y. et al. Deep dive into machine learning models for protein engineering. J. Chem. Inf. Model 60, 2773–2790 (2020). · DOI · Paper page
- Li, L. et al. Machine learning optimization of candidate antibody yields highly diverse sub-nanomolar affinity antibody libraries. Nat. Commun. 14, 1–12 (2023).
- Fannjiang, C. & Listgarten, J. Is novelty predictable? 1–30. Preprint at https://arxiv.org/abs/2306.00872 (2023). · arXiv
- Bailey, L. J. et al. Applications for an engineered Protein-G variant with a pH controllable affinity to antibody fragments. J. Immunol. Methods 415, 24–30 (2014). · DOI
- Olson, C. A., Wu, N. C. & Sun, R. A comprehensive biophysical description of pairwise epistasis throughout an entire protein domain. Curr. Biol. 24, 2643 (2014). · DOI · Paper page
- Brookes, D., Park, H. & Listgarten, J. Conditioning by adaptive sampling for robust design. In Proc. of the 36th International Conference on Machine Learning (eds Chaudhuri, K. & Salakhutdinov, R.) vol. 97, 773–782 (PMLR, 2019).
- Wu, N. C., Dai, L., Olson, C. A., Lloyd-Smith, J. O. & Sun, R. Adaptation in protein fitness landscapes is facilitated by indirect paths. Elife 5, e16965 (2016). · DOI · Paper page
- Li, Q. et al. Profiling protease specificity: combining yeast ER Sequestration Screening (YESS) with Next Generation Sequencing. ACS Chem. Biol. 12, 510–518 (2017). · DOI
- McInnes, L., Healy, J. & Melville, J. UMAP: uniform manifold approximation and projection for dimension reduction. Preprint at https://arxiv.org/abs/1802.03426 (2018). · DOI · Paper page
- Bedbrook, C. N. et al. Machine learning-guided channelrhodopsin engineering enables minimally invasive optogenetics. Nat. Methods 16, 1176–1184 (2019). · DOI · Paper page
- Rapp, J. T., Bremer, B. J. & Romero, P. A. Self-driving laboratories to autonomously navigate the protein fitness landscape. Nat. Chem. Eng. 1, 97–107 (2024). · DOI · Paper page
- Chen, L. et al. Learning protein fitness landscapes with deep mutational scanning data from multiple sources. Cell Syst. 14, 706–721.e5 (2023). · DOI
- Wells, J. A. Additivity of mutational effects in proteins. Biochemistry 29, 8509–8517 (1990). · DOI · Paper page
- Skinner, M. M. & Terwilliger, T. C. Potential use of additivity of mutational effects in simplifying protein engineering. Proc. Natl Acad. Sci. USA 93, 10753–10757 (1996). · DOI
- Tokuriki, N., Stricher, F., Schymkowitz, J., Serrano, L. & Tawfik, D. S. The stability effects of protein mutations appear to be universally distributed. J. Mol. Biol. 369, 1318–1332 (2007). · DOI · Paper page
- Gruver, N. et al. Effective surrogate models for protein design with Bayesian optimization. ICML Workshop on Computational Biology (2021).
- Zeng, H. & Gifford, D. K. Quantification of uncertainty in peptide-MHC binding prediction improves high-affinity peptide selection for therapeutic design. Cell Syst. 9, 159–166.e3 (2019). · DOI · Paper page
- Makowski, E. K. et al. Co-optimization of therapeutic antibody affinity and specificity using machine learning models that generalize to novel mutational space. Nat. Commun. 13, 3788 (2022). · DOI · Paper page
- Radivojević, T., Costello, Z., Workman, K., & Garcia Martin, H. A machine learning automated recommendation tool for synthetic biology. Nat. Commun. 11, 1–14 (2020). · DOI · Paper page
- Center for High Throughput Computing. Center for High Throughput Computing. https://doi.org/10.21231/GNT1-HW21 (2006). · DOI
- Basanta, B. et al. An enumerative algorithm for de novo design of proteins with diverse pocket structures. Proc. Natl Acad. Sci. USA 117, 22135–22145 (2020). · DOI · Paper page
- Chao, G. et al. Isolating and engineering human antibodies using yeast surface display. Nat. Protoc. 1, 755–768 (2006). · DOI · Paper page
- Freschlin, C. R., Fahlberg, S. A., Heinzelman, P. & Romero, P. A. Neural network extrapolation to distant regions of the protein fitness landscape. NCBI BioProject, Accession: PRJNA1117877 (2024). · DOI
- Freschlin, C. R., Fahlberg, S. A., Heinzelman, P. & Romero, P. A. Neural network extrapolation to distant regions of the protein fitness landscape. GitHub. https://github.com/RomeroLab/nn-extrapolation (2024). · DOI
Cited by
Discovery of highly active kynureninases for cancer immunotherapy through protein language model
Eom, H., Park, S., Cho, K. S., Lee, J., Kim, H., Kim, S., Yang, J., Han, Y.-H., Lee, J., Seok, C., et al. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkae1245 · Cited by 10
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
A Vast Navigable Protein Sequence Subspace
Sinai, S., Gane, A., Jain, N., Bryant, D., Bileschi, M., Bashir, A., Church, G. M., Kelsic, E., & Colwell, L. (2026) · doi:10.64898/2026.10.18.759989
Metadata
- Authors
- Freschlin, C. R., Fahlberg, S. A., Heinzelman, P., & Romero, P. A.
- Venue
- Nature Communications
- Published
- 2024-07-30
- Volume
- 15
- Issue
- 1
- Pages
- 6405-6405
- ISSN
- 2041-1723 (ISSN-L)
- OpenAlex
- W4401113423
- Cited by
- 48
- References
- 12
- FWCI
- 5.46
- Type
- journal article
- Language
- en
- Open access
- gold
- Licence
- creativecommons.org/licenses/by-nc-nd/4.0 (tdm)
- creativecommons.org/licenses/by-nc-nd/4.0 (vor)
- Funders
- U.S. Department of Health & Human Services | NIH | Center for Information Technology (5R35GM119854)
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Neural network extrapolation to distant regions of the protein fitness landscape. https://app.cochat.ai/papers/10.1038/s41467-024-50712-3
- The paper
- Freschlin, C. R., Fahlberg, S. A., Heinzelman, P., & Romero, P. A. (2024). Neural network extrapolation to distant regions of the protein fitness landscape. Nature Communications. https://doi.org/10.1038/s41467-024-50712-3
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.