CoChat

ProT-VAE: Protein Transformer Variational AutoEncoder for functional protein design

Open access
Sevgen, Emre, Moller, Joshua, Lange, Adrian, Parker, John, Quigley, Sean, Mayer, Jeff, Srivastava, Poonam, Gayatri, Sitaram, Hosfield, David, Dilks, Clayton, Buchanan, Claire, Speltz, Thomas, Korshunova, Maria, Livne, Micha, Gill, Michelle, Ranganathan, Rama, Costa, Anthony B., Ferguson, Andrew L. · 2025 · Proceedings of the National Academy of Sciences
Cited by 52

The DOI matches its Crossref record; 28 of the 28 references it cites were found in a registry or an index.

Abstract

Deep generative models have demonstrated success in learning the protein sequence to function relationship and designing synthetic sequences with engineered functionality. We introduce the Protein Transformer Variational AutoEncoder (ProT-VAE) as an accurate, generative, fast, and transferable model for data-driven protein design that blends the merits of variational autoencoders to learn interpretable, low-dimensional latent embeddings for conditional sequence design with the expressive, alignment-free featurization offered by transformer-based protein language models. We implement the model using NVIDIA’s BioNeMo framework and validate its performance in retrospective functional prediction and prospective functional design. The model identifies a phenylalanine hydroxylase enzyme with 2.5× catalytic activity over wild-type, and a γ -carbonic anhydrase enzyme with a melting temperature elevation of ΔT m = +61 ° C relative to the most thermostable sequence reported to date and activity in 23% v/v methyl diethanolamine at pH 11.25 and 93 ° C corresponding to industrially relevant conditions for enzymatic carbon capture technologies. The ProT-VAE model presents a powerful and experimentally validated platform for machine learning-guided directed evolution campaigns to discover synthetic proteins with engineered function.

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.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1073/pnas.2408737122

Cites

28 of 28 references verified0 citation errors · 0 not checkable

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.

A few free checks a day · papers already verified open instantly · results are public

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

Show all 28 references

Research works printed but not linked: 57

  1. 10.1038/s41592-019-0496-6 · DOI · Paper page
  2. 10.1038/s41592-020-0848-2 · DOI · Paper page
  3. 10.1002/prot.26194 · DOI · Paper page
  4. Z. Costello H. G. Martin How to hallucinate functional proteins. arXiv [Preprint] (2019). http://arxiv.org/abs/1903.00458. Accessed: 15 Feb 2025. · arXiv
  5. 10.1038/s41598-018-34533-1 · DOI · Paper page
  6. S. Sinai E. Kelsic G. M. Church M. A. Nowak Variational auto-encoding of protein sequences. arXiv [Preprint] (2017). http://arxiv.org/abs/1712.03346 (Accessed 15 February 2025). · arXiv
  7. 10.1016/j.cels.2024.07.005 · DOI
  8. 10.1038/s41598-022-05195-x · DOI · Paper page
  9. C. Angermueller “Model-based reinforcement learning for biological sequence design” in International Conference on Learning Representations 2020 (International Conference on Learning Representations 2020) pp. 1–23.
  10. A. Madani Progen: Language modeling for protein generation. arXiv [Preprint] (2020). http://arxiv.org/abs/2004.03497 (Accessed 15 February 2025). · arXiv
  11. 10.1038/s41587-022-01618-2 · DOI · Paper page
  12. P. Notin “Tranception: Protein fitness prediction with autoregressive transformers and inference-time retrieval” in International Conference on Machine Learning (Proceedings of Machine Learning Research 2022) pp. 16990–17017.
  13. 10.1038/s42256-022-00499-z · DOI · Paper page
  14. 10.1038/s41467-021-25976-8 · DOI · Paper page
  15. C. Dallago Benchmark tasks in fitness landscape inference for proteins. bioRxiv [Preprint] (2021). https://doi.org/10.1101/2021.11.09.467890 (Accessed 15 February 2025). · DOI · Paper page
  16. D. P. Kingma M. Welling Auto-encoding variational Bayes. arXiv [Preprint] (2013). http://arxiv.org/abs/1312.6114 (Accessed 15 February 2025). · arXiv
  17. S. Zhao J. Song S. Ermon “Infovae: Balancing learning and inference in variational autoencoders” in Proceedings of the AAAI Conference on Artificial Intelligence (Association for the Advancement of Artificial Intelligence 2019) vol. 33 pp. 5885–5892. · DOI · Paper page
  18. 10.1038/s42256-022-00532-1 · DOI
  19. 10.1063/5.0058639 · DOI
  20. A. Vaswani , Attention is all you need. Adv. Neural Inf. Process. Syst. 30, 6000–6010 (2017).
  21. T. Wang X. Wan “T-cvae: Transformer-based conditioned variational autoencoder for story completion” in Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI-19) (2019) pp. 5233–5239. · DOI
  22. J. Jiang G. G. Xia D. B. Carlton C. N. Anderson R. H. Miyakawa “Transformer VAE: A hierarchical model for structure-aware and interpretable music representation learning” in ICASSP 2020–2020 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP) (Institute of Electrical and Electronics Engineers 2020) pp. 516–520. · DOI
  23. C. Li Optimus: Organizing sentences via pre-trained modeling of a latent space. arXiv [Preprint] (2020). http://arxiv.org/abs/2004.04092 (Accessed 15 February 2025). · DOI · Paper page
  24. S. Park J. Lee Finetuning pretrained transformers into variational autoencoders. arXiv [Preprint] (2021). http://arxiv.org/abs/2108.02446 (Accessed 15 February 2025). · arXiv
  25. D. M. Arroyo J. Postels F. Tombari “Variational transformer networks for layout generation” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (Institute of Electrical and Electronics Engineers 2021) pp. 13642–13652. · DOI
  26. J. Henderson F. Fehr A variational autoencoder for transformers with nonparametric variational information bottleneck. arXiv [Preprint] (2022). http://arxiv.org/abs/2207.13529 (Accessed 15 February 2025). · arXiv
  27. S. Zhao J. Song S. Ermon InfoVAE: Information maximizing variational autoencoders. arXiv [Preprint] (2017). http://arxiv.org/abs/1706.02262 (Accessed 15 February 2025). · arXiv
  28. 10.1038/359851a0 · DOI · Paper page
  29. 10.1242/jcs.114.7.1253 · DOI · Paper page
  30. 10.1038/nature02178 · DOI · Paper page
  31. 10.1002/iub.1150 · DOI · Paper page
  32. 10.1074/jbc.273.27.16962 · DOI · Paper page
  33. 10.1093/nar/gki025 · DOI · Paper page
  34. 10.1038/s41467-022-29443-w · DOI · Paper page
  35. 10.1042/BCJ20160115 · DOI · Paper page
  36. R. S. Rowlett Structure and catalytic mechanism of the β-carbonic anhydrases. Biochim. Biophys. Acta Proteins Proteomics 1804 362–373 (2010). · DOI · Paper page
  37. 10.1016/j.bbapap.2009.08.026 · DOI · Paper page
  38. 10.1016/j.seppur.2012.06.022 · DOI
  39. 10.1074/jbc.M009182200 · DOI · Paper page
  40. 10.1002/j.1460-2075.1996.tb00588.x · DOI · Paper page
  41. 10.1021/bi000204s · DOI · Paper page
  42. 10.1073/pnas.1411461111 · DOI · Paper page
  43. 10.1128/jb.178.11.3270-3274.1996 · DOI
  44. 10.3390/su13010072 · doi:10.3390/su13010072
  45. 10.1021/acsmacrolett.0c00885 · DOI
  46. 10.1016/j.copbio.2022.102713 · DOI · Paper page
  47. 10.1021/acscatal.9b04321 · DOI · Paper page
  48. 10.1021/acs.jpcb.1c02449 · DOI
  49. 10.1016/j.sbi.2021.01.008 · DOI · Paper page
  50. 10.1021/acs.jcim.0c00073 · DOI · Paper page
  51. 10.1016/j.sbi.2021.03.009 · DOI
  52. 10.1093/nar/gkac1080 · DOI · Paper page
  53. S. Liu A text-guided protein design framework. arXiv [Preprint] (2023). http://arxiv.org/abs/2302.04611 (Accessed 15 February 2025). · arXiv
  54. 10.1016/j.sbi.2021.11.008 · DOI · Paper page
  55. B. Lai M. McPartlon J. Xu End-to-end deep structure generative model for protein design. bioRxiv [Preprint] (2022). https://doi.org/10.1101/2022.07.09.499440 (Accessed 15 February 2025). · DOI
  56. 10.1016/j.sbi.2023.102571 · DOI · Paper page
  57. M. Romiti P. Cooper Data from “RefSeq FTP.” NCBI Protein. https://www.ncbi.nlm.nih.gov/protein/. Accessed 25 August 2022.

Cited by

See all 52

Metadata

Authors
Sevgen, E., Moller, J., Lange, A., Parker, J., Quigley, S., Mayer, J., Srivastava, P., Gayatri, S., Hosfield, D., Dilks, C., Buchanan, C., Speltz, T., Korshunova, M., Livne, M., Gill, M., Ranganathan, R., Costa, A. B., & Ferguson, A. L.
Venue
Proceedings of the National Academy of Sciences
Published
2025-10-06
Volume
122
Issue
41
Pages
e2408737122-e2408737122
ISSN
  • 0027-8424 (ISSN-L)
  • 1091-6490
OpenAlex
W4414857930
Cited by
52
References
28
FWCI
7.92
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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: ProT-VAE: Protein Transformer Variational AutoEncoder for functional protein design. https://app.cochat.ai/papers/10.1073/pnas.2408737122
The paper
Sevgen, E., Moller, J., Lange, A., Parker, J., Quigley, S., Mayer, J., Srivastava, P., Gayatri, S., Hosfield, D., Dilks, C., Buchanan, C., Speltz, T., Korshunova, M., Livne, M., Gill, M., Ranganathan, R., Costa, A. B., & Ferguson, A. L. (2025). ProT-VAE: Protein Transformer Variational AutoEncoder for functional protein design. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.2408737122

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.