- News & Views
- Published:
Protein engineering
(2023)Cite this article
640 Accesses
7 Altmetric
Subjects
Functional proteins with limited homology to natural proteins are designed using a large language model.
Directed evolution has proven remarkably successful at finding variants of known proteins with enhanced properties1. Yet designing proteins that are not homologous to those found in nature is extremely challenging. The strategy of walking uphill on a rugged protein fitness landscape can stall at local optima, making it hard to discover diverse functional variants. Techniques such as DNA shuffling recombine parental variants and allow larger moves in sequence space, but diverse variants are rarely generated because the synthesis process favors sequence-similar parents.
This is a preview of subscription content, access via your institution
Access options
Subscribe to Nature+
Get immediate online access to Nature and 55 other Nature journal
Subscribe to Journal
Get full journal access for 1 year
$99.00
only $8.25 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Tax calculation will be finalised during checkout.
Buy article
Get time limited or full article access on ReadCube.
$32.00
All prices are NET prices.

References
Arnold, F. H. Angew. Chem. Int. Ed. Engl. 58, 14420–14426 (2019).
Madani, A. et al. Nat. Biotechnol. https://doi.org/10.1038/s41587-022-01618-2 (2023).
Romero, P. A., Krause, A. & Arnold, F. H. Proc. Natl Acad. Sci. USA 110, E193–E201 (2013).
Bryant, D. H. et al. Nat. Biotechnol. 39, 691–696 (2021).
Dauparas, J. et al. Science 378, 49–56 (2022).
Russ, W. P. et al. Science 369, 440–445 (2020).
Rives, A. et al. Proc. Natl Acad. Sci. USA 118, e2016239118 (2021).
Dohan, D. et al. In Proc. 27th ACM SIGKDD Conf. Knowledge Discovery & Data Mining 2782–2791 (ACM, 2021).
Repecka, D. et al. Nat. Mach. Intell. 3, 324–333 (2021).
Keskar, N. S. et al. Preprint at arXiv https://doi.org/10.48550/arXiv.1909.05858 (2019).
Ethics declarations
Competing interests
D.B. and L.J.C. have performed research as part of their employment at Google LLC. Google is a technology company that sells machine learning services as part of its business.
Rights and permissions
About this article
Cite this article
Belanger, D., Colwell, L.J. Hallucinating functional protein sequences.
Nat Biotechnol (2023). https://doi.org/10.1038/s41587-022-01634-2
Published:
DOI: https://doi.org/10.1038/s41587-022-01634-2
Associated Content
Read More
David Belanger
