Sampling protein language models for functional protein design
Cell Systems · 2026
DOI: 10.1016/j.cels.2026.101714 ↗Kurz gefasst
Fachartikel in der Zeitschrift Cell Systems, 2026. Verfasst von Jeremie Theddy Darmawan und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore. Electronic address: jeremie.darm…. Hauptthemen laut MEDLINE: Proteins, Protein Engineering. Die Arbeit nennt 77 Literaturangaben. Gefördert von: Engineering and Physical Sciences Research Council, The Alan Turing Institute.
Abstract (Originalsprache)
Protein language models have emerged as powerful tools for learning rich protein representations, improving tasks like structure prediction, mutation effect estimation, and homology detection. Their ability to model complex sequence distributions also holds promise for designing novel, functional proteins with applications in therapeutics, materials, and sustainability. Given the vastness of sequence space, efficient exploration is essential. However, most existing design approaches rely on single-mutant sampling strategies borrowed from natural language processing, which fail to capture epistatic interactions critical for function. Here, we develop an in silico framework to systematically compare sampling methods and introduce several approaches tailored for protein design. We show that sampling multiple mutations simultaneously substantially outperforms single-mutant approaches by better capturing epistatic effects. We evaluate these strategies across three protein families spanning eukaryotic, prokaryotic, and viral origins, examine key hyperparameters, and validate our findings on a de novo binder design task against a major therapeutic target.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore. Electronic address: jeremie.darmawan@u.nus.edu.
- Department of Computer Science, University of Oxford, Oxford, UK.
- Department of Systems Biology, Harvard Medical School, Boston, MA, USA. Electronic address: research@pascal-notin.com.
Förderung
- Engineering and Physical Sciences Research Council
- The Alan Turing Institute
Zitieren
APA
Darmawan, J. T., Gal, Y., & Notin, P. (2026). Sampling protein language models for functional protein design. Cell Systems. https://doi.org/10.1016/j.cels.2026.101714
Vancouver
Darmawan JT, Gal Y, Notin P. Sampling protein language models for functional protein design. Cell Systems. 2026. doi:10.1016/j.cels.2026.101714
BibTeX
@article{Darmawan2026,
author = {Darmawan, Jeremie Theddy and Gal, Yarin and Notin, Pascal},
title = {Sampling protein language models for functional protein design},
journal = {Cell Systems},
year = {2026},
doi = {10.1016/j.cels.2026.101714}
}
EUVIMED