Simulating 500 million years of evolution with a language model

More than 3 billion years of evolution have produced an image of biology encoded into the space of natural proteins. Here, we show that language models trained at scale on evolutionary data can generate functional proteins that are far away from known proteins. We present ESM3, a frontier multimodal generative language model that reasons over the sequence, structure, and function of proteins. ESM3 can follow complex prompts combining its modalities and is highly responsive to alignment to improve its fidelity. We have prompted ESM3 to generate fluorescent proteins. Among the generations that we synthesized, we found a bright fluorescent protein at a far distance (58% sequence identity) from known fluorescent proteins, which we estimate is equivalent to simulating 500 million years of evolution.

Authors

Institutions

Publication Details

Journal
Science
Published
2025-01-16
DOI
https://doi.org/10.1126/science.ads0018
Citations
482
Primary Topic
Genomics and Phylogenetic Studies
Type
article
Field-Weighted Citation Impact
379.64
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Simulating 500 million years of evolution with a language model

Halil Akin, Neil Thomas, Tom Sercu, Carolyn Kim et al.
482 citations
Science
Genomics and Phylogenetic Studies
379.64
article

Simulating 500 million years of evolution with a language model

Halil Akin, Neil Thomas, Tom Sercu, Carolyn Kim, Nicholas Sofroniew, Alexander Rives, Roshan Rao, Yousuf A. Khan, Deniz Oktay, Jun Gong, Jonathan Deaton, Patrick D. Hsu, Rohil Badkundri, Alexander Derry, Salvatore Candido, Robert Verkuil, Thomas Hayes, Marius Wiggert, Irhum Shafkat, Liam J. Bartie, Matthew Nemeth, Vincent Q. Tran, Zeming Lin, R. Molina, Chetan Mishra
article en
482 citations

Abstract

More than 3 billion years of evolution have produced an image of biology encoded into the space of natural proteins. Here, we show that language models trained at scale on evolutionary data can generate functional proteins that are far away from known proteins. We present ESM3, a frontier multimodal generative language model that reasons over the sequence, structure, and function of proteins. ESM3 can follow complex prompts combining its modalities and is highly responsive to alignment to improve its fidelity. We have prompted ESM3 to generate fluorescent proteins. Among the generations that we synthesized, we found a bright fluorescent protein at a far distance (58% sequence identity) from known fluorescent proteins, which we estimate is equivalent to simulating 500 million years of evolution.

ScienceVol. 387(6736)
Palo Alto Institute (US), New York Consortium in Evolutionary Primatology (US), Arc Research Institute, University of California, Berkeley (US)
Openalex Percentile: Top 0%
Genomics and Phylogenetic Studies
379.64
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.