July 2, 2024

News

AI model simulates 500 million years of protein evolution

EvolutionaryScale, a new AI research lab focusing on biology, launched Tuesday with a groundbreaking model capable of generating novel proteins by simulating hundreds of millions of years of evolution.

By MICHAEL CIRCLE
The Data Group News Service

EvolutionaryScale, a new AI research lab focusing on biology, launched Tuesday with a groundbreaking model capable of generating novel proteins by simulating hundreds of millions of years of evolution.

The company's ESM3 model created a new Green Fluorescent Protein (GFP), a process that would typically take 500 million years to occur naturally. This achievement demonstrates the model's potential to accelerate scientific discoveries across various fields, including drug development and materials science.

Trained on 2.78 billion proteins using massive computational power, ESM3 is the first generative model for biology that simultaneously processes protein sequence, structure, and function. This capability allows scientists to design and create new proteins from scratch.

EvolutionaryScale's founders previously developed ESM1, widely regarded as the first transformer language model for proteins. Their latest model, ESM3, builds upon this foundation to push the boundaries of AI applications in biology.

"ESM3 takes a step toward a future of biology where AI is a tool to engineer from first principles, the way we engineer structures, machines, and microchips, and write computer programs," said Alexander Rives, EvolutionaryScale co-founder and chief scientist.

The company is releasing an open version of ESM3 for scientific researchers and opening a closed beta API. EvolutionaryScale is also collaborating with Amazon Web Services and NVIDIA to expand the model's accessibility and optimize its performance.

To support its ambitious goals, EvolutionaryScale has secured over $142 million in seed funding. The round was led by Nat Friedman, Daniel Gross, and Lux Capital, with participation from Amazon, NVIDIA's venture capital arm, and angel investors.

As ESM3 becomes available to researchers worldwide, it has the potential to revolutionize biological research, from developing new cancer treatments to creating proteins for carbon capture. The model's ability to compress millions of years of evolution into a single computational process marks a significant milestone in the intersection of artificial intelligence and biology.

For more on this groundbreaking development, please visit EvolutionaryScale Launches with ESM3: A Milestone AI Model for Biology (aboutamazon.com).

***

If your business has a data problem, book your free consultation with The Data Group here!

Build a trust-worthy brand with The Data Group.

The Data Group empowers you to build a truly professional online data warehouse for your product team.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.