Our Technology

Generative AI: Design, Build, Test

We use evolutionary deep learning to mimic millions of years of evolution in the lab.

Proteins make living things work. Our mission is to create novel proteins that address critical problems.

To achieve this, we turned to evolution for guidance and found a transformative set of principles that explain how proteins function.

Evozyne uses that knowledge to make adaptive, high-performance proteins that solve long-standing therapeutic challenges.

A fundamentally different approach

Physics-based design informed by structure

Uses computational chemistry to make changes to the protein structure to guide new protein design.

Advantages

Does not always require high-throughput assays and effective at producing stable proteins.

Drawbacks

Requires an understanding of protein structures and how they fold—a challenge for many protein families.

Is computationally intensive.

Random design using directed evolution

Induces steps of random changes to the protein to steadily improve function.

Advantages

Effective at designing gradual improvement of functions starting from specific sequences.

Drawbacks

Very local exploration of sequence space.

Challenging to optimize over multiple parameters simultaneously.

Evozyne approach

Evolutionary design directed by Generative AI

Applies the rules of evolution to guide a rational search of a vast design space to create novel high-performance proteins.

AdvantageS

Explores the impact of changes throughout a protein to unlock the full potential of the structure.

Optimizes multiple parameters simultaneously.

Generates a large space of functional protein solutions with extraordinary diversity.

Our Rules

Generative AI simulates millennia of evolution.

Any difficult problem that you want proteins to solve requires a model.

We challenge proteins with a novel problem much like what nature might have done.

We accelerate the search for solutions using our models.

With every iteration, the models learn on what might have taken place over millions of years. This lets us leapfrog into new spaces of protein solutions that nobody has searched before.

Our Research

Decades of peer-reviewed data.

Our platform is based on decades of fundamental research into how and why proteins are built the way they are through the process of evolution.

Below are several key papers on the development of evolutionarily consistent models for protein structure and function.

Papers