Science & Publications

Leading with Peer-Reviewed Evidence

Evozyne grounds its platform in transparent, experimentally validated research. Our publications show that generative AI can produce proteins with properties far beyond natural evolution.

Foundational Papers

microPublication Biology Synthetic Rubisco activase proteins with increased thermotolerance Supported by the Gates Foundation

This study demonstrated ML-driven design of synthetic proteins with improved function and stability. It validated the predictive power of Evozyne’s early modeling frameworks.
PNAS ProT-VAE: Protein Transformer Variational AutoEncoder for functional protein design In collaboration with NVIDIA

This work showed significant gains in catalytic activity and a 61°C increase in melting temperature. It established a high-performance modeling backbone for therapeutic protein design.
Chemical Science Molecular latent space simulators
BioRxiv Evolution-based enzyme design
Physical Biology Computational design of hepatitis C virus immunogens
PNAS Mapping membrane activity using machine learning
CELL Origins of allostery and evolvability
CELL Evolvability of proteins

Why These Studies Matter

They provide third-party confirmation of Evozyne’s core technical claims.
They validate EvoGen as a foundation for immunology-focused therapeutic design.
They demonstrate performance leaps only achievable through generative AI.