@arcinstitute: Imagine programming protein, D...

Imagine programming protein, DNA, and RNA systems like you would write computer code, or even by natural language prompting of an AI agent. @brianhie and team just made this a reality with Proto: a high-level programming language for generative biology.

Proto works from four primitives: a sequence, some form of selective pressure as a constraint, random mutations as proposals, and optimization toward fitness. Nature has been running this loop via evolution for millions of years.

The team rebuilt five published design campaigns in Proto: symmetric protein oligomers, de novo monomers, CRISPR-Cas loci, multi-kilobase chromatin accessibility, and antibody CDRs. The models and optimizers differ in each, but all run on the same four primitives.

Proto can incorporate new models as they’re released. With AlphaGenome, the team designed introns with cell-line-specific splicing, validated in human cells, and combined models to build orthogonal promoter-repressor pairs.

To generate promoter-repressor pairs, Proto’s pipeline composed general-purpose generative and structural models to solve for both protein-protein and protein-DNA interactions simultaneously.

Proto also works with AI coding agents that write programs from natural language instructions. The team used this to specify hundreds of protein complexes across the human proteome, a multi-step signaling pathway, and a cancer-targeting lentiviral therapy.

Proto is open-source and the team hopes it becomes the tool researchers reach for whenever a new biological AI model comes out, either to plug it in themselves or find out if it's already there: proto.evodesign.org/landing

Get started using Proto with co-authors @aditimerch, @dan_guo_, @bviggiano_13, and Lucas Brennan-Almaraz: youtu.be/8UnOrC_0jmo

Check out the preprint: biorxiv.org/content/10.648…

Or go behind the scenes with the developers of Proto: arcinstitute.org/news/proto