AI Trained on Evolution’s Playbook Develops Proteins that Spur Drug and Scientific Discovery

September 25, 2024 • by Marc Airhart

EvoRank offers a new and tangible example of how AI may help bring disruptive change to biomedical research and biotechnology more broadly.

A colorful ribbon with elaborate twirls and twists represents the three-dimensional shape of a molecule

Using the MutRank framework trained with EvoRank, Danny Diaz and professor Andrew Ellington’s team are developing an improved version of a protein critical for the biomanufacturing of mRNA therapeutics and vaccines. In this example, the model recommends keeping the blue parts the same as the natural version of the protein and strongly considering mutating the red parts. Credit: Danny Diaz/University of Texas at Austin.


A spreadsheet with about a dozen rows and a couple of dozen columns. Each field contains a letter that reprsents an amino acid in a sequence.

Each row in this chart represents the same protein in a different organism, each with its own subtle tweaks. Each letter represents a different amino acid in a sequence (e.g., H = histidine). Researchers can look at one position in the amino acid sequence (e.g., the column outlined in red) and see how often evolution selected a particular amino acid for that position in the protein. Credit: University of Texas at Austin.

Share


A man in a plaid shirt smiles in front of a chalkboard with equations, as graphic elements of limestone and bursts frame the shot.

The Oden Institute

Transforming the Use of AI in Drug Design

Headshots of Dr. Gilpin, Dr. Kim and Dr. Baldini

Accolades

Three College of Natural Sciences Faculty Win NSF CAREER Awards