T. M. Chidyausiku

This paper introduces design rules for immunoglobulin-like domains and demonstrates the de novo creation of stable, functional scaffolds with improved biophysical properties.

The paper presents two neural network models for predicting and generating stable proteins based on amino acid sequences, showing improved stability in both expert-designed and model-generated proteins.

The paper uses deep networks to design new proteins by inverting structure prediction models, resulting in 27 stable folded proteins with diverse structures.