In the realm of protein engineering and design, the traditional approach of making incremental changes to protein structures and screening variants through experiments is being revolutionized by the integration of machine learning and AI. A recent study has challenged the long-held belief that altering the core regions of proteins would result in destabilization, opening up new possibilities for more efficient protein design. Researchers from the Centre for Genomic Regulation (CRG) and the Wellcome Sanger Institute conducted a groundbreaking experiment that identified key ‘rules’ governing protein structural stability over billions of years of evolution.
Published in Science under the title ‘Genetics, energetics, and allostery in proteins with randomized cores and surfaces,’ the study involved quantifying the stability of thousands of proteins with randomized cores and surfaces. By systematically randomizing the core and surface sequences of the SH3 domain, the research team discovered that many variants remained stable, challenging the notion of proteins as fragile structures. These findings suggest that protein stability is more akin to Lego than Jenga, where changes to the core can be predicted and managed effectively. Through the integration of machine learning algorithms, the researchers were able to expand their dataset and predict how mutations impact protein sequence and stability. By training energy models on experimental data and testing them against a diverse range of natural SH3 domain sequences, including those from bacteria, plants, insects, and humans, the study offers valuable insights into the fundamental principles underlying protein core stability and its implications for faster protein design.
Read more from genengnews.com