Harnessing AI/ML Tools to Design Proteins for Material Breakdown
Recent Nobel Prizes awarded for the use of AI/ML in protein structure prediction demonstrate the incredible potential these capabilities have in designing novel proteins. Biotechnological applications rely on proteins as the workhorses of many capabilities ranging from environmental sensing, human performance, biomanufacturing, and biomaterials. In fact, biotechnology applications often seek to improve these capabilities by designing enzymes with better properties (viz. increased stability or activity). With the availability of programs like AlphaFold, designer proteins are closer to reality. Yet, there remains a knowledge gap when it comes to correlating protein structure to function. For example, one limitation is that AI/ML does not perform as well in predicting the structural impact of mutations, which is a vital step in understanding a protein’s function. Algorithms are also highly dependent on the quality of the training data, which is lacking for families of enzymes (e.g. membrane proteins). Here, I discuss the application of AI/ML algorithms (particularly AlphaFold) and classical biomolecular simulations methods to the study of enzymes involved in the degradation of plastics, including determining protein structure and discovering non-native protein functions to elucidate enzyme promiscuity. Results are discussed in the context of experimental enzymatic activity observed for these enzymes. Our efforts will help enable understanding of enzymatic mechanisms for breakdown of materials that can be leveraged to design improved enzymes, and build toward multi-enzyme systems capable of degrading recalcitrant polymers for achieving the ability to repurpose materials in a circular economy.
IMPACT
Enable understanding of enzymatic mechanisms for breakdown of materials that can be leveraged to design improved enzymes, and build toward multi-enzyme systems for biotechnology capabilities.
PRESENTER
Small, Meagan
meagan.c.small.civ@army.mil
520-691-4921
DEVCOM ARL
CO-AUTHOR(S)
Small, Meagan C
meagan.c.small.civ@army.mil
Sarkes, Deborah A
deborah.a.sarkes.civ@army.mil
Pennington IV, Joseph M
joseph.m.pennington5.civ@army.mil
CATEGORY
Biotechnology
SECONDARY CATEGORY
AI/ML for HPC
SYSTEM(S) USED
Warhawk, Nautilus