Description:
- Biocatalysis and materials science; Bioremediation
- Food Industry
- Diagnostics and therapeutics
Abstract
USC researchers employed a maximum entropy (EMaxEnt) approach to accurately predict enzyme catalysis at the active site region, as well as stability at the more distant region. By inferring statistical energy from homologous sequences with the EMaxEnt principle, the procedure is able to elucidate enzyme architecture. The approach offers a powerful method for guiding enzyme design through evolutionary information.
Benefit
- Improved enzyme design
- Enzyme function design
- Accurately predicts enzyme design properties
- Synthesizes evolutionary approach with enzymology
- Minimal computational costs
Market Application
Enzyme design has emerged as a strategy for harnessing the biological power of enzymes to solve threats including the energy crisis, environmental pollution, and food shortages. To date, enzyme design techniques have utilized physics-based approaches, but their progress has been slow. Machine learning offers a promising new avenue for more efficient enzyme design; however, attempts have not yet accurately predicted the catalytic power of enzymes. Thus, an efficient and accurate machine learning technique for enzyme design is still needed.
Publications
Enhancing computational enzyme design by a maximum entropy strategy. Proc Natl Acad Sci U S A. (2022).
Other
Stage of Development
- Tested with a database of enzyme efficiency upon mutation