Engineered Thermostable Enzymes via Massively Parallel Design and Construction of Protein Libraries
Organizations: University of Texas at Austin and Twist Bioscience
This project, led by Dr. Andrew Ellington at the University of Texas-Austin, will establish and evaluate Deep Stability Scanning, a method that synergizes advances in DNA synthesis, machine learning-enabled protein design, and high-throughput screening, for protein engineering. The use of Deep Stability Scanning aims to accelerate the pace of engineering sensors and enzymes in support of chemical manufacturing.
This work addresses a central challenge in synthetic biology: the search for beneficial mutations – particularly in protein engineering – where a few amino acid mutations can be the difference between an economically competitive production organism and a costly prototype. Currently, the identification of new protein biosensors, catalysts, and regulators largely relies on mining new parts from phylogeny, or on the use of often time-consuming directed evolution methods for the optimization of protein parts.
Through the unique combination of several new technologies, including machine learning algorithms for identifying stabilizing mutations for the functional improvement of virtually any protein, and Mazzively Parallel© Protein Libraries, which rapidly builds thousands of barcoded protein expression plasmids with a 100x reduction in price. Together, these approaches should radically accelerate and streamline protein and metabolic engineering, with the goal of enabling any BioMADE member to order-on-demand an optimized protein library that would previously be a laborious multi-year project.
Year Launched: 2024
Funding source: BioMADE Open Project Call