Machine learning-guided design of modular, protein-based medical countermeasures

Protein therapeutics offer a significant opportunity to treat a broad range of chemical and biological threat agents. However, the timeline for developing medical countermeasures (MCM) is long, putting both warfighters and civilians at risk of technological surprise. The proposed work aims to dramatically expedite MCM development and is based on the combination of (i) massively parallel computational methods for designing de novo protein binders, assisted with artificial intelligence (AI) and machine learning (ML), (ii) cell-free expression systems for rapid synthesis and testing of proteins, including glycoproteins, (iii) high-throughput screening assays for protein binding and glycosylation, and (iv) a modular protein engineering platform for building multifunctional therapeutics with designer architectures. Overall, this work will create an integrated computational and experimental pipeline that offers unprecedented speed in developing potent, protein-based MCMs against novel threats.