Protein-nucleic acid (NA) interactions are central to numerous biological processes and to biotechnology. Yet, it is still difficult to predict structures and specificities of natural protein-NA complexes in silico, limiting understanding of their structure-function relationships and the ability to bioprospect valuable nucleic acid-binding proteins. When it comes to design tasks, scientists have been restricted to the repurposing of natural NA binding proteins through domain fusions, small sets of rational or laboratory-evolved mutations, and conservative manipulation of structures. This limits the potential application space and the efficacy of solutions.

Our lab’s goal is to develop new capabilities for computational prediction and design of protein-NA assemblies and apply these capabilities to address real-world problems. To do this we use a variety of tools including both physics-based and AI/ML approaches for protein design/modeling, high-throughput biochemistry, molecular biology, and sequencing technologies. Of particular interest is the development of cutting-edge protein functional assays to augment training of AI design models in low-data scenarios.