Computational Biophysics

Many phenomena in biology are mediated by proteins and other biomolecules that have very complex structures and dynamics. We leverage tools from statistical mechanics and machine learning to understand this complexity at the atomistic level, which often necessitates developing new sampling and analytical approaches. Most recently, we have studied how we can predict how evolution will influence protein structure and function and examined novel confirmations and mechanisms in kinases and beta-lactamases.

Protein Evolution

We have recently studied how accurately changes in the thermodynamics of proteins, including their folding and binding free energies, can predict their fitness. Our work has highlighted the need for more accurate high-throughput methods that more effectively account for epistatic infractions, as well as the fact that thermodynamics alone struggles to predict fitness landscapes.

Funding: NSF EPSCoR Genome2Phenome

Topological Descriptors of Protein Conformations

In collaboration with the Crawford Group, we have developed new topological descriptors that can be used to differentiate among protein conformations with high levels of statistical accuracy. SinatraPro enables us to decipher novel conformations from statistical fluctuations with high confidence.