Research Interests
1) Protein folding. We use all-atom molecular dynamics simulation to investigate protein folding mechanism from microscopic perspective. Current projects include using the Folding@Home distributed computing platform, along with newly-developed sampling algorithms to compute thermodynamics and kinetics for a large set of sub-millisecond folders.
2) Physics-based structure prediction. We are investigating how physics-based all-atom molecular dynamics simulations can be used with a mechanistic conformational search algorithm, called Zipping and Assembly (ZA), to predict protein folding pathways and native structures. Participation in the CASP assessment is an exciting opportunity to generate large data samples to quantify the role of local structural information in predicting native structure.
3) Multiscale Sampling Methods. We are currently developing a multiscale approach (MSFP) for sampling peptide chains that uses fragment-based potentials derived from all-atom MD, along with a coarse-grained model of the full chain, to efficient sample conformational and sequence space in large systems. This approach has a potentially wide range of applications, from predicting nascent chain sequences that preferentially bind to the ribosome exit tunnel, to quantifying microscopic cooperativities and affinities for biological peptide binding interactions, to drug design.
4) Co-translational, unfolded, and denatured states of proteins. We are applying simulation methods to study, in collaboration with experimentalists, unfolded and guanidine-denatured states, as well as the dynamics of nascent chains being translated by the ribosome.