Structural Dynamics of Viral Proteins: Development and Application of Multiscale Computational Methods for Studying Viral Capsids, Proteins and Membrane Systems

This proposal presents a vision for my research program over the next five years. My lab is engaged in the application and development of multiscale computational methods to study biomolecular systems, with main driver interests in non-enveloped virus capsids, biomembranes and protein dynamics. Through these studies we will provide modeling tools, gain insights into the regulation and mechanisms of biomolecular functions and apply deep learning methods to gain a new comprehension of protein communication and conformational changes. The main motivation in our viral capsid studies is to understand the determinants and molecular mechanisms of infection processes leading to capsid uncoating/disassembly. Our work on virus systems will expand into the genus of Enteroviruses, which include severe health threats such as poliovirus, EV-A71 and EV-D68. There are numerous high resolution structures of mature, uncoating intermediates and genome released states of enteroviruses which enable this research direction. Significant findings from these studies will include detailed structural and energetic information regarding viral infection related processes, which will be valuable in the development of anti-viral therapies against these agents. Our interests in biomembranes are centrally related to how bilayer lipid composition and shape affect the structure, dynamics and functional properties of lipids, peptides and proteins. Our work on virus capsids and biomembranes will involve multi-resolution approaches including atomistically resolved and coarse-grained models. For virus capsids, coarse-grained models will be advanced to study uncoating and genome release and software will be developed and distributed for this purpose. Our work on biomembranes will be supported by the continued development and enhancement of our BUMPy software for constructing curved membrane systems with biologically inspired shapes. We will also study protein components of the innate immune system, where we will employ supervised machine learning approaches to perform classification of simulations from different structural or biochemical states. Using these approaches will allow us to define new collective variables to compute pathways and energetics of molecular recognition and activation. For all the proposed studies we will employ advanced multiscale modeling methods and have existing or identified new experimental collaborators with mutual interest in these systems and questions, to partner with in these investigations.