. Molecular simulations complement experiments by revealing the microscopic dynamics underlying biological mechanisms and the forces promoting those dynamics. However, most biological processes involve time scales much longer than the time step of numerical integration. While there are many methods for bridging this separation of time scales to obtain equilibrium averages, further advances are needed to robustly estimate dynamical statistics. The proposed research seeks to develop general methods that can meet this need and to apply them to elucidating self-assembly mechanisms at both molecular and cellular length scales. Improving insulin therapies through rare-event analyses of short simulations. There is a pandemic in diabetes mellitus, with tremendous cost worldwide. The main treatment is insulin therapy, but it has a narrow therapeutic index, and its requirement for refrigerated transport and storage is prohibitively costly for much of the world. Insulin analogs have been engineered to have speci?c pharmacokinetics based on knowledge of insulin self- association, but an understanding of how insulin binds to the insulin receptor (IR) remains lacking. We seek to develop computational methods that can enable simulation and analysis of coupled folding and binding reactions and to combine these methods with recently obtained structures of IR bound to insulin and single-chain insulin (SCI) analogs to elucidate the microscopic origins of observed therapeutic activities. The study can thus ultimately lead to improved insulin therapies. We will also investigate the improved thermal properties of SCI analogs, in particular, their reduced tendency to form amyloid ?brils. The study thus also promises to yield insights into amyloid formation, with broad implications beyond insulin to neurodegenerative disorders like Parkinson's and Alzheimer's diseases. Modeling cytoskeletal processes leading to developmental patterning. Cytoskeletal dynamics underlie diverse processes, including developmental patterning, neuronal synapse formation, immunological recognition, wound healing, and tumor growth. These dynamics can be very hard to intuit because they involve balances of me- chanical forces, mechanochemistry, network assembly and dissasembly, and feedback to and from cell signaling molecules. Models thus play an important role in parsing contributing molecular processes and testing quanti- tative hypotheses. We will adapt a recently parameterized cytoskeletal model that is quantitatively predictive in vitro to elucidate mechanisms of developmental patterning in vivo. Namely, we will investigate how interactions between the small GTPase RhoA and actin assembly/dissasembly control pulsatile contractility, a widespread phenomenon that drives cortical ?ow, cell shape change, and tissue deformation. Then we will compare models for the localization of the evolutionarily-conserved RNA-binding protein Staufen during anterior-posterior speci- ?cation. In addition to aiding in understanding these key developmental processes, the simulations will yield a model that can be used to study cytoskeletal dynamics in a broad range of contexts with minimal modi?cation.
. The proposed research seeks to advance and deploy molecular simulation methods that can contribute to the NIH's mission to seek fundamental knowledge to enhance health, lengthen life, and reduce illness and disability. Atomic-resolution simulations will be used to understand the mechanism by which insulin binds to its receptor and itself to enable design of improved insulin therapies for treating diabetes melli- tus. Coarse-grained simulations of cytoskeletal networks will be used to parse the contributions of participating molecular processes to tissue patterning during development, with implications for understanding developmental defects and, more broadly, neuronal synapse formation, wound healing, and tumor growth.