DescriptionProteins custom-designed for specific molecular function have great promise to advance many areas of science and industry. High throughput methods – in particular, effective computational modeling of structure and function – are necessary to identify proteins with novel functions out of the vast number of candidate protein sequences. Nevertheless, state-of-the-art methods have only limited accuracy in predicting the functional impact of even a few mutations. To improve models for functional proteins, we are developing methods within the Rosetta computational protein design software suite that represent subtle structural fluctuations using flexible-backbone ensembles and integrate multiple functional constraints on proteins (i.e. catalytic conformations or binding partners). Promising initial results demonstrated improvement over standard fixed-backbone approaches in initial tests against large curated mutational datasets for experimentally determined binding affinities and high-throughput screening of protein-protein interactions. Our approach of using discrete ensembles to model flexible and dynamic systems is well suited to the distributed nature of high performance computing clusters. Our goal of predicting the functional effect of defined sequences is likewise well suited. In particular, the Open Science Grid will be very useful for our high-processing, low-memory requirements. The purpose of this Startup request is two-fold: benchmarking and optimizing computational methodologies to model functional proteins. First, we will streamline our methodology for XSEDE in preparation for a Research allocation application. Second, additional computing resources from XSEDE will greatly expand our ability develop methodology beyond the limitations of benchmarking on the computational resources at our home institution.
OrganizationUniversity of California, San Francisco
DepartmentBioengineering and Therapeutics
Sponsor Campus GridOSG-XSEDE
Principal Investigator
Samuel Thompson
Field Of ScienceMolecular and Structural Biosciences