DescriptionKnowledge of 3D protein structures is paramount towards our understanding of biochemistry. Currently, there are many more known protein sequences than 3D protein structures and experimentally determining their structure can be both expensive and time consuming. Therefore, extensive efforts have been made to model these structures using computational methods. Our group has developed the I-TASSER method, which constructs protein structure models by iteratively assembling structure fragments obtained by multiple threading algorithms. The method was stringently tested in the community-wide CASP experiments and has been widely used by the community, including more than 65,000 registered scientists from 122 countries. Despite its success, a major obstacle in the optimization of I-TASSER involves a dearth of computational resources available for use. We have access to a computing cluster composed of 1,100 cores. However, I-TASSER has seen a surge in users on our web server, and as a result, there have been over 2,000 jobs waiting or running at any one time, far exceeding our current capacity. Therefore, an increase in computational resources for our research interests would greatly benefit the further optimization of the I-TASSER method, as well as the biological and medical community as a whole. Over the course of this allocation, we expect to run approximately 400 I-TASSER jobs; each of these jobs would take 500 CPU hours, thus we would require roughly 200,000 CPU hours total. This will be of critical importance for the improvement of the I-TASSER methods and enhance its capacity to serve for the general biological community.
OrganizationUniversity of Michigan
DepartmentDepartment of Computational Medicine and Bioinformatics
Sponsor Campus GridOSG-XSEDE
Principal Investigator
Yang Zhang
Field Of ScienceMolecular and Structural Biosciences