Date of Award
Summer 2011
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Bioinformatics
First Advisor
Struble, Craig A.
Second Advisor
Sem, Daniel
Third Advisor
Merrill, Stephen
Abstract
Docking is a computational technique which predicts the interaction between a protein and a potential drug compound. Virtual screening is a tool, which employs docking, to investigate huge libraries of compounds and predicts potential drug molecules that bind favorably to the protein of interest. The size of one such commercially available library is about 13 million compounds. It would take approximately 400 years of CPU time to examine this library! As an alternative a high performance computing application with a distributed docking strategy is needed, which can efficiently predict the favorable compounds and can eventually be scaled for huge libraries.
In this thesis, IncreDock, a scoring based incremental docking software has been developed to improve the efficiency of virtual screening process. IncreDock provides two approaches to the distributed docking problem. First, it allows for a completely parallel implementation where the entire library is explored simultaneously in an unordered fashion. Second, and more important, is the ordered incremental parallel implementation where the library is explored in increments and a scoring function is used to determine the order of dockings.
IncreDock was used to perform docking studies on four different proteins using a library of 10,573 compounds. The results suggest that IncreDock is able to predict better compounds in early increments of dockings. IncreDock, thus, provides an effective initial strategy to sample out good ligands in less compute time and forms a good precursor to a tool that can proficiently investigate huge chemical libraries.