Year: 2009 Author: Viacheslav Fofanov, PhD Advisor: Marek Kimmel Committee Members: David Scott, Lydia Kavraki Thesis Title: Statistical Models in Protein Structural Alignments Abstract: Recent advances in protein structure determination techniques have greatly increased the rate at which high resolution protein structures become available. However, experimental protein function determination techniques remain expensive and time consuming. This makes computational methods, and particularly partial structural alignment-based approaches for function determination especially attractive. Because these approaches tend to generate substantial numbers of False Positives, a statistical model is necessary to make accurate functional predictions. Herein, we develop and test a statistical framework that takes into account both the unique properties of partial structural alignment algorithms and the geometric and chemical properties of the protein sub-structural being aligned.