Integration of Structural and Sequence Information for Homology-Based Modeling of Proteins

摘要

In the six month supplement to this project the authors continued developing the approaches defined in the original proposal (evolutionary profile), and the family pairwise search (FPS) method for defining sequence patterns. This work resulted in several fundamental publications regarding methods for making statistical evaluations of sequence match scores, and a submitted manuscript validating the evolutionary profile approach. Homology modeling allows one to predict the three-dimensional structure of a novel query sequence based on the known three-dimensional structure of a homologous template sequence. This approach requires that one find a three-dimensional structure that is homologous to the query sequence. The query sequence must then be mapped onto the template sequence such that each amino acid residue appears in the correct position in the modeled structure. Both the identification of homologous sequences and the mapping of the query onto the template structure are achieved through sequence comparisons. Standard sequence comparison methods use a single amino acid residue comparison matrix, such as BLOSUM 62 or PAM250, to identify proteins that are more similar than average. When the similarity is high enough, one can infer that these proteins are homologous and thus likely to have similar structure and function.