A Dirichlet Process Mixture of Hidden Markov Models for Protein Structure
Kristin Lennox, Department of Statistics, Texas A&M University
The goal of protein structure prediction is to determine the three-dimensional structure of a protein from its amino acid sequence. In the case of template-based prediction, a protein with unknown structure is similar in amino acid sequence to proteins with solved three-dimensional structures. This family of related proteins provides a template for the structure of the target protein. Template-based methods have historically been limited because protein families tend to be small and certain alignment positions within a family correspond to few or no solved structures.