|
Abstract
|
An active area of research in Bioinformatics is finding structural similarity of proteins by alignment. Among many methods, the popular one is to find the similarity based upon statistical features. This method involves gathering information from the com- plex biomolecule structure and obtaining the best alignment by maximizing the number of matched features. In this paper, after reviewing statistical models for matching the struc- tural biomolecule, we would utilize any available information of proteins to create more informative priors which would improve the MCMC mixing compared to previous investi- gations. In particular, we propose to use Delaunay Tetrahedralization (DT) in 3D-space such that the geometric structure, sequence information and amino acids type can be en- tered into the mathematical formulation of the Bayesian matching model of proteins. By enriching the model, we demonstrate that a more feasible empirical prior could be created in the initial stage of the MCMC algorithm. This method shows advantages over compet- ing methods in achieving a global alignment of proteins, accelerating the convergence rate and improving the parameter estimates.
|