Modeling protein structure alignment based on Markov random field theory

Abstract

The availability of three-dimensional spatial information about protein structures is expanding, as a result of the increased use of both x-ray crystallography and nuclear magnetic resonance spectroscopy for structure elucidation. Frequently, a newly determined structure is similar in fold to a known one, even when no sequence similarity is detectable. Therefore the importance of protein structure alignment parallels that of sequence alignment. Protein structure alignment is to detect the optimal substructure equivalence between two given three-dimensional protein structures. Comparison of three-dimensional protein structures may reveal common arrangement of secondary structures which provides clues to functional relationships between proteins. The location of particularly stable substructures may be highlighted and therefore leads to a better understanding of the principle of protein folding.