Sequence alignment algorithms and database search methods use BLOSUM and PAM substitution matrices constructed from general proteins. These de facto matrices are not optimal to align sequences accurately, for the proteins with markedly different compositional bias in the amino acid. In this work, a new amino acid substitution matrix is calculated for the disorder and low complexity rich region of Hub proteins, based on residue characteristics. Insights into the amino acid background frequencies and the substitution scores obtained from the Hubsm unveils the residue substitution patterns which differs from commonly used scoring matrices .When comparing the Hub protein sequences for detecting homologs, the use of this Hubsm matrix yields better results than PAM and BLOSUM matrices. Usage of Hubsm matrix can be optimal in database search and for the construction of more accurate sequence alignments of Hub proteins.
A Study of Hough Transform-based Fingerprint Alignment Algorithms
