Sequence comparison is one of the major tasks in bioinformatics, which could serve as evidence of structural and functional conservation, as well as of evolutionary relations. There are several similarity/dissimilarity measures for sequence comparison, but challenges remains. This paper presented a binomial model-based measure to analyze biological sequences. With help of a random indicator, the occurrence of a word at any position of sequence can be regarded as a random Bernoulli variable, and the distribution of a sum of the word occurrence is well known to be a binomial one. By using a recursive formula, we computed the binomial probability of the word count and proposed a binomial model-based measure based on the relative entropy. The proposed measure was tested by extensive experiments including classification of HEV genotypes and phylogenetic analysis, and further compared with alignment -based and alignment-free measures. The results demonstrate that the proposed measure based on binomial model is more efficient.

Key words: Word count; Binomial model; Sequence comparison; Classification; Phylogenetic analysis.

This article can be cited as:
X. Liu, Q. Dai, L. Li, Z. He, An Efficient Binomial Model-Based Measure for Sequence Comparison and its Application, J. Biomol Struct Dyn 28(5) 833-843 (2011)