A linear-time algorithm for studying genetic variation

Abstract

The study of variation in DNA sequences, within the framework of phylogeny or population genetics, for instance, is one of the most important subjects in modern genomics. We here present a new lineartime algorithm for finding maximal k-regions in alignments of three sequences, which can be used for the detection of segments featuring a certain degree of similarity, as well as the boundaries of distinct genomic environments such as gene clusters or haplotype blocks, k-regions are defined as these which have a center sequence whose Hamming distance from any of the alignment rows is at most k, and their determination in the general case is known to be NP-hard. © Springer-Verlag Berlin Heidelberg 2006.