Multiple alignment of biological sequences with gap flexibility

Abstract

We present a new heuristic approach to obtaining multiple alignments of biological sequences, such as DNA and proteins. Pairwise alignments of sequences are represented as special directed acyclic graphs, which allow the encoding of several optimal alignments in one compact structure. In this graph a node is a character from a sequence and edges link consecutive characters. Gaps and mismatches are not fixed, and can be better positioned when other graphs (i.e. pairwise alignments) are considered. This is done by a heuristic that joins graphs in their order of score. From the final graph a multiple alignment is obtained by an algorithm that runs in O(m + n log n) time, where n is the number of nodes and m the number of edges in the graph. Using the ideas here presented we aligned six protein sequences, and the result compares favorably with another published alignment. Our implementation is also capable of performing DNA fragment assembly, and we present a simple comparison with a popular fragment assembly package.