A multiple divide-and-conquer (MDC) algorithm for optimal alignments in linear space

Dynamic programming algorithms are often used to find the similarities of sequences as well as to deliver the actual alignment of two sequences. Two kinds of alignments are used to compare sequences: local alignments and global alignments. The local alignments attempt to locate conserved regions, while the global alignments identify overall relationship between two sequences. While dynamic programming algorithms are relatively time consuming, the space required is often the limiting factor when aligning long sequences. A linear space algorithm for computing maximal common subsequences, proposed by Hirschberg, was applied by Myers and Miller to deliver optimal alignments in linear space. The authors have improved the Myers and Miller algorithm by introducing a multiple divide and conquer technique that reduces the algorithm`s running time while maintaining its linear space property. Efficient sequence alignment algorithms have been an active topic in computational biology.