Abstract

We show space-economical algorithms for finding maximal unique matches (MUM's) between two strings which are important in large scale genome sequence alignment problems. Our algorithms require only O(n) bits (O(n/log n) words) where n is the total length of the strings. We propose three algorithms for different inputs: In case the input is only the strings, their compressed suffix array, or their compressed suffix tree. Their time complexities are O(n log n), O(n log^ε n) and O(n) respectively, where ε is any constant between 0 and 1. We also show an algorithm to construct the compressed suffix tree from the compressed suffix array using O(n log^ε n) time and O(n) bits space.

Citations