Indexing a sequence for mapping reads with a single mismatch

Mapping reads against a genome sequence is an interesting and useful problem in computational molecular biology and bioinformatics. In this paper, we focus on the problem of indexing a sequence for mapping reads with a single mismatch. We first focus on a simpler problem where the length of the pattern is given beforehand during the data structure construction. This version of the problem is interesting in its own right in the context of the next generation sequencing. In the sequel, we show how to solve the more general problem. In both cases, our algorithm can construct an efficient data structure in O(n log(1+ε) n) time and space and can answer subsequent queries in O(m log log n + K) time. Here, n is the length of the sequence, m is the length of the read, 0<ε<1 and is the optimal output size.