libFLASM: a software library for fixed-length approximate string matching

Abstract Background Approximate string matching is the problem of finding all factors of a given text that are at a distance at most k from a given pattern. Fixed-length approximate string matching is the problem of finding all factors of a text of length n that are at a distance at most k from any factor of length ℓ of a pattern of length m. There exist bit-vector techniques to solve the fixed-length approximate string matching problem in time $\mathcal {O}(m\lceil \ell /w \rceil n)$ O ( m ⌈ ℓ / w ⌉ n ) and space $\mathcal {O}(m\lceil \ell /w\rceil)$ O ( m ⌈ ℓ / w ⌉ ) under the edit and Hamming distance models, where w is the size of the computer word; as such these techniques are independent of the distance threshold k or the alphabet size. Fixed-length approximate string matching is a generalisation of approximate string matching and, hence, has numerous direct applications in computational molecular biology and elsewhere. Results We present and make available , a free open-source software library for solving fixed-length approximate string matching under both the edit and the Hamming distance models. Moreover we describe how fixed-length approximate string matching is applied to solve real problems by incorporating into established applications for multiple circular sequence alignment as well as single and structured motif extraction. Specifically, we describe how it can be used to improve the accuracy of multiple circular sequence alignment in terms of the inferred likelihood-based phylogenies; and we also describe how it is used to efficiently find motifs in molecular sequences representing regulatory or functional regions. The comparison of the performance of the library to other algorithms show how it is competitive, especially with increasing distance thresholds. Conclusions Fixed-length approximate string matching is a generalisation of the classic approximate string matching problem. We present , a free open-source software library for solving fixed-length approximate string matching. The extensive experimental results presented here suggest that other applications could benefit from using , and thus further maintenance and development of is desirable.