Searching gapped palindromes in DNA sequences using Burrows Wheeler type transformation

Abstract

Recent developments in sequencing algorithms have made it easy for scientist to consider many high throughput biological applications. Still, some sequences like gapped palindrome, in which the empirical methods do not work well. Gapped palindromes are those structures that have spacer (set of characters) between left and right palindromic arms. We introduce a fast algorithm for searching two natural classes of gapped palindromes in DNA sequences by using Burrows Wheeler type transformation technique. Our contributions are two fold: First, we introduce a novel gapped palindrome searching algorithms designed for two natural classes of gapped palindromes: Length Constrained and Long Armed. Second, we conduct a comparative performance analysis with existing algorithm for length constrained gapped palindromes.Our algorithmic technique is based on Burrows-Wheeler Transform, in which computational complexity depends only on the length of the sequence. In addition, our proposed algorithms run in time O(n) and are easy to implement. When compared with other algorithm relying on suffix arrays, proposed algorithm is found to be faster by approximately 58% for finding all possible length constrained gapped palindromes in biological sequences.