BLSSpeller: exhaustive comparative discovery of conserved cis-regulatory elements

• DOI: 10.1093/bioinformatics/btv466
• Publication Date: 2015-08-08T23:59:31Z
• Authors: Dieter De Witte; Jan Van de Velde; Dries Decap; Michiel Van Bel; Pieter Audenaert; Piet Demeester; Bart Dhoedt; Klaas Vandepoele; Jan Fostier
• Keywords: DNA binding site; Multiple sequence alignment

Abstract Motivation: The accurate discovery and annotation of regulatory elements remains a challenging problem. growing number sequenced genomes creates new opportunities for comparative approaches to motif discovery. Putative binding sites are then considered be functional if they conserved in orthologous promoter sequences multiple related species. Existing methods usually rely on pregenerated sequence alignments, which difficult obtain more diverged species such as plants. As consequence, misaligned often remain undetected. Results: We present novel algorithm that supports both alignment-free alignment-based the motifs exhaustively enumerated words over IUPAC alphabet screened conservation using branch length score. Additionally, confidence score is established genome-wide fashion. In order take advantage cloud computing infrastructure, MapReduce programming model adopted. method applied four monocotyledon plant it shown high-scoring significantly enriched open chromatin regions Oryza sativa transcription factor inferred through protein-binding microarrays O.sativa Zea mays. Furthermore, recover experimentally profiled ga2ox1-like KN1 Z.mays. Availability implementation: BLSSpeller was written Java. Source code manual available at http://bioinformatics.intec.ugent.be/blsspeller Contact: [e-mail] or [e-mail] Supplementary information: data Bioinformatics online.