Sequence information gain based motif analysis. Issue 1 (December 2015)

• Record Type: Journal Article
• Authors: Maynou, Joan; Pairó, Erola; Marco, Santiago; Perera, Alexandre
• Is Part Of: BMC bioinformatics. Volume 16:Issue 1(2015)
• Page Start: 1
• Page End: 13
• Publication Date: 2015-12
• Subjects: Bioinformatics -- Periodicals; Computational biology -- Periodicals; 570.285
• DOI: 10.1186/s12859-015-0811-x
• Languages: English
• ISSNs: 1471-2105

Abstract

Abstract Background The detection of regulatory regions in candidate sequences is essential for the understanding of the regulation of a particular gene and the mechanisms involved. This paper proposes a novel methodology based on information theoretic metrics for finding regulatory sequences in promoter regions. Results This methodology (SIGMA) has been tested on genomic sequence data forHomo sapiens andMus musculus . SIGMA has been compared with different publicly available alternatives for motif detection, such as MEME/MAST, Biostrings (Bioconductor package), MotifRegressor, and previous work such Qresiduals projections or information theoretic based detectors. Comparative results, in the form of Receiver Operating Characteristic curves, show how, in 70 % of the studied Transcription Factor Binding Sites, the SIGMA detector has a better performance and behaves more robustly than the methods compared, while having a similar computational time. The performance of SIGMA can be explained by its parametric simplicity in the modelling of the non-linear co-variability in the binding motif positions. Conclusions Sequence Information Gain based Motif Analysis is a generalisation of a non-linear model of the cis-regulatory sequences detection based on Information Theory. This generalisation allows us to detect transcription factor binding sites with maximum performance disregarding the covariability observed in the positions of the training set of sequences. SIGMA is freely available to the public athttp://b2slab.upc.edu .