Methodology article

Chromatin-driven de novo discovery of DNA binding motifs in the human malaria parasite

Elena Y Harris², Nadia Ponts¹, Karine G Le Roch^1* and Stefano Lonardi^2*

• * Corresponding authors: Karine G Le Roch karine.leroch@ucr.edu - Stefano Lonardi stelo@cs.ucr.edu

Author Affiliations

¹ Department of Cell Biology and Neuroscience, University of California, Riverside (CA) 92521 USA

² Department of Computer Science and Engineering, University of California, Riverside (CA) 92521 USA

For all author emails, please log on.

BMC Genomics 2011, 12:601 doi:10.1186/1471-2164-12-601

Published: 13 December 2011

Abstract

Background

Despite extensive efforts to discover transcription factors and their binding sites in the human malaria parasite Plasmodium falciparum, only a few transcription factor binding motifs have been experimentally validated to date. As a consequence, gene regulation in P. falciparum is still poorly understood. There is now evidence that the chromatin architecture plays an important role in transcriptional control in malaria.

Results

We propose a methodology for discovering cis-regulatory elements that uses for the first time exclusively dynamic chromatin remodeling data. Our method employs nucleosome positioning data collected at seven time points during the erythrocytic cycle of P. falciparum to discover putative DNA binding motifs and their transcription factor binding sites along with their associated clusters of target genes. Our approach results in 129 putative binding motifs within the promoter region of known genes. About 75% of those are novel, the remaining being highly similar to experimentally validated binding motifs. About half of the binding motifs reported show statistically significant enrichment in functional gene sets and strong positional bias in the promoter region.

Conclusion

Experimental results establish the principle that dynamic chromatin remodeling data can be used in lieu of gene expression data to discover binding motifs and their transcription factor binding sites. Our approach can be applied using only dynamic nucleosome positioning data, independent from any knowledge of gene function or expression.