From reads to regions: a Bioconductor workflow to detect differential binding in ChIP-seq data.

Abstract

Chromatin immunoprecipitation with massively parallel sequencing (ChIP-seq) iswidely used to identify the genomic binding sites for protein of interest. Most conventionalapproaches to ChIP-seq data analysis involve the detection of the absolutepresence (or absence) of a binding site. However, an alternative strategy is toidentify changes in the binding intensity between two biological conditions, i.e., differential binding (DB). This may yield more relevant results than conventional analyses,as changes in binding can be associated with the biological difference beinginvestigated. The aim of this article is to facilitate the implementation of DB analyses,by comprehensively describing a computational workflow for the detection ofDB regions from ChIP-seq data. The workflow is based primarily on R softwarepackages from the open-source Bioconductor project and covers all steps of theanalysis pipeline, from alignment of read sequences to interpretation and visualizationof putative DB regions. In particular, detection of DB regions will be conductedusing the counts for sliding windows from the csaw package, with statistical modellingperformed using methods in the edgeR package. Analyses will be demonstratedon real histone mark and transcription factor data sets. This will providereaders with practical usage examples that can be applied in their own studies.