ChIPComp

ChIPComp is an R software package for rigorous statistical analysis comparing multiple ChIP-seq datasets. ChIP-seq technology, which measures protein binding or histone modification across the genome, needs a comprehensive method that accounts for control experiment data, signal-to-noise ratios, biological variations, and multifactor experimental designs. ChIPComp introduces a statistical approach to quantitatively compare multiple ChIP-seq datasets and identify genomic regions with differential protein binding or histone modification.

The method begins by detecting and uniting peaks from all datasets to create a single set of candidate regions. It assumes the read counts from the immunoprecipitation (IP) experiment at these candidate regions follow a Poisson distribution. The Poisson rates are modeled to reflect experimental artifacts and biological signals specific to each experiment. By estimating the biological signals and comparing them through hypothesis testing within a linear model framework, ChIPComp can accurately and robustly detect differences across datasets.