ChIPComp

ChIPComp performs quantitative comparison of multiple ChIP-seq datasets to identify differential protein-DNA binding and histone modifications.

Key Features:

• Peak detection and union formation: Detects peaks across all datasets and unifies them into a single set of candidate regions for comparison.
• Poisson distribution modeling: Assumes read counts from immunoprecipitation (IP) experiments at candidate regions follow a Poisson distribution.
• Biological signal estimation: Models underlying Poisson rates as functions of artifacts and experiment-specific biological signals to separate biological variation from technical noise.
• Hypothesis testing in a linear model framework: Compares estimated biological signals using hypothesis testing within a linear model to detect differential binding or histone modifications.
• Control data and complex designs: Incorporates control data and accommodates complex experimental designs and comparisons across conditions.

Scientific Applications:

• Comparative genomics: Identifying differentially bound regions or histone modifications across cell types, developmental stages, or treatment conditions.
• Epigenetic research: Investigating changes in chromatin structure associated with gene regulation via histone modifications.
• Transcription factor binding studies: Analyzing variation in transcription factor binding sites across experimental conditions.

Methodology:

Initial peak detection across all datasets to form candidate regions; modeling IP read counts at candidate regions with a Poisson distribution whose rates are modeled as functions of artifacts and experiment-specific biological signals; and hypothesis testing of estimated biological signals within a linear model framework.