MMDiff: quantitative testing for shape changes in ChIP-Seq data sets.

Area of Science:

• Genomics and Epigenetics
• Computational Biology
• Bioinformatics

Background:

• Cell-specific gene expression relies on epigenetic modifications and transcription factor binding.
• Quantitative comparison of ChIP-Seq data is challenging due to limitations in current count-based analysis methods.
• Existing methods often overlook peak shape information, focusing solely on fragment counts.

Purpose of the Study:

• To introduce MMDiff, a novel computational method for analyzing ChIP-Seq data.
• To address the limitations of current methods by incorporating peak shape information.
• To provide a robust and broadly applicable tool for detecting differences in ChIP-Seq datasets.

Main Methods:

• MMDiff quantifies differences by analyzing changes in signal profile shapes within ChIP-Seq data.
• The method was validated using simulated datasets and compared against four alternative approaches.
• MMDiff was applied to empirical datasets, including histone modifications (H3K4me3, H3K27ac) and transcription factor binding (CTCF).

Main Results:

• MMDiff demonstrated superior performance, especially when peak profiles exhibited changes between samples.
• The analysis of H3K4me3 data showed reproducible results and identified functionally significant histone modification changes.
• MMDiff proved complementary to count-based methods for H3K27ac and CTCF binding data, also detecting changes in homotypic binding events.

Conclusions:

• Higher-order features of ChIP-Seq peaks contain crucial information beyond total counts for differential analysis.
• MMDiff effectively utilizes these features, offering a valuable new approach for ChIP-Seq data analysis.
• The developed method fills a critical gap in the analysis of differential histone modifications and transcription factor binding.