An introduction to computational tools for differential binding analysis with ChIP‐seq data

Abstract

Background: Gene transcription in eukaryotic cells is collectively controlled by a large panel of chromatin associated proteins and ChIP-seq is now widely used to locate their binding sites along the whole genome. Inferring the differential binding sites of these proteins between biological conditions by comparing the corresponding ChIP-seq samples is of general interest, yet it is still a computationally challenging task. Results: Here, we briefly review the computational tools developed in recent years for differential binding analysis with ChIP-seq data. The methods are extensively classified by their strategy of statistical modeling and scope of application. Finally, a decision tree is presented for choosing proper tools based on the specific dataset. Conclusions: Computational tools for differential binding analysis with ChIP-seq data vary significantly with respect to their applicability and performance. This review can serve as a practical guide for readers to select appropriate tools for their own datasets.

RepViz: a replicate-driven R tool for visualizing genomic regions

“…Additionally, the tool allows a replicate-driven inspection of the length of the called peak. This is useful because several peak callers tend to combine clusters of sharp peaks to broader peaks [11, 12]. Finally, the third function of RepViz visualizes the gene track to display the genes in the region of interest, such as gene promoters or their vicinity.…”

“…Currently, genomic visualization of the sequencing data is especially important in the analysis of chromatin data, such as ChIP-seq and ATAC-seq. Specific histone modification markers with distinct dynamics require custom parameterization in calling the differential signal and, therefore, constitute a more complex situation compared to, for example, RNA-seq analysis [11, 12]. Accordingly, the selection of a proper peak calling or differential peak calling tool and parameters for specific histone modification markers is often a complex and iterative process in which visualization has an important role.…”

MAnorm2 for quantitatively comparing groups of ChIP-seq samples

“…We selected two representative computational tools for group-level differential ChIP-seq analysis to compare with MAnorm2. The two tools, named ChIPComp [18] and PePr [28], represent two broad classes of methods for differential ChIP-seq analysis [6,14]. More specifically, ChIPComp requires its users to provide pre-defined peaks for each single ChIP-seq sample while PePr has no such requirement (we could see that MAnorm2 and…”

“…Despite the importance of differential ChIP-seq analysis on group level, it remains a highly challenging computational task to reliably assess on a genome-wide scale the statistical significances of observed differences in ChIP-seq signal intensities between groups of samples, owing to the high level of noise and variability intrinsic to ChIP-seq data [6,14]. In general, the success of a group-level differential ChIP-seq analysis relies on a robust approach for normalizing multiple ChIP-seq samples, as well as a sophisticated statistical model for accurately assessing the variability of ChIP-seq signals across samples of the same group (referred to as within-group variability) [15,16].…”

Computational approaches for characterizing the tumor immune microenvironment

“…A comprehensive review of peak‐calling algorithms is available elsewhere . In addition to detection of enriched signals, a common goal of these experiments is differential peak analysis between biological conditions . Computational tools originally developed for RNA‐seq data are often used for this purpose, including DESeq2 and edgeR .…”