Software for motif discovery and ChIP-Seq analysis
Finding Peaks and Differential Peaks with Replicate ExperimentsThe following outlines HOMER's recommended approach to identifying peaks that are statistically enriched across replicates. As you might expect, the use of replicate experiments is highly encouraged and can help reduce the number of false positives in your data. This page covers both the initial identification of peaks (i.e. IP vs. input) and also the identification of differential peaks (IP vs. IP).
There are two general (but related) approaches to identifying differential peaks. The first is to use the getDifferentialPeaksReplicates.pl command, which attempts to automate the steps described below into a single command to generate a peak file. The second is to prepare your own regions and read counts and then use getDiffExpression.pl directly to calculate the differential enrichment. That's covered in more detail (along with differential expression) in the next section, and is recommended if you want a little more fine control on how things are quantified. In both cases, HOMER uses R/Bioconductor and DESeq2 (by default) to perform the differential enrichment calculations.
Looking for overlapping peaks (i.e. Venn Diagram) or differential peaks without replicates:
The information below is catered to the analysis of peaks using replicate experiments. For more basic operations/analysis without replicates, check out the page covering the mergePeaks and getDifferentialPeaks programs.
Required 3rd party software:
Can't figure something out? Questions, comments, concerns, or other feedback: