README.md

In plbaldoni/mixHMM: Differential Peak Calling via Hidden Markov Model With Mixture of Negative Binomial Distributions

mixHMM

The goal of mixHMM is to detect differential peaks from ChIP-seq data of multiple conditions with replicates (under development).

Installation

You can install the released version of mixHMM from this repository with:

devtools::install_github("plbaldoni/mixHMM")

Example

# Loading example dataset 
data(H3K36me3)

# ChIP read counts, as well as the model offset, group indicator, and number of mixtures.
ChIP = assay(H3K36me3)
offset = ZIMHMM::createOffset(ChIP,method = 'loess',span = 1)
group = rep(1:3,each=3)
B = 2^(length(unique(group)))-2

# Calling peaks
peakcall = mixHMMConstr(ChIP = ChIP,offset = offset,group = group,B = B,control = controlPeaks(rep(1e-6,4))) 

# Plotting peak calls
plotPeaks(peakcall,ranges = c(which.min(abs(start(rowRanges(H3K36me3))-17800000)),which.min(abs(end(rowRanges(H3K36me3))-17910000))),
          ChIP = ChIP,group = group,offset = offset)

plbaldoni/mixHMM documentation built on Nov. 8, 2019, 8:05 p.m.