Quantification of transcription factor footprints in chromatin accessibility data, using the annotation of methylTFR.
Tn5 insertion densities replace methylation levels. For every binding site of a motif the insertions are counted per base and stacked on the site centre, and the deviation score is the insertion density of the central window over the density of the outer flanks, corrected by the same ratio on an expected profile. The expected profile comes either from the GC bin composition of the binding sites, as in methylTFR, or from a k-mer model of the Tn5 sequence preference.
Installation
remotes::install_github("EpigenomeInformatics/chromTFR")Reading a dataset requires ChrAccR, and the k-mer model requires a BSgenome package. Both are suggested rather than required, so the package installs without them.
Usage
library(chromTFR)
dsa <- loadAccDataset("dsATAC_filtered")
peaks <- getAccRegions(dsa, regionType = ".peaks.cons", extend = 500)
ins <- getTn5Insertions(dsa, getAccSamples(dsa)[1], regions = peaks)
accDeviationScore(accProfile(ins, prepareTFBS(tf_bindsites[["CTCF"]])))Citation
citation("chromTFR")The vignette A worked example with the ChrAccR example data runs the whole workflow on the public ChrAccRex dataset, from the DsATAC object to the footprints.
See the vignettes for the full workflow, and system.file("scripts", package = "chromTFR") for template scripts covering the steps from a dataset to the figures.