The domainCalling package is a R/Bioconductor package, and it has been tested on R 3.4.0/Bioc 3.5. This package detects broad range of non-coding transcripts (from total RNA-seq) or binding regions, particularly from CutAndRun-seq (CAR-seq) data.
You need the csaw, edgeR and ChIPseeker packages. A TxDb and compatible BSgome org.xx.db package for annotation.
This function uses window-based counting method provided by the csaw package. It first counts reads overlapping with a sliding window from non-background and background (if there is any) samples. Secondly it filters out uninteresting windows if the average abounance of the non-background samples is (1) lower then 3 fold change to the backgroud or (2) does not exceed the threshold. Finally, the retained windows are merged with neighbors within designated perimeters. The merged regions is the potential binding regions of the protein.
Providing a data.frame (or DataFrame) containing three columns named sample_name
, file_bam
, and spike_bam
the user can
use getSampleInfo() to find information from the bam files including fragment size, library size, spike-in counts,
single-ended or pair-ended. These inforamtion is needed to run csawDomainCalling().
Need to add an example here
The package provides a tool to estimate the normalization factor using spike-in sequence data.