getCrossCorrelationScores: QC-metrics from cross-correlation profile, phantom peak and...

In carmencita/ChIC: Quality Control Pipeline for ChIP-Seq Data

Description

We use cross-correlation analysis to obtain QC-metrics proposed for narrow-binding patterns. After calculating the strand cross-correlation coefficient (Kharchenko et al., 2008), we take the following values from the profile: coordinates of the ChIP-peak (fragment length, height A), coordinates at the phantom-peak (read length, height B) and the baseline (C), the strand-shift, the number of uniquely mapped reads (unique_tags), uniquely mapped reads corrected by the library size, the number of reads and the read lengths. We calculate different values using the relative and absolute height of the cross-correlation peaks: the relative and normalized strand coefficient RSC and NSC (Landt et al., 2012), and the quality control tag (Marinov et al., 2013). Other values regarding the library complexity (Landt et al., 2012) like the fraction of non-redundant mapped reads (NRF; ratio between the number of uniquely mapped reads divided by the total number of reads), the NRF adjusted by library size and ignoring the strand direction (NRF_nostrand), and the PCR bottleneck coefficient PBC (number of genomic locations to which exactly one unique mapping read maps, divided by the number of unique mapping reads).

getCrossCorrelationScores

Usage

getCrossCorrelationScores(
  data,
  bchar,
  annotationID = "hg19",
  read_length,
  savePlotPath = NULL,
  mc = 1,
  tag = "ChIP"
)

Arguments

data	data-structure with tag information read from bam file (see readBamFile())
bchar	binding.characteristics is a data-structure containing binding information for binding preak separation distance and cross-correlation profile (see spp::get.binding.characteristics).
annotationID	String, indicating the genome assembly (Default="hg19")
read_length	Integer, read length of "data" (Defaul="36")
savePlotPath	if set the plot will be saved under "savePlotPath". Default=NULL and plot will be omitted.
mc	Integer, the number of CPUs for parallelization (default=1)
tag	String,can be used to personalize the prefix of the filename for the cross- correltion plot (default="ChIP" and "Input" in case of cross-correlation plot for the input')

Value

finalList List with QC-metrics

Examples

## This command is time intensive to run

## To run the example code the user must provide a bam file and read it 
## with the readBamFile() function. To make it easier for the user to run 
## the example code we provide a bam file in our ChIC.data package that has 
## already been loaded with the readBamFile() function.

mc=4
print("Cross-correlation for ChIP")
## Not run: 
filepath=tempdir()
setwd(filepath)
data("chipSubset", package = "ChIC.data", envir = environment())
chipBam=chipSubset

## calculate binding characteristics 

chip_binding.characteristics<-spp::get.binding.characteristics( chipBam, 
    srange=c(0,500), bin = 5, accept.all.tags = TRUE)

crossvalues_Chip<-getCrossCorrelationScores( chipBam , 
    chip_binding.characteristics, read_length = 36, 
    annotationID="hg19",
    savePlotPath = filepath, mc = mc)

## End(Not run)

carmencita/ChIC documentation built on April 28, 2021, 7:20 p.m.