smoothRatiosByChromosome: Backgound correction and signal smoothing per chromosome

In jianhong/NADfinder: Call wide peaks for sequencing data

Backgound correction and signal smoothing per chromosome

Description

Split the ratios by chromosome and do background correction and signal smoothing.

Usage

smoothRatiosByChromosome(
  se,
  chr = paste0("chr", c(seq_len(21), "X", "Y")),
  ratioAssay = "ratio",
  backgroundCorrectedAssay = "bcRatio",
  smoothedRatioAssay = "smoothedRatio",
  zscoreAssay = "zscore",
  backgroundPercentage = 0.25,
  chrom.level.background = TRUE,
  ...
)

Arguments

se	An object of RangedSummarizedExperiment with log2-transformed ratios, CPMRatios or OddRatios. Output of log2se
chr	A character vector, used to filter out seqnames. It should be the chromosome names to be kept.
ratioAssay	The name of assay in se, which store the values (log2-transformed ratios, CPMRatios or OddRatios) to be smoothed.
backgroundCorrectedAssay, smoothedRatioAssay, zscoreAssay	character(1). Assays names for background corrected ratios, smoothed ratios and z-scores based on background corrected ratios.
backgroundPercentage	numeric(1). Percentage of values for background, see zscoreOverBck. The percentage of values lower than this threshold will be treated as background, with 25 percentile as default.
chrom.level.background	logical(1): TRUE or FALSE, default to TRUE, use chromosome-level background to calculate z-score
...	Parameters could be passed to butterFilter.

Value

A SimpleList of RangedSummarizedExperiment with smoothed ratios.

Author(s)

Jianhong Ou, Haibo Liu and Julie Zhu

Examples

data(single.count)
se <- single.count
dat <- log2se(se, nucleolusCols="N18.subsampled.srt.bam", genomeCols="G18.subsampled.srt.bam", 
transformation="log2CPMRatio")
dat1 <- smoothRatiosByChromosome(dat, N=100, chr = c("chr18", "chr19"))
dat2 <- smoothRatiosByChromosome(dat, N=100, chr = c("chr18", "chr19"), 
                                 chrom.level.background = FALSE)

jianhong/NADfinder documentation built on Oct. 29, 2023, 11:08 a.m.