differentialNbinomWaldTestUMI4C: Differential UMI4C contacts using DESeq2 Wald Test

In UMI4Cats: UMI4Cats: Processing, analysis and visualization of UMI-4C chromatin contact data

Description

Using a UMI4C object, infers the differences between conditions specified in design of the smooth monotone fitted values using a Wald Test from DESeq2 package.

Usage

differentialNbinomWaldTestUMI4C(
  umi4c,
  design = ~condition,
  normalized = TRUE,
  padj_method = "fdr",
  query_regions = NULL,
  padj_threshold = 0.05,
  penalty = 0.1,
  alpha = 20
)

Arguments

umi4c	UMI4C object as generated by makeUMI4C or the UMI4C constructor.
design	A formula or matrix. The formula expresses how the counts for each fragment end depend on the variables in colData. See DESeqDataSet.
normalized	Logical indicating if the function should return normalized or raw UMI counts. Default: TRUE.
padj_method	The method to use for adjusting p-values, see p.adjust. Default: fdr.
query_regions	GRanges object or data.frame containing the coordinates of the genomic regions you want to use to perform the analysis in specific genomic intervals. Default: NULL.
padj_threshold	Numeric indicating the adjusted p-value threshold to use to define significant differential contacts. Default: 0.05.
penalty	Amount of smoothing to be applied to the estimated functional parameter. Default: 0.1.
alpha	Approximate number of fragments desired for every basis function of the B-spline basis. floor((max(number of fragments)) / alpha) is passed to create.bspline.basis as nbasis argument. 4 is the minimum allowed value. Default: 20.

Details

This function fits the signal trend of a variance stabilized count values using a symmetric monotone fit for the distance dependency. Then scales the raw counts across the samples to obtain normalized factors. Finally, it detects differences between conditions applying the DESeq2 Wald Test.

Value

UMI4C object with the DESeq2 Wald Test results.

Examples

 files <- list.files(system.file("extdata", "CIITA", "count", package="UMI4Cats"),
                     pattern = "*_counts.tsv.gz",
                     full.names = TRUE
 )
# Create colData including all relevant information
colData <- data.frame(
  sampleID = gsub("_counts.tsv.gz", "", basename(files)),
  file = files,
  stringsAsFactors = FALSE
)

library(tidyr)
colData <- colData %>%
  separate(sampleID,
           into = c("condition", "replicate", "viewpoint"),
           remove = FALSE
  )
  
# Make UMI-4C object including grouping by condition
umi <- makeUMI4C(
  colData = colData,
  viewpoint_name = "CIITA",
  grouping = NULL,
  bait_expansion = 2e6
)

umi_wald <- differentialNbinomWaldTestUMI4C(umi4c=umi,
                                            design=~condition,
                                            alpha = 100)

UMI4Cats documentation built on Dec. 31, 2020, 2:01 a.m.