getAvgProfileWithCI: Compute the average profile and its confidence interval for a...

In pgpmartin/GeneNeighborhood: Analyze the Neighborhood (Upstream/Downstream Neighbors) of Genes

Description

Compute the average profile and its confidence interval for a selection of features

Usage

getAvgProfileWithCI(
  FeatureProfiles,
  selFeatures = NULL,
  pos = NULL,
  conf = 0.95,
  type = c("mean", "freq", "avgnorm", "sum", "n")
)

Arguments

FeatureProfiles	Matrix (features in row, position in columns) or RleList with coverage profiles for a set of features
selFeatures	Character or integer vector defining a selection of features from FeatureProfiles
pos	(Optional) Integer vector with values defining the genomic positions represented in FeatureProfiles
conf	Numeric in ]0,1] representing the confidence level used to compute the confidence interval
type	Character string indicating the method used to aggregate values from all the selFeatures. Methods are: "mean" (Default). Compute the average value at each position. "freq" Compute the proportion of signal at each position. Sum of "freq" signal accross positions is 1. "avgnorm" Same as "freq" but mutiplied by the number of positions. Mean of "avgnorm" signal accross positions is 1. "sum" Compute the sum of signal at each position. "n" Compute the number of non missing observations at each genomic positions.

Value

a data frame with 4 columns:

• Position: pos values

• Profile: values obtained by aggregating as defined by type on all selFeatures

• Upper: Upper bound of the confidence interval of Profile

• Lower: Lower bound of the confidence interval of Profile

Examples

## Extract the profiles around (+/-50bp) the first 200 genes. We bin the genes in 3 bins only.
  top200Prof <- annotationCoverageAroundFeatures(Genegr,
                                                 features=1:200,
                                                 sidedist = 50,
                                                 usePercent = TRUE,
                                                 nbins=3)
## Assemble the metagene profiles
  Prof <- assembleProfiles(top200Prof)
## Select a set of 50 random genes within the top200
  set.seed(123)
  randGenes <- sample(names(Genegr)[1:200], 50)
##Get the average profile for these genes
  avgProf_sense <- getAvgProfileWithCI(Prof$Profiles_Sense,
                                       selFeatures = randGenes,
                                       pos = c(-50:0, 1:3, 0:50))
  head(avgProf_sense)
## A quick look at the profile:
  plot(1:105, avgProf_sense$Profile,
       type="l", axes = FALSE,
       xlab="Position", ylab="Percent")
  lines(1:105, avgProf_sense$Upper, lty=2, col="grey")
  lines(1:105, avgProf_sense$Lower, lty=2, col="grey")
  axis(side=2)
  axis(side=1, at=c(1,51,55,105),
       labels=c("-50bp", "TSS", "TES", "+50bp"),
       las=2)

pgpmartin/GeneNeighborhood documentation built on Sept. 2, 2021, 6:37 a.m.