Obtain Levene residuals

Share:

Description

Obtain absolute or squared Levene residuals for each CpG given a series of methylation arrays

Usage

1

Arguments

data

object of class matrix of M values, with rows corresponding to features of interest such as CpG sites and columns corresponding to samples or arrays

design

the design matrix of the experiment, with rows corresponding to arrays/samples and columns to coefficients to be estimated. Defaults to the unit vector.

type

character string, "AD" for absolute residuals or "SQ" for squared residuals. Default is "AD".

Details

This function will return absolute or squared Levene residuals given a matrix of M values and a design matrix. This can be used for graphing purposes or for downstream analysis such a gene set testing based on differential variability rather than differential methylation. If no design matrix is given, the residuals are determined by treating all samples as coming from one group.

Value

Returns a list with three components. data contains a matrix of absolute or squared residuals, AvgVar is a vector of sample variances and LogVarRatio corresponds to the columns of the design matrix and is usually the ratios of the log of the group variances.

Author(s)

Belinda Phipson

References

Phipson, B., and Oshlack, A. (2014). A method for detecting differential variability in methylation data shows CpG islands are highly variably methylated in cancers. Genome Biology, 15:465.

See Also

varFit

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
# Randomly generate data for a 2 group problem with 100 CpG sites and 5 arrays in each group
y <- matrix(rnorm(1000),ncol=10)

group <- factor(rep(c(1,2),each=5))
design <- model.matrix(~group)

# Get absolute Levene Residuals
resid <- getLeveneResiduals(y,design)

# Plot the first CpG
barplot(resid$data[1,],col=rep(c(2,4),each=5),ylab="Absolute Levene Residuals",names=group)