RUVfit: Remove unwanted variation when testing for differential...

In Oshlack/missMethyl: Analysing Illumina HumanMethylation BeadChip Data

Remove unwanted variation when testing for differential methylation

Description

Provides an interface similar to lmFit from limma to the RUV2, RUV4, RUVinv and RUVrinv functions from the ruv package, which facilitates the removal of unwanted variation in a differential methylation analysis. A set of negative control variables, as described in the references, must be specified.

Usage

RUVfit(
  Y,
  X,
  ctl,
  Z = 1,
  k = NULL,
  method = c("inv", "rinv", "ruv4", "ruv2"),
  ...
)

Arguments

Y	numeric matrix with rows corresponding to the features of interest such as CpG sites and columns corresponding to samples or arrays.
X	The factor(s) of interest. A m by p matrix, where m is the number of samples and p is the number of factors of interest. Very often p = 1. Factors and dataframes are also permissible, and converted to a matrix by design.matrix.
ctl	logical vector, length == nrow(Y). Features that are to be used as negative control variables are indicated as TRUE, all other features are FALSE.
Z	Any additional covariates to include in the model, typically a m by q matrix. Factors and dataframes are also permissible, and converted to a matrix by design.matrix. Alternatively, may simply be 1 (the default) for an intercept term. May also be NULL.
k	integer, required if method is "ruv2" or "ruv4". Indicates the number of unwanted factors to use. Can be 0.
method	character string, indicates which ruv method should be used.
...	additional arguments that can be passed to RUV2, RUV4, RUVinv and RUVrinv. See linked function documentation for details.

Details

This function depends on the ruv package and is used to estimate and adjust for unwanted variation in a differential methylation analysis. Briefly, the unwanted factors W are estimated using negative control variables. Y is then regressed on the variables X, Z, and W. For methylation data, the analysis is performed on the M-values, defined as the log base 2 ratio of the methylated signal to the unmethylated signal.

Value

A list containing:

betahat	The estimated coefficients of the factor(s) of interest. A p by n matrix.
sigma2	Estimates of the features' variances. A vector of length n.
t	t statistics for the factor(s) of interest. A p by n matrix.
p	P-values for the factor(s) of interest. A p by n matrix.
Fstats	F statistics for testing all of the factors in X simultaneously..
Fpvals	P-values for testing all of the factors in X simultaneously.
multiplier	The constant by which sigma2 must be multiplied in order to get an estimate of the variance of betahat.
df	The number of residual degrees of freedom.
W	The estimated unwanted factors.
alpha	The estimated coefficients of W.
byx	The coefficients in a regression of Y on X (after both Y and X have been "adjusted" for Z). Useful for projection plots.
bwx	The coefficients in a regression of W on X (after X has been "adjusted" for Z). Useful for projection plots.
X	X. Included for reference.
k	k. Included for reference.
ctl	ctl. Included for reference.
Z	Z. Included for reference.
fullW0	Can be used to speed up future calls of RUVfit.
include.intercept	include.intercept. Included for reference.
method	Character variable with value indicating which RUV method was used. Included for reference.

Author(s)

Jovana Maksimovic

References

Gagnon-Bartsch JA, Speed TP. (2012). Using control genes to correct for unwanted variation in microarray data. Biostatistics. 13(3), 539-52. Available at: http://biostatistics.oxfordjournals.org/content/13/3/539.full.

Gagnon-Bartsch, Jacob, and Speed. 2013. Removing Unwanted Variation from High Dimensional Data with Negative Controls. Available at: http://statistics.berkeley.edu/tech-reports/820.

See Also

RUV2, RUV4, RUVinv, RUVrinv, topRUV

Examples

if(require(minfi) & require(minfiData) & require(limma)) {
# Get methylation data for a 2 group comparison
meth <- getMeth(MsetEx)
unmeth <- getUnmeth(MsetEx)
Mval <- log2((meth + 100)/(unmeth + 100))
group <- factor(pData(MsetEx)$Sample_Group)
design <- model.matrix(~group)
# Perform initial analysis to empirically identify negative control features 
# when not known a priori
lFit <- lmFit(Mval,design)
lFit2 <- eBayes(lFit)
lTop <- topTable(lFit2,coef=2,num=Inf)
# The negative control features should *not* be associated with factor of 
# interest but *should* be affected by unwanted variation 
ctl <- rownames(Mval) %in% rownames(lTop[lTop$adj.P.Val > 0.5,])
# Perform RUV adjustment and fit
fit <- RUVfit(Y=Mval, X=group, ctl=ctl)
fit2 <- RUVadj(Y=Mval, fit=fit)
# Look at table of top results
top <- topRUV(fit2)
}

Oshlack/missMethyl documentation built on March 26, 2023, 1:50 p.m.