adjusting for batch using ComBat (sva) - singularity error
0
0
Entering edit mode
@annebozack-8680
Last seen 4.9 years ago
United States

I'm adjusting for batch effects in 450K data using ComBat in the sva package. When I run ComBat, I get the following error:

> batch = factor(pheno$plate)
> modcombat = model.matrix(~1, data=pheno)
> methCombat = ComBat(dat = as.matrix(meth), batch = batch, mod = modcombat)

Found7batches
Adjusting for0covariate(s) or covariate level(s)
Found216170Missing Data Values 
Standardizing Data across genes
Error in solve.default(crossprod(des), crossprod(des, y1)) : 
  Lapack routine dgesv: system is exactly singular: U[4,4] = 0

The error arises due to the treatment of NAs in the methylation dataset when ComBat fits a model in the presence of missing values using the Beta.NA function. Specifically, the error is caused by this part of the ComBat function:

cat("Standardizing Data across genes\n")
if (!NAs) {
    B.hat <- solve(crossprod(design), tcrossprod(t(design), 
        as.matrix(dat)))
} else {
    B.hat <- apply(dat, 1, Beta.NA, design)
}

Other users have found this error due to their data structure; however, I do not get the error when running the ComBat after removing incomplete rows from the methylation dataset:

> methComplete = meth[complete.cases(meth),]
> methCombat = ComBat(dat = as.matrix(methComplete), batch = batch, mod = modcombat)

Found7batches
Adjusting for0covariate(s) or covariate level(s)
Standardizing Data across genes
Fitting L/S model and finding priors
Finding parametric adjustments
Adjusting the Data

Any suggestions would be appreciated!

microarray combat sva batch effects • 2.4k views
ADD COMMENT

Login before adding your answer.

Traffic: 864 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6