Limma package for DNA copy number data
Entering edit mode
djtml ▴ 10
Last seen 6.0 years ago
United Kingdom


I have a question about using the Limma package, lmfit() and ebayes in R which I was hoping someone could answer.

Briefly, I want to use the Limma package and the eBayes fit and BH p.value adjust method to be able to identify differences in DNA copy number values for a particular gene between ovarian cancer cell lines which have a mutated or normal copy of another gene (for example is there a copy number difference in gene X between cell lines which are mutant or wild type for gene Y?).

The code seems to run and produces a table containing p.values, adjusted p.values etc.. but a warning message appears stating:

"In ebayes(fit = fit, proportion = proportion, stdev.coef.lim = stdev.coef.lim, : Estimation of var.prior failed - set to default value"

I've used this approach for differential gene expression analysis and it worked just fine

My question(s) is this: why is this warning message produced for these datasets, and can the output results be trusted? Is there another approach for this type of analysis that I should consider?


R limma • 914 views
Entering edit mode

Hi! Did you finally use this approach? Did it work? I'm interested in applying it.

Entering edit mode
Aaron Lun ★ 28k
Last seen 13 hours ago
The city by the bay

I don't think you have to worry about it. The var.prior data object is only required to compute the B-statistics, i.e., the log-odds of differential expression for each gene. You'll probably be using the BH-adjusted p-values to determine significance, and they should be unaffected.


Login before adding your answer.

Traffic: 393 users visited in the last hour
Help About
Access RSS

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

Powered by the version 2.3.6