Ebayes Erorr: No Residual Degrees Of Freedom In Linear Model Fits.

Question

[Bingley]

Hi,

I got an error "No residual degrees of freedom in linear model fits" when I was using eBayes to look for the differentially expressed gene. The microarray data is 4 arrays for 4 conditions. I'd like to compare them against each other.

I understand this error might be the sample size, as I can see the $df.residual is a 0 vector which leads to this error.

$df.residual [1] 0 0 0 0 0 32316 more elements ...

$sigma [1] NA NA NA NA NA 32316 more elements ...

Is there any way around? Any other good way to estimate the differentially expressed genes in this kind of situation. Thank you very much.

Answer 1

[User 59]

I'm pretty sure you need replicates to make use of this approach, and you don't appear to have any.

See also:

https://stat.ethz.ch/pipermail/bioconductor/2009-May/027834.html

You'll have to return to looking at log2 fold changes between the samples.

Answer 2

[Bingley]

Many thanks, Daniel.