[BioC] limma's eBayes error: No residual degrees of freedom in linear model

[Gordon Smyth]

>[BioC] limma's eBayes error: No residual degrees of freedom in linear model
>Li,Qinghong,ST.LOUIS,Molecular Biology Qinghong.Li at rdmo.nestle.com
>Tue Nov 15 22:09:13 CET 2005
>
>Hi BioC,
>
>I was runing eBayes and got the above error. I searched the old archives 
>of BioC, and has found similar problem poseted by Ken Ninh: 
>http://files.protsuggest.org/biocond/html/4652.html
>
>I checked the summary(fit$df.residual), all zero's. But the 
>fit1<-lmFit(normData, design) and fit2<-contrasts.fit(fit1, cont.matrix) 
>ran properly. I checked normData with boxplots, and they looked fine and 
>well normalized. Here is my design matrix:
>  > design
>       C.M.1 C.F.2 C.M.3 C.F.3 R.M.1 R.F.2 R.M.3 R.F.3 (C/R: 
> control/treatment; F/M: male/female; 1,2,3 are sibling pairs)
>1     1     0     0     0     0     0     0     0
>2     0     1     0     0     0     0     0     0
>3     0     0     1     0     0     0     0     0
>4     0     0     0     1     0     0     0     0
>5     0     0     0     0     1     0     0     0
>6     0     0     0     0     0     1     0     0
>7     0     0     0     0     0     0     1     0
>8     0     0     0     0     0     0     0     1
>attr(,"assign")
>[1] 1 1 1 1 1 1 1 1
>attr(,"contrasts")
>attr(,"contrasts")$TBS
>[1] "contr.treatment"
>
>contrast matrix
>
> > cont.matrix
>       Diff
>C.M.1   -1
>C.F.2   -1
>C.M.3   -1
>C.F.3   -1
>R.M.1    1
>R.F.2    1
>R.M.3    1
>R.F.3    1
>
>What could be the possible reasons for the error and how to fix that?
>
>Thanks
>Johnny

Dear Johnny,

I have to tell you that what you're doing, i.e., the design matrix you've 
created, is not very sensible statistically. Hence the non-useful results 
you are getting from limma. Here are some steps that you can take to do 
something about it:

1. Consult someone with statistical experience at your organization who can 
tell you about replication and degrees of freedom for error.

2. To get meaningful help from this list, you need to explain a little more 
about your experiment. In particular you need to explain what you are 
hoping to learn scientifically from your data and what comparisons are of 
interest to you.

3. Explain what documentation you have read and what examples you are 
attempting to follow here. That would help us understand what you need to 
know, and may also help us to improve the documentation.

Best wishes
Gordon