[BioC] issue with lmFit in limma

[James W. MacDonald]

Hi Aaron,

Skewes,Aaron wrote:
> Hi,
> 
> I am attempting to process a few Agilent 4x44 based arrays using limma. When I try to estimate expression levels with lmFit in limma, I get the following message:
> 
> Warning message:
> In lmFit(MA, method = "robust") :
> Some coefficients not estimable: coefficient interpretation may vary.
> 
> And all the coeff's are NA. I am not at all sure what the problem is? Can someone give me an idea?

All of the coefficients are NA? Not just some of them?

The code you show below doesn't jibe with the warning you show above 
(method = "ls" is the default, so lmFit(MA) should never result in a 
warning that quotes method = "robust").

I am not sure you have given us enough information. I think at the least 
we need to see the actual input/output stream from R when you run your 
code, as well as the number of samples, coefficients you are trying to 
estimate, and a sessionInfo().

Best,

Jim


> 
> Code snipet:
> 
> #Normalize data with background subtraction
>         MA<-normalizeWithinArrays(RG.Nctr , method="loess", bc.method="rma" )
>         MA<-normalizeBetweenArrays(MA , method="quantile")
> 
>         #get exp values
>         fit <- lmFit(MA)
> 
> Thanks,
> Aaron
> 
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald, M.S.
Biostatistician
Hildebrandt Lab
8220D MSRB III
1150 W. Medical Center Drive
Ann Arbor MI 48109-0646
734-936-8662