[BioC] edgeR GLM error

Gordon K Smyth smyth at wehi.EDU.AU
Thu Sep 20 01:38:40 CEST 2012


Dear Javier,

I don't see the slightest evidence of any error or problem.  In fact, you 
say that everything is running properly.

You might find it helpful to have a look at the posting guide:

  http:://www.bioconductor.org/doc/postingGuide.html

Best wishes
Gordon


On Wed, 19 Sep 2012, Javier Simón-Sánchez wrote:

> Hello Gordon,
>
> Again, thank you for your recent help on this. As I tyold you, I managed to
> un my analysis. However, I get the impression that the model is adjusting
> propoerly.
>
> I want to run Cases vs Controls controling for differences beteween brain
> fregions but IM getting ~300- associated p values after correcting for
> multiple testing. The code im using is:
>
> library(limma)
> library(edgeR)
> x <- read.delim("/myexpressiondata.txt",row.names="Symbol")
> group <- factor
> (c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0))
> d <- DGEList(counts=x,group=group)
> cpm.d <- cpm(d)
> d <- d[rowSums(cpm.d > 1) >= 4, ]
> d <- calcNormFactors(d)
> region <-
> factor(c("Caudate","Frontal","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal","Caudate","Frontal","Hippocampus","Putamen","Temporal"))
> design <- model.matrix(~region+d$samples$group)
> d <- estimateGLMCommonDisp(d, design,verbose=TRUE)
> d <- estimateGLMTrendedDisp(d, design)
> d <- estimateGLMTagwiseDisp(d, design)
> glmfit.tagwise.d <- glmFit(d,design,dispersion=d$tagwise.dispersion)
> lrt.tagwise.d <- glmLRT(glmfit.tagwise.d)
> topTags(lrt.tagwise.d)
>
>
>
> Thanks!
> On Thu, Sep 13, 2012 at 2:43 AM, Gordon K Smyth <smyth at wehi.edu.au> wrote:
>
>> Dear Javier,
>>
>> This error has been discused a number of times on this list.  The solution
>> is to upgrade edgeR to the current devel version.
>>
>> Also please see the Bioconductor posting guide:
>>
>>  http://www.bioconductor.org/**help/mailing-list/posting-**guide/<http://www.bioconductor.org/help/mailing-list/posting-guide/>
>>
>> Best wishes
>> Gordon
>>
>>  Date: Tue, 11 Sep 2012 13:04:55 +0200
>>> From: Javier Sim?n-S?nchez <simonsanchezj at gmail.com>
>>> To: bioconductor at r-project.org
>>> Subject: [BioC] edgeR GLM error
>>>
>>> Hello,
>>> My name is Javier Simón Seanchez and I'm a post-doc at the VUmc in
>>> Amsterdam.
>>>
>>> The reason of this e-mail is that im running edgeR in an expression
>>> dataset
>>> and getting the following error when calculating the GLM common
>>> dispersion:
>>>
>>> *Error in beta[k, ] <- betaj[decr, ] :
>>>  NAs are not allowed in subscripted assignments
>>> *
>>>
>>> Im running cases versus controls and I want to modulate for different
>>> tissues. How can I overcome this error?
>>>
>>> Thanks a lot in advance
>>>
>>> --
>>> Javier Simon-Sanchez
>
> -- 
> Javier Simon-Sanchez

______________________________________________________________________
The information in this email is confidential and intend...{{dropped:5}}


More information about the Bioconductor mailing list