[BioC] lmfit "No residual degrees of freedom"

[David martin]

Ok thanks.

On 01/17/2011 02:25 PM, Sean Davis wrote:
> On Mon, Jan 17, 2011 at 8:09 AM, David martin<vilanew at gmail.com>  wrote:
>
>> Hello,
>> I'm having problems computing comparisons between different groups (this is
>> not array data and i'm not using limma). Any on what the error must be ?
>>
>>> mydata.eset
>> ExpressionSet (storageMode: lockedEnvironment)
>> assayData: 2754598 features, 3 samples
>>   element names: exprs
>> protocolData: none
>> phenoData
>>   sampleNames: sampleA sampleB sampleC
>>   varLabels: Group Sample
>>   varMetadata: labelDescription
>> featureData: none
>> experimentData: use 'experimentData(object)'
>> Annotation:
>>
>>
>> fit<- lmFit(mydata.eset, design)
>> ....
>> contrast.matrix
>>> contrast.matrix
>>                     group1-group2
>> group1                                   1
>> group2                                      -1
>> group3                                       0
>>                     group1-group3
>> group1                                   1
>> group2                                       0
>> group3                                      -1
>>                     group2-group3
>> group1                               0
>> group2                                   1
>> group3                                  -1
>>
>> fit2<- contrasts.fit(fit, contrast.matrix)
>> fit2<- eBayes(fit2)
>> Error in ebayes(fit = fit, proportion = proportion, stdev.coef.lim =
>> stdev.coef.lim) :
>>   No residual degrees of freedom in linear model fits
>>
>>
> Hi, David.
>
> You have three samples and three conditions.  That is not enough samples to
> perform the tests that you want to perform.  Your only recourse here,
> really, is to make ratios among your three arrays and rank genes based on
> fold change.  Alternatively, you can go run some biological replicates.
>
> Sean
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
>