bioinformatics: Mo allocation DESeq nbinomTest

[coutellec]

Dea all,

I encounter a memory problem with nbinomTest in DESeq: R is unable to allocate 221.1 Mo to a vector.

My dataset is 48000 contigs x 16 libraries.

I tried several computers, without success.
Thanks in advance for any help on this,
Cheers
Marie-Agnès

[Simon Anders]

Do you maybe have some really large numbers in your data? If so, try using the GLM approach instead of nbinomTest.

[pinki999]

Hi,

I too have the same problem. I am working with 50 replicates for each condition and I encounter memory allocation problem with nbinomTest. How exactly can this GLM approach be used here?

Thanks

[Simon Anders]

See "example(nbinomGLMTest)" for an example:

cds <- newCountDataSet( countTable, conditions )
cds <- estimateSizeFactors( cds )
cds <- estimateDispersions( cds, method="pooled" )
fit1 <- fitNbinomGLMs( cds, count ~ condition )
fit0 <- fitNbinomGLMs( cds, count ~ 1 )
pvals <- nbinomGLMTest( fit1, fit0 )
padj <- p.adjust( pvals, "BH" )