Hi All,

I have been using using DESeq2 to find differentially expressed genes between healthy and diseased groups but I do not have house keeping genes in the data set. Is it still possible to do DESeq calculations without house keeping genes? and I do not have all the genes expressed across all subjects/samples (about ~4,000 genes in the data, at least 1 gene has a sample with zero counts). Sequencing depths for my libraries vary between 1M and 25M reads. Does DESeq handle such huge variation in sequencing depths and also imbalance in the number of samples between groups (5 healthy and 7 diseased subjects)? 

I have the following error message:

> dds <- DESeq(dds)

estimating size factors

Error in estimateSizeFactorsForMatrix(counts(object), locfunc = locfunc,  : 

every gene contains at least one zero, cannot compute log geometric means

Another question - is it possible that some genes are turned on exclusively in healthy but not in diseased subjects and vice-versa. Is there a way to look for such genes as well?

Thank you for your help.

Gajender

What kind of data is this? Can you say more? It’s not typical that there would be no genes with positive counts across 12 samples, even with sequencing depth of 1 million for the lowest sequenced sample.