Hello!
I am trying to run DESeq2 using HTseq count as input but I am running in the following error every time, I already checked and input data is correct.
Error in Ops.factor(a$V1, l[[1]]$V1) :
level sets of factors are different
Calls: DESeqDataSetFromHTSeqCount → sapply → sapply → lapply → FUN → Ops.factor
I have SINGLE file per treatment (two treatments) and they have same number of lines (outpu htseq)
Thank you,
Arpita
Hey Arpita,
Its hard to say, what the reason is. Please make an error report or share your history here. I assume that there is something wrong with your factor/level setup.
Hey,
I somehow managed to get over that error in DESeq2 but now I am getting the following as new error:
2 samples with counts over 18838 genes
using disperion fit type: parametric
estimating size factors Error in checkForExperimentalReplicates(object, modelMatrix) : The design matrix has the same number of samples and coefficients to fit, so estimation of dispersion is not possible. Treating samples as replicates was deprecated in v1.20 and no longer supported since v1.22. Calls: DESeq → DESeqParallel → checkForExperimentalReplicates
Thank you,
Hey, please look at this post: Bioconductor Forum
Basically you have only one file per condition, which is problematic for DESeq because you cannot estimate the variation anymore.
So in order to fix this you can (1) use a edgeR or limma, (2) redefine your conditions in DESeq2, or (3) include more data.
Hey,
Thanks for such a prompt response.
I also wanted to know if I can do GSEA directly without doing DESeq2?
Regards,
Arpita
You would need to normalize your data first, which you can to with bamcoverage (see this tutorial). If you normalized your data then you would probably calculate the log2 fold-change. With that you would have an idea about the enrichment between your conditions. I am not sure if Galaxy has a tool for it.