Hello!

I have some problems with run_edgeR or run_DESeq2 functions from systempipeR package. I 've done all required steps described in vignette and  when I've tried to do run_edgeR by command:

edgeDF <- run_edgeR(countDF = countDF, targets=targets, cmp=cmp[[1]], independent=FALSE, mdsplot="")

I've had an error like this:

   Error in `[.data.frame`(countDF, , names(samples)) : 
    undefined columns selected

Maybe, I've created targets.txt file in wrong way or somthing else. So, I have no problems using my set of BAM files  in DESeq2 package and edgeR package, but I need to use run_edgeR (or run_DESeq2) from systempipeR because I want to use GO term enrichment analysis of DEGs from this package.

Thank you in advance,

Alex Gopanenko

Yes, most likely something is wrong with your count object (countDF) and/or targets file. BTW, if you just want to run the GO analysis then all you need is a list containing character vectors of gene set labels and each list component has some name meaningful to you. 

To identify the source of your current error, you may want to run the following checks:

(1) Check whether all samples in cmp[[1]] exist in Factor column of your targets file (data.frame).

targets$Factor %in% unique(as.character(cmp[[1]]))

All of them need to be TRUE.

(2) Check whether all column names in your countDF exist in SampleName column of your targets file (data.frame).

colnames(countDF) %in% targets$SampleName

(3) Check whether your countDF contains the feature names in the rowname slot and the columns contain count values. For this you could just inspect the first few rows:

head(countDF)

Thomas