I have 40 liver cancer samples. Among them I have two groups. GroupA (26 samples) and GroupB (14 samples) and I have a total of 10,000 lncRNA transcripts. I'm trying to do differential analysis between GroupA vs GroupB using
With edgeR analysis I see that there are around 300 upregulated lncRNA transcripts in GroupA vs GroupB. But I see that all these upregulated lncRNA transcripts are not expressed in all GroupA samples (highly expressed only in few samples), and moreover I see that these are not at all expressed in GroupB which means zero counts in GroupB.
I took a specific lncRNA transcript and made a box plot which looks like below:
There are many other upregulated lncRNA transcripts expressed only in few samples of GroupA and no expression in GroupB for differential analysis between GroupA vs GroupB.
What I should do now? Can I carry on with further analysis with those 300 upregulated lncRNA transcripts? Or Is there any other way to get true differentially expressed genes?