May I clarify a part of the DESeq2 package? Is it true that DESeq2 counts for linear (only) transcripts and excludes any possibility in circular RNA (circRNA)? In which, an increase/decrease in circRNA has no effect on the number of counts produced in the results/output? Or does DESeq2 not able to distinguish the two, will assume circRNA as linear fragments, and include it into the counts?
DESeq2 has not concept of linear vs circular transcripts. You provide the software with a matrix of counts and it compares across samples, adjusting for e.g. library size.
If you used Rsubread featureCounts to generate counts, the mapped reads are just assigned to overlapping genes. It doesn't matter whether the genes are located on a linear sequence or on a circular sequence.
This is good information. So what you mean is that Rsubread will use the annotations (from the gtf file) to gather information of a location, and count the number of reads that cover that area yes? In which, the reads could encompass either linear or circular RNA? Am I right to say that if the DESeq2 shows like an increased log2foldchange in a treated condition (as compared to a control), it could mean either the linear RNA is increasing, or both linear & circRNA? I know this might sound silly but don't mind me asking - Will there be any hypothetical idea that the log2foldchange would stay consistent (linear RNA does not change through qRT-PCR) but circRNA is actually increasing (let's consider the scenario that the circRNA has a robust expression)?
Thank you.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
That I can understand. I believe the Rsubread is the one featuring the reads. But wanted to clarify at least before I lay the information on my colleagues.
Thanks by the way.