Hi,

I am trying to study the enrichment of transcripts at different subcellular fractions in E. coli. I collected three biological replicates, and for each replicate, I have the following samples:

SampleA: total RNA. SampleB: RNA isolated from specific fractions obtained by resolving sample A in sucrose gradients. SampleC: RNA obtained by pulling down an RNA binding protein of interest from sample B.

Briefly, in order to quantify the relative enrichment of each transcript within samples B and C, I plan to compare the transcriptomes of these samples to the total RNA by running deseq2 of sampleB vs sampleA and sampleC vs sampleA. The log2 fold-change would indicate the relative enrichment of each gene in samples B and C in relation to sampleA. Would you agree with this?

Furthermore, I am interested in comparing the enrichment of transcripts in sampleB vs those in sampleC. Specifically, I am interested in identifying transcripts that are underrepresented in sampleC in relation to sampleB. Considering that these two samples are not different experimental timepoints or conditions but different fractions, would running deseq2 of sampleB vs sampleC be suitable for this comparison? If not, could anyone suggest an alternative approach to tackle this second task?

Thanks.

Mikel

This has been asked a few times before on the support forum, but I can imagine it's difficult to search for the same question. The crux is this:

(C - A) - (B - A) = C - B

There is no point including the A samples if they are the same for C and for B. It reduces to simply C - B.