Question: DESeq2: Tests of log2 differences
gravatar for fabian
2.4 years ago by
fabian20 wrote:

Dear all

I have sequencing data reflecting the amount of nascent mRNA and I have from the same biological samples data reflecting the polyA fraction. both for two conditions (treated and unteated). I plan to do differential gene expression analysis on the nascent RNA-seq (to get an idea of differential transcription). In a next step I would like to test the polyA fraction sequencing data if they differ from the nascent RNAseq data, which could be indicative of differential decay/stability. To this end I do not want to test the hypothesis that the log2FC(polyA gene) != 0 but log2FC(polyA gene) != log2FC(nascent gene).

I did find the  lfcThreshold, altHypothesis options. But I can not provide a lfcThreshold for each gene individually. Furthermore I do not have an unequal altHypothesis.

I thought about doing it with contrasts but could not figure out how. If I would do something like '~ nascent + condition' I would fake additional replicas.

Any idea how to do this properly is highly welcomed. Thank you very much in advance.

ADD COMMENTlink modified 2.4 years ago • written 2.4 years ago by fabian20

Dear Michael,

thank you very much your helpful and prompt help. The suggested approach seems to fit indeed. I see only one potential problem.  Since I do have nascent RNA and polyA signal from the same biological replica, so I do have paired data. If I now go along your suggested line, I somehow pretend to have 3 replica x 2 condition x 2 assays = 12 biological replicas. But in fact I do only have 6 biological replicas. wouldn't that make the whole approach assailable since I underestimate the true biological variance?

Thank you again for your precious time and help.


ADD REPLYlink written 2.4 years ago by fabian20

You have 6 biological replicates, but you have 12 experiments. You can add a term for replicate, to control for within-replicate differences. And then you will only be performing inference on whether the nascent/polyA signal is different across condition, controlling for replicate. See this section in the vignette:

ADD REPLYlink written 2.4 years ago by Michael Love25k

Those replies come almost as fast as if I would call you.

Thank you very much for your assistance.

ADD REPLYlink written 2.4 years ago by fabian20
Answer: DESeq2: Tests of log2 differences
gravatar for Michael Love
2.4 years ago by
Michael Love25k
United States
Michael Love25k wrote:

Does this answer? 

DESeq2 testing ratio of ratios (RIP-Seq, CLIP-Seq, ribosomal profiling)



ADD COMMENTlink written 2.4 years ago by Michael Love25k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 221 users visited in the last hour