We are running into a problem when calling differentially expressed genes (DEGs) using an ANODEV with DESeq (or any other application).
Let me state our data then our dilemma.
Our data:
128 samples, sampled across 3 time points: timeA, timeB, timeC. (384 samples total = 128 x 3)
The 128 samples can be split into 2 conditions (disease (n=64) and controls, n=64) at each time point).
A design matrix would look something like this:
Total------------------TimeA(n=128)---TimeB(n=128)--TimeC(n=128)
DISEASE(n=192)-----n=64-------------n=64-------------n=64
CONTROL(n=192)----n=64-------------n=64-------------n=64
Our dilemma:
What we aim to do is to test for DE of transcripts across all 3 time points (using DESeq ANODEV) but we want to be able to identify at which time points these transcripts are being DE for both cases and controls separately. In other words, compare which transcripts are DE, with respect to specific time points, between cases and controls. Generally one would employ a post-hoc test to reveal the time point (timeA vs timeB, timeB vs. timeC etc...) where these transcripts are DE.
From our knowledge there is no implementation of post-hoc testing after ANODEV in DESeq or any other RNAseq analysis. In this scenario, what is the advantage to running an ANODEV when one needs to run a post-hoc test afterwards?
What insight can anyone offer us regarding DESeq towards understanding which transcripts are DE at our various time points for cases and controls?
Best,
Michael
Let me state our data then our dilemma.
Our data:
128 samples, sampled across 3 time points: timeA, timeB, timeC. (384 samples total = 128 x 3)
The 128 samples can be split into 2 conditions (disease (n=64) and controls, n=64) at each time point).
A design matrix would look something like this:
Total------------------TimeA(n=128)---TimeB(n=128)--TimeC(n=128)
DISEASE(n=192)-----n=64-------------n=64-------------n=64
CONTROL(n=192)----n=64-------------n=64-------------n=64
Our dilemma:
What we aim to do is to test for DE of transcripts across all 3 time points (using DESeq ANODEV) but we want to be able to identify at which time points these transcripts are being DE for both cases and controls separately. In other words, compare which transcripts are DE, with respect to specific time points, between cases and controls. Generally one would employ a post-hoc test to reveal the time point (timeA vs timeB, timeB vs. timeC etc...) where these transcripts are DE.
From our knowledge there is no implementation of post-hoc testing after ANODEV in DESeq or any other RNAseq analysis. In this scenario, what is the advantage to running an ANODEV when one needs to run a post-hoc test afterwards?
What insight can anyone offer us regarding DESeq towards understanding which transcripts are DE at our various time points for cases and controls?
Best,
Michael
Comment