Hi all,
I'm confused about how to analyze my data. It is a RNAseq experiment looking at unique mRNA changes in a specific cell population over time.
The design matrix is something like 2X4. IMPORTANT: I generate 2 samples from one piece of tissue, the cell-specific mRNA (NOTE: I do an immunoprecipitation protocol (think RiboTag or TRAP), and not single-cell RNAseq) and Total mRNA. I then have 4 time points (say 0,2,4,6 days). We want to know how mRNA expression for a gene or multiple genes change over time specifically in our cell-type of interest.
My biggest concern and question is whether or not each pair of samples (generated from the same animal) need to be analyzed together (IE sample 1 cell-specific mRNA vs. sample 1 total mRNA, sample 2 cell-specific mRNA vs. sample 2 total mRNA, etc.), or they can be treated like a classical RNAseq experiment (treatment vs control)?
The confusion comes when I want to go further with the analysis and compare the cell-specific mRNA vs Total mRNA of each time point and test for significant changes in gene expression over time. We want to do significance testing, and I'm not sure we can test for significant changes if we simply lump the replicates of cell-specific mRNA of a time point together and compare to the lump of replicates of Total-mRNA of the same time point. It can provide DGEs, but what about now comparing over time? Is it still possible (and statistically correct/accurate) to look at significant changes?
My inexperienced knowledge would say yes. DESeq2 provides standard errors for each log fold change (LFC). So couldn't I take the LFC of an individual gene and its standard error and do significance testing of gene expression changes of the same gene across time?
This seems like a question with a simple solution, and there very well could be. However, it's caused some disagreement and confusion between myself and another. I'd like to better understand this in order to strengthen the study as much as possible.
Please let me know if you have any questions or I need to fill in any gaps. I apologize ahead of time if my post is confusing.
Thanks!
