Entering edit mode
20 months ago
bioinf.questions • 0
I have two similar datasets containing the same cell types which I am attempting to compare with SCENIC. I'd like to know the difference in regulatory network activity between the two datasets.
What would be the best way to go about this? Should SCENIC be run individually and a comparison somehow made, or would it be best to run SCENIC once on the combined datasets?
What is wrong with doing both?
In my opinion, running the data from two different experiments separately makes more sense because when we combine them together, there will be chances that you would lose the specificity.
For example, If you rank the predicted regulons for the combined dataset you might get end up with the regulon for belongs to one dataset because of read count variation OR other sequencing-related differences. Of course, you can represent the results and compare the commonalities and specificities (visualization) in the end but the actual dataset-specific regulon would be different if you process it separately.