How to use Poisson distribution to get meaningful expression counts?
I have a list of enhancers elements that I care about (about 70,000 in number, each exactly 3000 bp long genomic location). I also have total RNA-seq reads from an experiment (3 replicates, mouse), which I have mapped and gathered counts for these enhancer elements using featurecounts. Now, I want to know which of these enhancers have some meaningful expression level. For this, my plan is to get the basal expected expression level, and then if the expression level for an enhancer is greater than this, I would call it as being truly expressed. I have heard that Poisson distribution can be a good tool for such count-based modeling, but I am not sure how to go about it. Could anyone provide any help/ideas?
Thank you in advance!
• 42 views