Apologies in advance if this is a repeat post. I couldn't quite find the answer I was looking for, so let me know if it already exists. My question is about the statistical processes used for detecting differential gene expression. I am currently using JunctionSeq but know this is built on the methods of DESeq and DEXSeq (but I haven't used them before).
So the first step is counting the number of occurrences of each gene/feature between conditions A and B. The next step involves modeling/fitting a negative binomial distribution for each condition. In plain English, how is this done?
Once the models are fitted, the Wald test is applied. But I need help understanding how this test is run with this data. What is the test statistic? Or can someone point me to a good resource?
Thanks so much.