Hi,

I have some difficulty to understand the parameters of the negative binomial model in the context of gene expression.

The negative binomial distribution models the number of failures in a sequence of independent and identically distributed Bernoulli trials before a specified (non-random) number of successes r. It has 2 parameters:

r : number of successes until the experiment is stopped

p : success probability in each trial

1) In a context of gene expression and rnaseq: what is success, failure and trial ?

2) The notation of the classical model is NB(r, p). How in DEseq2 it becomes NB(s,d) with s=normalized count and d=dispersion.

Thanks for your help.

You may find this useful: https://github.com/mikelove/DESeq2/blob/2c3c14cbfa09300e503e08302c60ddf112beacfd/R/core.R#L2193-L2202