# Poisson and Gamma distribution for testing randomness

In genetics I want to test whether InDel (insertion and deletion in DNA) sizes occurs with the same probability.

I heard that I should gamma distribution to model it. I found

library(goft)
gamma_test()


in R and I rejected hypothesis that length of InDels are gamma distributed, but what does it mean? Doest it mean that are no random or what?

Second case. I want to check whether in some bins of chromosomes, some mutations occur randomly so I should use Poisson distribution. Which test with which parameter should I use to check randomness of mutation occurrence in chromosome across each non-overlapping window (bin)?

• Welcome to CV, @bwczech! Could you be a bit more precise in describing what you are trying to do? When you did the gamma test, how did you set the range of parameters tested? Did you (or the test you used) do some kind of optimization? Perhaps this is documented in gamma_test(), but in any event, you should know what was actually being tested before you trust the results. The greater the detail you offer in your post, and the more focused the question, the higher the chance of a useful response. – Peter Leopold May 13 at 15:47
• Ok, sorry. I mean, e.g. we have a file with InDels length — we have about 10 000 InDels length ranged from 0-180. I want to check whether those lengths occur with the same probability. Now I think that using chi square test to compare with uniform distribution based on histogram will do it what I want. What do you think? – bwczech May 13 at 15:54