I have a network which I clustered into 11 smaller clusters. I have a set of genes and I would like to check if statistically significant number of these genes fall on any specific cluster. One method which I thought of was to find which percentage of a total cluster were made up of my genes of interest. For example, say 60% of the genes in a cluster were the genes in my list. However, I was hoping that there was a better way to do this. This study on honey bees has done something of the kind.

I would appreciate if there was a way to do something similar. I do not know how to replicate what they have done in the study. Any help is appreciated.

  • $\begingroup$ Do you have problem with the hypergeometric test or with the later methods they do? $\endgroup$
    – llrs
    Commented Sep 27, 2018 at 7:48
  • $\begingroup$ @Llopis I have never done hypergenometric tests before and I am not sure how to do it. $\endgroup$ Commented Sep 27, 2018 at 17:17

1 Answer 1


A hypergeometric test is related to the fisher test, in a table with one dimension being some factors and the other one some other factors it is tested if they are independent. In R you con do this:

data <- matrix(c(0, 1, 2, 5), ncol = 2, nro = 2)

Or you can use the specific function phyper, to see the relation between them you can read this question and answer


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.