Over representation analysis for gene sets without child sets

[suzialeksander]

From the Helpdesk email:

Hi,

I want to perform over-representation analysis only for GO gene sets that do not have any child set. Is this approach fundamentally wrong?

Along with identifying the largest GO term, I also want a more resolved look at the over-representation among the smaller GO BP gene sets. For example, once I know that Apoptotic Process is over-represented, I want to compare the over-representation among the smallest gene sets under this term.

Thank you for your help regarding this query.

[suzialeksander]

@thomaspd