Test if subsample from a large population is over-represented by extreme values

hi everyone,

I want to do a statistical test for significance, but I'm having difficulty deciding which test is most suitable. I was hoping to get some help.

My data:
I have changes in gene expression after a pertubation, to be precise, a log2-fold change i.e. log2((expression after pertubation)/(expression before pertubation)).

What I want to test:
From this data set I have selected a set of genes (based on some hypothesis), which I expect might have either relatively low fold changes or relatively high a fold changes compared to the full set of genes.

So basically I selected a subset of my data and want to know if this is significantly enriched for extreme values (either high or low). However, if this is the case my background distribution (all genes) will have a different shape than my selected set of genes, which is an assumption of many nonparametric statistical tests.

Any advice on what test is most suitable?

Thanks in advance :tup:,