I have a list of DNA motif found in 5' UTR common in 2 expression groups (using FIRE). Now, we want to test if this motif is really enriched in 5' UTR than in 3' UTR in these specific expression groups. I am planning on using Fisher's exact test, and count genes with this motif in 5' UTR and 3' UTR (i scan both the forward and reverse orientation of the motif in 5' UTR and 3' UTR). Will my contingency table violate the assumption of independence, as I am using the same set of genes to test (although different location)?
I appreciate your suggestion.