I am analysing two categories of protein coding genes comparing their relative characteristics. What I am observing is that one of these, in comparison to the other, is enriched in pyrimidine runs (two ore more consecutive pyrimidines). I have counted the number of these runs (I do not call them dimers because I have no data underlining a cross-link between two or more pyrimidine bases) per protein coding-gene performing then a Mann Whitney test that indicates that one of these two group of genes is enriched with respect to the other. I am hypothesising that this enrichment could be an early indicator of DNA to undergo local conformation changes. By your experience, could this hypothesis be reliable or I am totally misleading what I am observing?