 Sometimes when I talk about natural selection, I talk about it as being like, like gardening, right? Like you can imagine a plot of land where you've got some flowers growing, right? And seeds are randomly falling in that plot of land. Let's say dandelion seeds or other weeds, right? And if you don't care, if you're the sort of gardener who doesn't care about weeds in your garden, you'll just let those seeds sprout and grow and, you know, you'll have some mixture of dandelions and your cultivated flowers in your garden, right? Natural selection is like a gardener who weeds those seeds out. And so if you look at two gardens, right? One of them has a lot of weeds and one of them doesn't. And they're, in all other respects, matched to one another, right? Then you can say something about how diligent the gardener is, right? By comparing them. And that's essentially what we do in the genome. So we look for regions of the genome where the new mutations have been carefully eliminated, right? By natural selection. And those are regions where we can say selection is strong. We've actually played this game a lot over time. Over time, we've played the game of using these signatures of natural selection to identify particular regions in the genome that might be important. What we're doing here was slightly different. We were trying to say how can we use that in order to get a global measure of how important each one of these epigenomic data sets is, right? So rather than pinpointing specific regions of the genome, which are important, and some of that comes out as a byproduct of this analysis, rather than that being our main goal, our main goal was to say, how much do we learn genome-wide by doing, you know, an attack-seek experiment, or a chip-seek experiment, or from RNA-seek data?