For example. We have a set of samples, where genes are with increased expression. But when we look at the values for each sample we see that for some of them expression is super high, which results in overall increase. Have you ever seen something like that in articles? I need a reference.