Colour me skeptical (but curious) about the conclusions as it presently stands. Let me reason it out:
The quantity of mRNA of any gene is dependent on its synthesis and decay rates. This means that genes that decay at a rate lower than the mean decay (basically perish slower* than average), will seemingly appear to be highly expressed after passage of unit time.
Figuring out this pattern is not easy, as it relates to the complex relationship between the quantity of the decay rate of a specific gene, the expression levels of that gene (think of order of reaction from chemical kinetics), the overall decay of the entire sample (either of which is not necessarily linear), etc.
The authors point this out and talk about it a bit in the discussion, but I'm not sure it's completely convincing. And hence, if what we observe is nothing but a normalization artifact.
*how slow is another ball-game (read: order of reaction from chemical kinetics) altogether
