Nonsense Induced Transcriptional Compensation (NITC)!

A bunch of new studies just dropped in Nature this month that go a long way toward explaining a lot of results that seemed like nonreproducible nonsense.

The best place to start is this summary!

Here is the paradox mentioned above in my dummies view of it: in higher organisms you can "knock out" a gene by messing up it's structure or you can "knock down" a gene/gene product by messing with it's regulator or silencing it's RNA production.

You'd think that decreasing the amount of gene by either mechanism would have the same effects (cause, obviously, the gene doesn't actually do anything --- only the protein and how much of it is around is what is important) -- and sometimes it does.

But, other times the knockout and the knockdown will have VERY different effects on the phenotype. This will look like
1) Either the knockdown or the knockout didn't work
2) The person measuring the phenotype is:

 But  now there appears to be a mechanism -- it looks like short "nonfunctional" products from knockouts (often called nonsense mutations) may lead to compensation by causing upregulation of similar genes. Which -- from an evolutionary standpoint seems to make a lot of sense, right? Why have 30,000 genes or whatever if one single amino acid variant could call a stop codon and shut the whole organism down?

The evidence is described in zebrafish in this and this great new study(ies?).