Thursday, August 28, 2008

Useful RNAs?

The latest issue of Nature contains a news feature by Anna Petherick [Genetics: The production line]. The article is mostly about a new regulatory RNA called HOTAIR but it's the implications of this discovery that bother me.

Let's look at the question being posed ...
If more than 90% of the genome is 'junk' then why do cells make so much RNA from it?
One of the answers being promoted by many molecular biologists is that this RNA is mostly functional and it represents a massive new level of control that has hitherto gone unnoticed. That doesn't mean that we never knew about regulatory RNAs because, in fact, we've known about them for over three decades. The important point of this opinion is that these RNAs are abundant and it "explains" the presence of so much non-coding DNA in eukaryotic genomes.

The other answer to the question is that transcription is sloppy and it frequently makes mistakes. That's why there's a very low level of transcription from all parts of the genome. It's junk RNA. This explanation seems to be widespread in the molecular biology community but it doesn't get much press because there are few papers that discuss this hypothesis [What is a gene, post-ENCODE?] [Junk RNA].

The probem in this field is that it's difficult to publish a paper that proves a negative (but see Brosius (2005)) and it's easy to publish a paper showing that a particular non-coding RNA has a function. The rare examples of those with function get all the publicity and obscure the fact that 99% of these transcripts may not have a function.

The article continues with ...
It is hard to comprehend the upheaval that RNA has been causing in molecular biology over the past few years. Once viewed as a passive intermediary, it was thought to faithfully carry genetic messages from the DNA sequence to the protein-making machinery, where things were made that actually got things done. Biologists were comfortable in the knowledge that only 1–2% of the human genome made protein-coding RNA in this way, and most of the rest was filler. So when, in 2005, geneticist Thomas Gingeras announced that some cells churn out RNA molecules from about 80% of their DNA, he astonished scientists attending the Biology of Genomes meeting at Cold Spring Harbor Laboratory in New York. Why should cells bother with so much manufacturing if, as it seemed, such a tiny fraction was involved in the important business of protein-making?
I wasn't at this meeting but I'd be very surprised if the scientists were "astonished." I'm pretty sure most of them thought that this was an artifact of some kind, probably due to accidental transcription.

This is a case where the author of the article could have benefited from interviewing more of the skeptics.
Over the past three years or so the case for this 'pervasive transcription' has strengthened. The phenomenon has now been ascribed to mice, fruitflies, nematode worms and yeast. These studies, and Gingeras's original reports, came from microarrays — a technology that relies on the tendency of nucleic acids to find their complementary cousins in a solution. Gingeras works for the microarray manufacturer Affymetrix in Santa Clara, California. But not everyone has been persuaded of the extent of pervasive transcription, in part because microarrays are subject to background 'noise'. Even using no RNA, control chips will give off some signals, and results can be a matter of interpretation.
Yes, false positives may account for some of the observations but I think most scientists recognize that the microchips are actually detecting rare transcripts. The question is whether these rare transcripts are biologically significant or whether they are artifacts like most of the alternative splice variants that made all the headlines a few years ago.

If the transcripts are accidental and nonfunctional then the fact that we see this in mice, fruit flies, nematodes, and yeast isn't a surprise. It is not evidence that the transcripts are functional. We would like to see evidence that most of these transcripts are (1) evolutionarily conserved, (2) reproducibly synthesized from a functional promoter, and (3) abundant enough in vivo to make a difference,
John Mattick, the director of the Centre for Molecular Biology and Biotechnology at the University of Queensland in Brisbane, Australia, has no such qualms. He is a long-time advocate of non-coding RNA's importance. The doubters, he says, "keep regressing to the most orthodox explanation [that the long RNAs are junk]. But they can't just sit on their intellectual backsides and tell us to prove it."
John Mattick is one of the most vocal cheerleaders for non-coding RNA. He maintains that huge amounts of it are biologically functional. His statement is a tacit confession that he has no proof of his claims. What in the world is wrong with asking for "proof" (evidence) whether sitting on our backsides or standing?

Is Mattick advocating science by assertion? It certainly seems that way in many of his papers.

Brosius, J. (2005) Waste not, want not – transcript excess in multicellular eukaryotes. Trends in Genetics 21:287-288 [DOI: 10.1016/j.tig.2005.02.014]


  1. "What in the world is wrong with asking for 'proof' (evidence) whether sitting on our backsides or standing?"

    What in the world, indeed?

    "The probem in this field is that it's difficult to publish a paper that proves a negative..."

    Especially "this has no function", isn't that the same sort of claim the IDers get chastised for? I can't see right now how X could be true, so it must be false.

  2. The situation is similar to the problem with adaptationism. It's not scientific to just assume that something has a function because it exists. The onus is on those who make the claim for function because there's a scientifically reasonable alternative; namely, no function.

  3. Proteins themselves can (demonstrably) have no function. Morphological traits can have no function.
    What's so special about RNA?
    larr's right. The problem of adaptationism. at all levels, seeing every trait as adaptive.

  4. I might add that maybe the 'function' of these ncRNAs is their capacityfor function. Evolution needs material to work on - and sometimes the 'junk' in the soup is just what mother nature ordered. Without all these ncRNAs around, there would be no opportunity for them to ever acquire function.

    Or am I missing something?

  5. jonathan asks,

    Or am I missing something?


    Natural selection is incapable of predicting the future. You can't select for something that may become beneficial in several thousand years.