On the mortality of television sets: function in the human genome according to the free gospel of Genome Biology*


With great fanfare comes great cynicism, and so it should: science is built on a tug-of-war between novel claims and kneejerk skeptism, and the probity that follows. When a sound bite leaked out of last year's ENCODE publications, of which Genome Biology was a participatory journal, that '80% of the human genome has a function', evolutionary biologists were cynical all right, and they took to the journals to say so.

While some of the disagreement hinged on the semantics of the word 'function', a key sticking point was the scientific validity of declaring a stretch of the genome functional when there is no evidence for evolutionary constraint on its constituent DNA sequence. In other words, how much function can a DNA sequence really have when it is liable to change willy nilly, with no fitness penalty to the organism?

Seems like a reasonable argument. But what if it is wrong? What if, in certain circumstances, the make-up of a DNA sequence can change when its function does not?

lncRNAs enter the fray

Researchers in the field of non-coding RNA have for many years contested claims that most long non-coding RNAs (lncRNAs) can be dismissed as junk. Sure, a few lncRNAs might have a specific purpose but, for the most part, those long RNAs that are functional encode proteins, or so the argument goes. To say otherwise would beg the question: of what use could such a large quantity of lncRNAs really be?

And yet, inconveniently, functions for lncRNAs keep being discovered. An increasing number are being identified as competing endogenous RNA, or 'ceRNAs', which means that they are involved in the regulation of microRNA levels. Others still are being found to anchor together various combinations of proteins, RNA and DNA, using their unique ability to combine a specific tertiary structure with intermolecular base pairing.

Of this latter class, a number of lncRNAs have been shown to bind to the genome in order to regulate gene expression, with examples including PINT (and its mouse homolog Pint), which is newly described in a Genome Biology article.

PINT: the covert evolutionary operator

One of the many interesting things about PINT and Pint is that, although the two homologs are conserved at the level of function, an evolutionary analysis based on the conservation of DNA sequence would not detect them as homologs – as is the general pattern for human lncRNAs. But a synteny analysis supports the notion that PINT and Pint truly are homologs.

So when we are dealing with lncRNAs, perhaps evolution cannot be measured on the basis of DNA sequence constraint, even though function is still tied into the idea of evolutionary pressures.

And this is a problem for the ENCODE critics, because a significant portion of the disputed 80% is designated 'functional' on the basis of encoding a lncRNA.

PINT even cures cancer

How better to end a post about hyperbolic statements that overegg scientific studies than with a… hyperbolic statement that overeggs a scientific study? OK, so we don't know yet that PINT cures cancer, but the Genome Biology article does show it to be a candidate tumor suppressor, as a linchpin in a p53-pathway feedback loop. For a digest of PINT and p53, see Ramin Sheikhattar's Research Highlight.

*with apologies to Dan Graur and colleagues

View the latest posts on the On Biology homepage



If I understand the above commentary correctly, the claim is made that as far as genes specifying lncRNAs, evolutionary rules do not apply. In other words, in such sequences all mutations are acceptable, selection does not operate, the pattern and rate of nucleotide substitution are not affected by effective population size, and the evolutionary process is free of considerations on fitness effects. In other words, a new evolutionary theory is required to explain the evolutionary behavior of lncRNAs. Perhaps, intelligent design?

Moreover, the rules of evidence should be suspended. If 0.02% of the human DNA does something, then it is obvious that 80% do the same. This has been done already by a lead author of ENCODE. In an interview with Scientific American (https://www.scientificamerican.com/article.cfm?id=hidden-treasures-in-junk-dna) published on September 18, 2012, Ewan Birney proclaims: “[I]f we take quite a conservative view from our ENCODE data, we end up with something like 8 to 9 percent of the bases of the genome involved in doing something like regulation.” So, there is hard evidence for 8-9% functionality. Then the question arises where does the 80% come from? Birney continues: “And that 9 percent can’t be the whole story. The most aggressive view of the amount we’ve sampled is 50 percent. So certainly it’s going to go above 9 percent, and one could easily argue for something like 20 percent. That’s not an unfeasible number.” So, by extrapolation, we have 20%. Again, where does the 80% come from. I still don’t know. Interestingly, after admitting that only 20% of the genome is functional, Birney advocates the term “junk DNA” to be “totally expunged from the lexicon.” But wait, if 20% is functional, then 80% isn’t. You can only “expunge” the term “junk DNA” if you assume that 20% is larger than 80%.

Finally, the onion test is still a problem for ENCODE. Because, the assumption that humans have no “junk DNA” can only be reconciled with comparative data by resorting to special pleading, i.e., that all organisms that have more DNA then humans (e.g., onions, salamanders, apples, amoebae, lungfishes) possess junk DNA, but humans (being created separately) don’t.

In summary, my heart sings whenever a new lncRNA is discovered. I do not believe for a moment, however, that they invalidate the rules of population genetics and molecular evolution, nor the rules of evidence and extrapolation.

Naomi Attar’s gospel is not only evolution-free, but also logic-free.

Dan Graur
University of Houston

Naomi Attar

Dear Dan,

Thanks for your comment and for reading my blog post.

I think that unfortunately you misunderstood the meaning of my blog post somewhat, and I am sorry if I left room for this misinterpretation. I certainly did not say that evolutionary rules do not apply – in fact, if you reread my post you will note that I wrote that ‘function is still tied into the idea of evolutionary pressures’. As for the notion that I meant to support intelligent design, nothing could be further from the truth – you will not meet anyone more opposed than I am to the poisonous pseudoscience of ID.

The point I meant to make is not that evolution isn’t absolutely key, which of course it is, but that selective constraint might be harder to measure for some lncRNAs than for ORFs. For ORFs, it is clear that a given mutation to the primary sequence will be deleterious. But for some lncRNAs, their function might be robust to changes in the primary sequence – as seems to be the case for the interesting novel lncRNA (‘Pint’) described in the new Genome Biology article. I don’t know why this might be but would speculate that it is something related to conservation at the level of secondary or tertiary structure. There was actually a small part of Pint that did show selection at the level of primary sequence, but as it was a small portion of the transcript this would not have been enough to detect a homolog using standard primary sequence-based methods.

Of course this is just one example, and I completely agree that it doesn’t prove anything about any other lncRNAs, and certainly not 80% of the genome (again, if you reread the blogpost, you will see that I myself describe the ‘80%’ figure as ‘hyperbole’ and ‘overegging’). It just raises an interesting question about whether we are actually currently able to determine the % of the genome that is functional, if there are types of evolutionary constraint that we are currently not very good at measuring.

So please don’t mistake me as an apologist for the ‘80%’ claim, I am not. On the contrary, I thought the articles published by you and others, such as Ford Doolittle, made for very interesting and thought-provoking reading, and were a necessary counter to the ENCODE publications, a point that again I think I made in the opening paragraph of the blog post.

I would also maintain that I am not ‘logic-free’, but you are of course welcome to differ with me on that point.


Comments are closed.