Monday, November 14, 2011

Jonathan Wells Talks About Sequence Conservation

Paul McBride (paulmc) tried to convince the readers on Uncommon Descent that there was evidence for junk DNA. One of the lines of evidence has to do with sequence conservation. If most of the genome sequences are not conserved between species this strongly suggests that they have no function, although it doesn't rule out a function that is independent of sequence.

Wells addresses this argument in: Jonathan Wells on Darwinism, Science, and Junk DNA. Before analyzing his response, it's worth reviewing what he wrote in The Myth of Junk DNA.

In chapter 5, Wells talks about sequence conservation as evidence of function—specifically the fact that the sequences of some potential pseudogenes are more conserved that would be expected if they were really pseudogenes [Junk & Jonathan: Part 8—Chapter 5]. That's an important argument and, if true, it would point to a function. The irony is that Wells doesn't believe in common descent so, from his perspective, these are not conserved sequences due to negative natural selection. Nevertheless, he is happy to use evolutionary arguments whenever it suits him.

What about the reverse argument; namely, that lack of sequence conservation is a indication of lack of function? Wells is not so happy to use that argument, as he explains in his latest posting.
Not so. Although sequence conservation in divergent organisms suggests function, the absence of sequence conservation does not indicate lack of function. Indeed, according to modern Darwinian theory, species diverge because of mutational changes in their functional DNA. Obviously, if such DNA were constrained, then evolution could not occur.

In 2006 and 2007, two teams of scientists found that certain non-protein-coding regions that are highly conserved in vertebrates (suggesting function) are dramatically unconserved between humans and chimps (suggesting… rapid evolution!). More specifically, one of the teams showed that one unconserved region contains an RNAcoding segment involved in human brain development.
Wells is arguing that evolution requires change. Therefore, sequence conservation is not what "modern Darwinian theory" would predict. Instead, it predicts that sequences in different lineages should change!

He picks two papers from the scientific literature that purport to show rapid sequence changes in regions that are functional. Wells is trying to discredit the idea that lack of conservation implies lack of function and he's using a few specific examples to make the case.

That pig won't fly. I looked at the first paper. The authors identified functional noncoding regions of the genome by looking for sequence conservation. They found that certain regions were well conserved in most vertebrates but showed evidence of more rapid substitution in the human lineage than in other lineages. These changes could be due to positive selection for a new function in humans but it could also indicate loss of function and degeneration of the sequences by random genetic drift. The authors prefer to explain the rapid changes by biased gene conversion, possibly due to loss of function followed by selection for G/C-rich substitutions in a G/C isochore.

In any case, the idea that lack of sequence function indicates selection for function rather than evidence for nonfunction cannot be sustained by cherry-picking a few controversial examples from the literature. Wells' argument is silly, and it's made even more silly when you realize that he doesn't actually believe in evolution.

Furthermore, the analysis by “paulmc” assumes that the only thing that matters in nonprotein-coding DNA is its nucleotide sequence. This assumption is unwarranted. As I pointed out in Chapter Seven of my book, non-protein-coding DNA can function in ways that are largely independent of its precise nucleotide sequence. So absence of sequence conservation does not constitute evidence against functionality.
This point is valid but it's not consistent with his first point. Either lack of sequence conservation indicates lack of sequence-dependent function or it doesn't. If it does, contrary to his first point, then one can still explain sequences that aren't conserved by invoking functions that require bulk DNA but not specific sequences.

Wells spends very little time on these kinds of functions, which is surprising since they represent the only real possibility to establish that junk DNA isn't really junk. He discusses enhancers, chromatin organization, centromeres, nucleus organization, and the idea that junk DNA could act as a lens in mouse eyes [Junk & Jonathan: Part 10—Chapter 7]. None of these make the case for most of our genome having a function.

If junk DNA is really a myth then there should be some good explanations for why we have massive amounts of DNA in our genome that has all the characteristics of nonfunctional DNA (lack of sequence conservation). Wells does not give us these explanations. Instead, as we see in his posting, he wants to deny that there's a connection between lack of sequence conservation and lack of function. But he's happy to accept evidence that sequence conservation implies function.

Isn't that strange coming from someone who doesn't believe in common descent, natural selection, and "conservation"?


  1. Although sequence conservation in divergent organisms suggests function, the absence of sequence conservation does not indicate lack of function. Indeed, according to modern Darwinian theory, species diverge because of mutational changes in their functional DNA. Obviously, if such DNA were constrained, then evolution could not occur.

    Not sure quite what this is saying. Darwinian theory demands that functional DNA cannot be constrained, and therefore it is unconstrained? Creationists have a problem with questions of degree. Can't it be fairly-constrained, or not-very?

    If ALL functional DNA were FULLY constrained, and all DNA were functional DNA, I guess evolution would stall. But since there is more than one way to skin a functional cat, then evolution can occur. Phew!

    There is more than one gene, more than one possible functional version of every gene, and variable degrees of constraint both per locus, and at any given locus in time, dependent upon epistasis, pleiotropy, duplication, environmental change ... it's all relative.

  2. As it happens there is a new paper on the % of the human genome that is functional (in Genome Research.) They guess 10-15%. They review evidence that there is fairly high turnover in non-coding sequence and take some account of this. It makes sense that there will be some species-specific sequence that is functional, but it shouldn't be very much. Some of the work they cite deals with transcription factor binding sites, and there are probably plenty of those that have no functional significance, which would explain their rapid turnover.

  3. The Genome Research article is freely available at