The thick plottens

Pngarrison helpfully pointed us a day or two ago to a recent article in the context of the randomness of variation. It does raise some interesting issues from The Hump’s perspective. So I’ll very briefly summarise it in the knowledge that it’s open-access, and those with a better background can bypass my meanderings.

The study takes cystic fibrosis transmembrane conductance regulator (CFTR) as a typical example of a very ancient and essential gene (ie present for “hundreds of millions of years”), that has gained many variants through point mutations – otherwise known as single nucleotide polymorphisms (SNPs).

The article explains the plasticity of the gene – many variants work well – but also how catastrophic the loss of function is (as every respiratory physician knows from experience). It examines the relative ease by which accumulating mutations will suddenly render the gene non-functional.

The meat of the dish is the detailed examination of one problem: how over such a long evolutionary period this single gene has not simply degenerated by Muller’s ratchet to the point where higher life simply runs out of steam. One can multiply that problem by the very many similarly conserved ancient genes. It looks in detail at the conventionally accepted mechanisms such as sexual recombination, purifying selection, and neutral drift, and concludes that they would be fundamentally incapable of preventing this degeneration.

In other words, the pattern of variation observed now cannot be random, and they propose that there must be a mechanism in place that simply prevents many of the possible variants from ever forming, thus enabling purifying selection and so on to keep abreast of the job. Now, the controversial issue here, though such controversy is now becoming routine, is that it gives solid evidence of the non-comprehensiveness of Neodarwinian theory, even including Neutral theory. The authors insist that there are further mechanisms to be discovered, which would clearly affect our understanding of biology at every level.

I understood enough of the argument to see why they require this new understanding. In passing, however, they exclude other possible explanations, such as “directed evolution” and “Intelligent Design”. I think the paper would have been improved by more discussion of just what they were discounting in these, and why, since their argument was basically “We’ve shown these theories don’t explain the issue”, with bare references to sources on the theories in question.

My understanding would be that by “directed evolution” they mean something like James Shapiro’s natural genetic engineering in which mutations have some adaptive direction, and I can understand how their data might suggest that what variation there is in CFTR shows no sign of such adaptiveness. Incidentally, this only shows that in such cases NGE has no place; the evidence for it in other cases remains, and so we have only added one other non-Darwinian mechanism to the complicated mix that’s emerging.

On Intelligent Design, it’s hard to be sure what they’re actually denying, since they only reference Stephen Meyer’s Darwin’s Doubt, whose conclusions on mechanisms for detailed instantiation of design are intentionally minimal. But one can suggest they’re making a case for a plausible naturalistic explanation that Meyer didn’t discuss (though he wasn’t, of course, addressing genetic degeneration at all as far as I know).

This explanation of theirs turns out to be that, hidden deep in evolutionary history, life evolved a mechanism that would limit the nature and variety of SNPs in such essential genes, in the same way that it seems to have found a way of preventing the degeneration of the DNA code itself. This would explain the pattern of the data, ie that many of the variants that would inevitably accumulate deleteriously simply haven’t appeared, and so presumably are actively prevented from doing so.

The conclusion that there must be such a mechanism seems reasonable, and must change how evolution is viewed, by suggesting that evolution itself has evolved beyond the plausible simplicity of the Darwinian model. The more such a thing is accepted, the less fortuitous the “endless forms most beautiful” seem, and indeed the less fortuitous they are.

What is easy to forget, though, is that an actual mechanism has not yet been found, before the means of its arrival has been hypothesised. And the means suggested for its arrival, according to this paper, appear simply to be the standard Neodarwinian mechanisms that it claims to be incomplete now. Except that these simple Darwinian mechanisms have been endowed by the authors with almost supernatural powers of teleological foresight, that Darwin’s theory was formulated to avoid. Look at these quotes from the article:

The pathway was best interpreted as another device that co-evolved with both the genetic code and codon usage to help preserve the exome of higher organisms…

Based on the analysis presented here and the vital imperative to avert “meltdown” of protein coding DNA, we speculate that the introme, itself, might not only serve as an evolutionary strategy to support the generation of diversity (through alternative splice variants, microRNAs, gene network regulation, etc.), but as a specific alternative to meddling with the exons.

…but would be essential for selfish genomes attempting to cope with environmental challenge over an evolutionary timeframe.

The first quote implies that “life” was somehow aware of the need to preserve the exome and devised a strategy specifically to do so. Well, OK, at a stretch one might re-interpret that teleological design-language to mean that organisms happening to develop this grand new mechanism did fortuitously preserve their exome and de eventu didn’t die out and take the whole biosphere with them. That was another bit of luck, then.

The second quote, though, is even harder to extricate from its language of “evolutionary strategy” and the “vital imperative to avoid meltdown.” Quite apart from “the introme’s” quite unconscious realisation that “meddling with exons” would all end in tears.

But the third quote, starting from the already anthropomorphic “selfish gene” analogy, extends their wilfulness to an “attempt to cope” not in their own time, but “over an evolutionary timeframe”. In other words, the picture is like one of these ancient sci-fi civilisations planning over millions of years to survive the loss of their atmosphere or some other predicted catastrophe.

But is it not the whole point of evolution that it makes short-term fixes for today’s problem, and cannot make provision for the future? “Just good enough for today” is one of the commonly used arguments against divine design. But the only way the blind watchmaker knows his gene is at the point of meltdown is that his organism becomes extinct. At that point it’s a little late to start evolving a sophisticated mechanism to ensure your descendants stick around for 3 billion years. And if randomness were able to simulate such wise foresight, you’d have thought it could just as easily have found a way to ensure that all mutations were adaptive and all species successful, thus saving much bloodshed and heartache for egocentric genes.


I have three points, in conclusion, to make on this interesting, and increasingly typical, piece of research, other than the obvious one that overly simple theories should have had their day in the light of complicated reality.

Firstly, on randomness, which is where we came in. Taken overall, the study shows that the SNPs hitherto thought to be random are clearly not – they are only a subset of what is possible, the rest being universally prevented by whatever the magic mechanism is, ostensibly with the longevity of Gaia in mind. In this case randomness was an assumption found wanting on closer examination. And that endorses the point I’ve been making over a few posts that randomness is nothing more than a pattern one hasn’t yet understood. At this point in evolution, such mutations are constrained by a front-loaded mechanism, just as Shapiro’s NGE has some kind of bias inherited from the past.

Secondly, arising from that, it’s a little disingenuous to explain the teleology that one now observes (preventing genetic meltdown is just as teleological as designing mutations) by pushing the random process back into an unreachable past. “Evolution is cunningly constrained now, because random mutation, selection and drift at the dawn of life happened to create those sophisticated mechanisms, though nobody can now say how.” Somebody’s already coined the term “randomness of the gaps”, and there’s a danger of an endless regress in that: perhaps some other even earlier evolutionary strategy governed how the selfish genes became alerted to the genetic danger they were in so they could start to save the planet (for entirely self-centred reasons, of course). Perhaps molecules randomly became self-replicating. Perhaps only useful proteins are likely to be found by evolution. Perhaps the environment naturally favours rational intelligence. Perhaps the dice are naturally loaded.

Thirdly, on my other leitmotif of the theory-laden nature of science (and indeed, all knowledge), it goes without saying that many people reading this article from a thoroughly Neodarwinian mindset would only be able to perceive through those eyes. There are various ways that could be achieved.

For example, most paradigmatically one can always find some way of accommodating the data to the existing theory (or any existing theory). In effect, the article creates just such an epicycle by saying, “Look, we’ve found a non-Darwinian phenomenon, but we suggest it could have evolved in a Darwinian way, so nothing’s really changed.”

Alternatively, one could shrug off the problem as an unimportant anomaly which, “undoubtedly” will be explained by the standard theory one of these days.

Or one could simply deny the findings of the paper altogether. At best, the authors don’t understand evolution. Or maybe they’ve even skewed their data to make a name for themselves. Or best of all, since the research team is based in Birmingham, Alabama, you just have to suspect that some of them may be Southern Baptists and therefore closet creationists, if indeed they have real degrees at all.

But I come to it with a different theory-load. I think it’s intriguing work that shows the secondary causes to be a bit more adequate for the job of populating God’s cosmos as he purposed – but it does little to explain sufficiently how those secondary causes came to be what they are.

Jon Garvey

About Jon Garvey

Training in medicine (which was my career), social psychology and theology. Interests in most things, but especially the science-faith interface. The rest of my time, though, is spent writing, playing and recording music.
This entry was posted in Creation, Science. Bookmark the permalink.

4 Responses to The thick plottens

  1. pngarrison says:

    I’ll be curious to see what Lou has to say about this, but my guess is he will say “yes, recombination and purifying selection” can account for the observations is this paper.

    The bias to transition mutations is textbook stuff. I haven’t kept up with DNA polymerase studies recently, but I would guess the bias toward transitions is just a structural matter. Trying to cram 2 purines together is going to distort things relative to a purine and a pyrmidine (purines have 2 fused rings and pyrimidines only one ring), and 2 pyrimidines will leave a gap, probably filled with H2O. The most common point mutation (by several fold) is a transition is C->T, because cytosine spontaneously deaminates, especially when it is methylated, and a T is the final result.

    There are also selective reasons for synonymous mutations being retained. Highly expressed genes have codons corresponding to abundant tRNAs. You can usually increase the expression of a cloned low expression gene by changing the codons to optimal ones in the species where you are expressing it. I have seen a few papers with evidence that specific codons actually are selected for. tRNAs have a surprising array of functions, and there may be other reasons for preferring a particular codon. There are also transcription and splicing regulatory sites in coding regions that could account for the 3rd position nucleotides being selected.

    There is no mystery about why Y chromosomes have less diversity. Without recombination, if a Y mutation occurs that clobbers male fertility (about he only function on the Y), that whole Y and all its variants will be lost. They write as if fitness if only 1 or 0, and a mutation has to be lethal to have an effect on fitness. Fitness is quantitative and the result of the mass of variants that get together in an individual. If the sum of the effects of those variants is no offstring or fewer offspring, the copies of all those variants will be lost, not just one.

    Talking about 2 people who share a common ancestor 50,000 years ago is simple minded. Beyond a few thousand years ago, everyone who has any current descendants is an ancestor of everyone alive today. Those two people would be unlikely to have any DNA segments at all from their ancestor 50,000 years ago, even though they are his descendants, because they have a large number of other ancestors. It is well known that the coalescent age for autosomal genes is a million or more years. There are very old variants, some them so old they are present in chimps too, but there are also countless young variants that have occurred in the massive population expansion of the last few thousand years. Each ond of these recent variants is rare, but all of them together represent a significant fraction of all the SNPs in the species.

    There is a genetics literature on the problem of mutation load. I have read things by J. F. Crowe and by Kondrashov, but these guys don’t even reference any of it. Those papers are hard core population genetics that I don’t know enough to evaluate. I don’t know what the answer to that problem is, but I don’t think it’s in this paper.

  2. Lou Jost says:

    Preston, I have only skimmed the paper; hopefully I’ll be able to read it carefully later. I like your comments. I’d add that I partly agree with Jon that “randomness” can cover ignorance of mechanisms in some cases, and as you mention, we do in fact know some physical mechanisms that bias mutations. But that does not contradict standard evolutionary theory.

    • Lou Jost says:

      I forgot to mention this article that appeared in the sidebar of the article you showed us:
      It gives evidence for natural selection controlling mutation rate, and for drift limiting this effect, as expected according to standard population genetics.

    • pngarrison says:

      I’m in the process of reading a review on the basis of mutation rate variation in humans in this year’s Annual Review of Genomics and Human Genetics. It seems to be a pretty good overview of the area. Of course there’s bound to be a lot about this that we don’t know, since they just acquired the ability to determine mutation rate directly in a family a few years ago.

      I’m the lone “professional” in the U106 Y-SNP Haplogroup discussion, and they would be surprised to hear the Y variant rate is much lower than the autosomes, as they are dealing with a rapidly expanding group of SNPs from next-generation sequencing. I haven’t yet coughed up the $500 bucks it takes to get 10 million bases of your Y sequenced.

Comments are closed.