Saturday, 13 October 2012

A Critique of Richard Dawkins' The Greatest Show on Earth: The evidence for evolution - cont...

6. Embryogenesis defies an evolutionary origin

I apologise to regular readers for the delay in writing again, but I’m reading Jerry Coyne’s Why evolution is true, and I’ll be writing a critique of that as soon as time allows. But first I must finish commenting on Richard Dawkins' book.

As I've mentioned previously, Dawkins recognises that the complexity - and, quite frankly, the stunning achievement - of embryogenesis (embryological development) is a major challenge to the theory of (macro)evolution.

we find it hard to imagine, even in principle, how we might set about writing the instructions for building a body in the way the body is in fact built, namely by what I have just called 'self-assembly', which is related to what computer programmers sometimes call a 'bottom up' as opposed to 'top-down' procedure'. (p217)

He attempts to circumvent the challenge by trying to argue that it is all achieved merely by the operation of 'local rules', i.e. that it wouldn't need an overall plan (and hence no designer). But there are two fundamental issues which cannot be so readily brushed aside:

  • To build something as complex as a functioning organism, with many disparate yet interdependent parts, local rules by themselves are insufficient - there has to be some higher level of organisation as well to ensure overall compatibility and function.
  • For anything biological to work, implementation of those local rules is carried out by biological macromolecules (mostly proteins and nucleic acids); and, to work, those macromolecules must have closely-defined sequences (of nucleotides or amino acids).

The information required at either of these levels cannot have been generated in an evolutionary manner - by natural selection acting on randomly occurring mutations.

a. Embryogenesis requires higher level organisation

We know from everyday experience that a coherent end-product - one that functions well, or even at all - is not going to emerge from the bottom-up.

To use a simple mechanical example of constructing a car, or even just its engine: There’s no point in e.g. a main bearing of the engine being machined to perfection, unless there is also a matching surface on the crankshaft, and lined up accurately with other bearings, and the crankshaft connected to pistons (with other sets of matching bearings), which fit tightly (via piston rings of appropriate materials and construction) into the correctly orientated cylinders, synchronised with the camshaft etc. - you get the idea.

Put another way: Karl Benz didn’t come up with an automobile by starting with the detail of the engine - he had an overall plan, and he then designed and manufactured the components (which needed to coordinate with each other) in such a way as to implement that plan.

And it’s no different with embryogenesis; in fact there's another reason why overall planning is necessary - because, as Dawkins says, it's all achieved by self-assembly rather than by an external craftsman.

Each organ or tissue requires its specialised cells (e.g. lens, neuron, muscle fibre, erythrocyte) which must be produced only by that organ/tissue; and the organs/tissues must be formed in their right places - they would be useless if not detrimental in the wrong place - and connected appropriately.

How is this done in the course of embryogenesis?

Dawkins outlines the approach by reference to the worm Caernorhabditis elegans which has been studied closely, including its embryological development. He explains that, starting from the initial fertilised egg, at each cell division, each daughter cell is slightly different: each is different in terms of the genes it has switched on or off (gene regulation is much more sophisticated than simply on or off, but that's another story), and progressively the various daughter cells diverge morphologically.

And how is this done? - By a hierarchy or cascade of regulatory genes.

Dawkins is well aware (p358) of the Hox genes which occur in all animals. They play a crucial role in early embryogenesis - organising the overall body plan - for example each controls the development of a particular section (e.g. segment in insects) along the length of an animal's body. They do this initially by turning on the appropriate network or hierarchy of genes that form the organ, and continue to have a role in regulating the action of those genes. Incorrect expression of Hox genes can lead to major disruption of embryogenesis, such as a fully formed organ appearing in the wrong place on the animal's body.

The Hox genes themselves are controlled by a series of 'gap' genes (so called because if one is missing it leaves a gap in the resulting embryo) and pair-rule genes. And these gap and pair-rule genes are themselves controlled by mRNA that comes from the unfertilised egg.

Dawkins is right when he says that development proceeds through asymmetric cell division - i.e. before the cell divides each end is slightly different (e.g. in concentration of a protein or other chemical) and this leads to the differing gene expression in the daughter cells. Is this through the operation of local rules? - Well, yes, but only through a hierarchy of gene regulation which is extremely complex and we're only just beginning to unravel it. In other words, higher level instructions are required to ensure that the lower level development can take place.

And - lest any should think (despite my comments below) that it would be relatively easy to add another layer of expression at the bottom of this genetic hierarchy - there is a confounding twist: It will be apparent from what I've said above that the genes that organise development of the gonads (e.g. Hox genes and those 'below' them) are themselves controlled by cells produced within the gonads (via the maternal mRNA). So here is another example of chicken-and-egg scenarios which we find all too often in biological systems (the interdependence of proteins and nucleic acids in the synthesis of each other, is a very obvious one) and which completely defy an evolutionary origin.

b. Embryological development is mediated by very specific macromolecules

So what about implementation of those local rules?

Dawkins should be ashamed of himself for the way he glosses over the biochemical realities - relying on his readers' ignorance of biochemistry to get way with it. Here I can only outline what's involved - for more information you could look at Wikipedia's article on the Regulation of gene expression which includes the following that is particularly relevant here:

Furthermore, in multicellular organisms, gene regulation drives the processes of cellular differentiation and morphogenesis, leading to the creation of different cell types that possess different gene expression profiles, and hence produce different proteins/have different ultrastructures that suit them to their functions (though they all possess the genotype, which follows the same genome sequence).

This article - and others linked to it - gives an indication of the complexity of gene regulation.

But the main point I want to make here is to emphasise the nature of the regulatory proteins.

For example, a Hox gene (a DNA sequence) codes for a Hox protein (a sequence of amino acids) called a transcription factor which selectively binds to a specific regulatory DNA sequence associated with several other completely different genes. (Again, in most cases there are several transcription factors involved in regulating a particular gene, rather than just one.)

In an earlier post Half-truths about proteins I commented on the specificity required of an amino acid sequence just to ensure that it will fold into a 3-dimensional structure - that criterion alone is enough to defeat an evolutionary origin of proteins. Regulatory proteins must not only fold in this way, but part of their outer surface, once folded, must be of the correct shape and chemical composition (derived from its constituent amino acids) to bind to a specific sequence of nucleotides (and, in most cases, interact correctly with other factors involved in regulating that particular gene).

So what would be required for the evolution of just one new protein?

  • Obviously, we need a nucleotide sequence to arise by chance that, once translated into an amino acid sequence, will result in a protein that will fold, not only that, but will serve a useful function - one that will benefit the organism. That in itself is so improbable that it should not be taken seriously - but evolutionary texts do or, rather, like Dawkins, they uncritically assume it must be possible, because they aren’t willing to contemplate the alternative of design.
  • But that's only the beginning. Because, of course, a random nucleotide sequence will not be used to make a protein - it must also have the nucleotide sequence that means it is recognised as a gene and translated. But if the protein product has no value then there is no reason to recognise it as a gene - so the regulatory sequence must arise (by chance) in close association with, and at about the same time as, the sequence that arises (by chance) to code for a useful protein. Why is it that evolutionary texts never even mention this?

(Evolutionary text books roll out the fairytale of new genes arising by gene duplication - that while one copy retains the original function the other is free to ‘experiment’ to find a new useful function. But what that scenario completely overlooks is that until the duplicate finds a new function there is no reason to produce the protein, so it’s likely the control sequence will degrade, and once that’s happened even if a potentially useful sequence should arise there’ll be no way the organism could ‘know’ it.)

And what if, to realise the potential of the new protein (so that natural selection can act to retain it), just one regulatory protein were also required? What would need to arise (by chance, at more or less the same time) is, as well as the gene (with its regulatory region) for the protein, an independent gene that codes for a protein that selectively binds to the regulatory region of the end-product protein!

It is clear that the hierarchy of gene regulation in embryogenesis involves far more complex arrangements than that, especially when you consider that self-assembly sometimes requires production of additional proteins to transport the end-product proteins (before they can have any use). So it’s not surprising that Dawkins doesn’t want to delve into the details of embryogenesis, and would rather divert readers’ attention with fanciful talk of starlings and origami!


Andrew Chapman said...

David, do you believe that all the information to believe the adult organism is contained in the zygote? It probably has to be under philosophical naturalism/materialism, but not with some other ways of looking at reality. It just seems very far-fetched from at least 4 points of view:

1) The total amount of information required to specify the adult: 100 trillion cells in an adult man or woman, and however many zillions to specify a cell.
2) So far as we know the genetic code is just a template. It produces proteins and RNA, but these are just the raw materials. It's like the raw materials for a house or car or whatever. That's only a small way to the end goal. (I don't mean a person or creature is just that, but we do include structure, mechanisms etc). Gene regulation factors are also just molecules. How does cell 3,586,245 at cell division stage 20 know which regulation factors are to be turned on?
3) How is a differentiating parent cell meant to know what it's offspring looks like: a neuron, a skin cell, and thousands of others?
4) Can one even conceive of a way in which instructions for a cell 20 stages down could be passed down through the intermediate 19 stages, without them being implemented in the intermediate stages, and then being implemented in the 20th one?

I see that Jonathan Wells of the Discovery Institute has been persuaded by Rick Steinberg that not all the information is in the zygote:

'I used to believe that all the information required for embryo development was in the fertilized egg – even though not all of it was in the DNA. Several years ago, Rick Steinberg convinced me otherwise, and for what follows I am indebted to him.' at 40.00 to 45.50. He says that the new information is not material, does not come from matter (45.10). As a biologist he doesn't know where it comes from. As a theologian he thinks that the information comes from the divine ideas. Steinberg seems to be a non-Chrisian platonist, whereas I guess from the term 'theologian' that Wells is a Christian. I don't see any biblical objection to mind being prior to matter. Indeed, if anything, that makes more sense, since God is spirit, and in 1 Thessalonians 5 we have spirit, soul, body in that order.


David Swift said...


Thanks for your comment and link to Jonathan Wells’ presentation.

My default position is reductionist - I expect embryological development to be exclusively through physico-chemical processes. So I expect all of the necessary information to reside ultimately in DNA even if some of it is expressed indirectly e.g. though RNAs or structural features that are transmitted from one generation (whether of cell or organism) to the next. So my starting point is reductionist, but I’m open to change my mind in the light of evidence. (Just as my starting position was to expect evolution to be a satisfactory natural explanation for the development of life, but I now see the evidence is overwhelmingly against it.)

I think it is conceivable to have a mechanism that switches on/off genes in a systemic way (and not just on/off options) through a series of generations, e.g. in the development of the different tissues of an organism; and we see something of this in epigenetic mechanisms. Such a system must, I think, be complex and sophisticated; but we are beginning to discover something about these, including for example the many different types of RNAs (far more than when I first learned any biochemistry) involved in expressing genes (as well as protein transcription factors). As I see it, it’s not that a cell ‘knows’ it (or its progeny) is going to develop into a particular organ; each cell is simply following genetic/molecular mechanisms rather like blindly following instructions for making something without knowing what it will be in advance.

As I said in my post, for me these discoveries underline the fact that there is no way such sophisticated (and successful) mechanisms could have arisen opportunistically or by trial-and-error.

And although Jonathan Wells now doesn’t accept the ‘central dogma’ of DNA->RNA->proteins->morphology, I’m not persuaded by his examples. For example, just because RNA transcripts are processed and there can be a range of end-product RNAs, to me doesn’t negate the general process of DNA->RNA, it just happens to be more complex than simple one to one. And I expect we will find a mechanistic explanation. Same for processing of protein end-products. Even the spatial information in daughter cells - I don’t see why this can’t be organised in advance; so I don’t see that there would need to be an input of new information each cell-division or generation. But perhaps I should reserve judgment until I’ve read his paper(s)!

Regarding your last point: as I say on my website ( I do not think the Bible should be used to support or oppose any scientific theory.

Andrew Chapman said...

Thanks, David, for the reply. I appreciate your willingness to change your mind in the light of evidence, but why be reductionist in the first place? A conventional atheist seems almost forced into naturalism. There would be no place for mind in the universe prior to the evolution of life, for a Dawkins at least. But I don't see why a Christian should be, particularly.

As it happens, for myself, I began to believe that there must be intelligence in the universe before life began, before I began to believe in God. So I was mulling very vague concepts of an intelligent universe without a God. One can be vaguely Buddhist I think, or Teilhardian, and hold this kind of position.

I read Rupert Sheldrake's book A New Science of Life around 1986. He was a Cambridge botanist. There is a very interesting chapter on the Unsolved Problems of Biology. From memory, one section was on morphogenesis, and another was on the ability of wounds to heal. This last creates even greater difficulties for the idea of pre-programmed instructions. Just take a cut on one's arm. How do the skin cells know where to go, so to say? How do they know what one's arm should look like when restored? They can't be programmed for every possible accident.

DNA is just a molecule, so far as we know - I too am happy to work with a reductionist framework, when dealing with stuff, and keep going as far as it will take us. I think people go too far in saying it carries a message. It's a template, which maps bases to amino acids and forms proteins and DNA. So the end product is more molecules. They don't know anything (I assume!) - they are just subject to forces which propel them one way or another. Physico-chemical forces can form crystals, but not arms and legs, surely?

So I don't see anything in what we know about DNA that would suggest it is able to delineate even the basic structural form of the body, let alone the fine details of a lens. I was looking at a technical paper on eye morphogenesis the other day which used the word 'orchestrated'. I think it's natural when looking at one of these wonderful videos of a baby growing in the womb to perceive that it is being controlled by somebody or something. That doesn't mean it is, but it suggests that it could be considered as a possibility.

An atheist has to reject that possibility, probably, for the reason I gave above - although Sheldrake contents himself with some sort of evolutionary mind, growing up in tandem with the physical world, so to say. The Christian and other theists can be at least open to the idea that mind is prior to matter. When God created, I guess He had in mind what it would like before it came into being. Perhaps that idea continues to be formative in the natural world.

I don't see what's wrong with bringing the bible into science, in conversation with another believer. There is no point in doing so when talking with somebody who doesn't believe it, but I don't yet see the objection to doing so between believers. Even those on the more liberal end of the scale would probably get something out of Genesis 1 - that God, and His word, and His Spirit, are prior to the creation - in which case mind and form could be prior to matter, it seems to me.


David Swift said...

Andrew, thanks for your further post, just a few brief comments in response:

I think we need to be wary of a God of the Gaps mentality.

I think we need to distinguish between methodological naturalism (I don’t see how a scientist can practise science without adopting methodological naturalism, so I disagree with those (some of whom support ID) who oppose it) and ideological naturalism - that only naturalistic explanations are acceptable.

Here’s some recent work (published this month) that has uncovered at least one of the ways cells can divide asymmetrically to give rise to daughter cells with different molecular signatures, and hence different subsequent development.

I think Augustine was wise when he said (I bring his comment up to date) that Christians should not try to interpret the Bible in line with a scientific explanation (not just to oppose one), because if that scientific theory is subsequently seen to be false then it needlessly brings discredit on the Bible and its real message. I can’t find the original, but here’s a reference to it.
If only the church in Galileo’s time had heeded that advice! The churchmen of the time were intelligent well-educated people, but they wrongly thought the Bible supports geocentricism - and we’re still living with the fallout of their mistake!
And that’s my last comment about the Bible, because my blog and website are supposed to be about the science of evolution.