My friend Razib, who is one of my
fellow ScienceBloggers sent me href="http://blog.beliefnet.com/cityofbrass/2009/01/is-evolution-sufficient.html">a
link to an interest attempt by creationist at arguing why evolution can’t
I say interesting, because it’s at least a little bit unusual in
its approach. It’s not just the same old regurgitation of the same talking
points. It’s still not a great argument, but it’s at least not as boring as
staring at the same stupid arguments over and over again. Alas, it’s not entirely
new, either. It’s an argument that the mutation rates required for humans to have evolved from a primate ancestor would have to be impossibly high.
In this case, the guy starts off with a gambit that always annoys the heck
out of me. That is, he starts by claiming to be a scientific guy who doesn’t
believe in all that ID rubbish.
Partly because I just went to Africa and flew over (but did not have a chance to actually visit) the Great Rift Valley (also known as the Cradle of Mankind), I’ve been thinking a bit about evolutionary theory of late. I am a scientist, but not a biologist – my field is medical physics. Still, my background makes me biased towards the scientific establishment and I am an ardent believer in the scientific method. While technically as a deeply religious person I do believe in “intelligent design” (in the abstract) I don’t believe in Intelligent Design as promoted by the various evolution-denialists in the political arena. I am quite strictly against introducing religious theories such as ID into the science curriculum.
Of course, from that point, it’s straight into the
ID rubbish for the rest of the argument.
He tries to use a mathematical argument for why evolution can’t work. But before
he really gets to it, he leads into it by saying something that I find remarkably strange.
From an engineering and aesthetic perspective, I have trouble with the idea that a system so complex as DNA and gene expression can be so rigid. My intuition is that DNA does not posess enough degrees of freedom to “encode” life as we know it. But how can I test that intuition without getting a PhD in genetics? I think I’ve come up with a way, though of course it is crude and rife with bland layman assumptions. Still, bear with me (and I hope to attract some attention to this from experts so we can refine it).
I really don’t understand what he means by this.
Coming at it from an engineering viewpoint, it makes no sense. DNA is not a simple, rigid system. In fact, part of what makes it so fascinating is its very fluidity. When we look at real gene sequences, to see how things have changed,
we find some really amazing things that demonstrate just how fluid DNA is. For one example, there’s the regulatory DNA – that is, there are non-protein-coding sequences that act as regulators for turning on and off the production of
proteins by other, coding regions. The location of a regulating region relative to the coding region that it regulate isn’t fixed. – different regulating sequences appear different distances from what they regulate.
Looked at from a mathematical viewpoint, it doesn’t make sense either. In the context of the argument that he’s going to move on to make, DNA is mathematically an information carrier – and talking about its “rigidity” in that context is meaningless.
Looked at biologically, it doesn’t make sense either. DNA is not rigid, biologically – that’s why mutations happen. If DNA were rigid, if it were able to reliably copy without error or change, and it always expressed itself in exactly the same way, there wouldn’t be any mutation at all.
Anyway, moving on…
Let’s take some basic numbers. There are about 20,000 genes in the human genome, with an average size of 50 kilobases (ie, 50,000 base pairs. Remember DNA is a double-helix, unlike RNA). Also, we are often told that humans and chimpanzees differ in their genomes by only 1%. Actually, that figure is only for genes where humans and chimps totally differm but there are some genes where the variation between teh species might not be so absolute. I’ll use 5% instead. Finally, we know that according to best estimates, humand and chimpanzees diverged from their common ancestor about 5 million years ago.
This is already pretty bad. If you’re just speculating, pull rough guesstimates
out of thin air is reasonable. It’s never good, but if you’re playing
with ballpark estimates, and you’ve got no way of finding out what the correct numbers are, then estimates are fine. But that’s not the case here. This is a number that can be looked up with a google search. But he doesn’t bother – he just takes a
rough ballpark number that he vaguely remembers.
But it’s worse than that. If you’re being scientific, and you’re making an
honest scientific argument, then if you use wild-ass estimates, you use
conservative estimates. That is, you don’t skew anything in your
favor. If you’re not sure, you use the honest estimate that’s worst for
your argument. But instead, he takes his vague memory of human/chimp DNA similarity,
and then multiplies it by five based on flimsy handwaving, in a way that makes it more favorable for his argument.
Taking these bits of data, we can actually estimate the required rate of evolution in terms of point mutations in DNA needed to turn a chimp into an ape. Of course, humans did not evolve from chimps, so we would then halve the rate we calculate to get the change from the common ancestor of both to humans (or chimps). So, let’s do the math.
Actually, let’s not do the math yet. Before we do the math, we need to
make sure that we understand what we’re doing, in order to make sure that we
do the correct math.
And he’s not. The key phrase: “required rate of evolution in terms of point mutations“. He’s going to only consider point mutation – one very restrictive kind of mutation. And he’s only going to consider independent point mutations.
That’s another example of stacking the numbers to benefit his argument. In reality, there are simple point mutations. There are also copies,
transpositions, deletions, breaks. There are point mutations that effectively
produce large numbers of new point mutations – for example, a frame shift. But
he completely discards all of those.
Ok, so now that we know that he’s stacked his math to try to make things
work out the way that he wants them to, let’s look at his math. (Note: I’ve change the formatting a bit to make it easier to read in a blockquote. The original text used indents, rather than bullets.)
- 20,000 genes x 50,000 base pairs = 1 billion base pairs
- 5% difference between humans and chimps = 50 million base pairs
- rate of change = 1/2 * 50 million base pairs / 5 million years = 5 base pairs per year
So, to go from the common ancestor of humans and chimps, to modern humans, the average rate of mutation required would be 5 base pairs per year. This seems like a very high rate to me; if we discretize into generations of 25 years each, then we are talking about 125 base pairs every generation.
125 base pairs per generation. Stacking things very much in his favor,
by increasing the magnitude of difference between human and chimp DNA; by limiting mutation to only independent point mutations; and by choosing an extremely long generation length (really, historically, when have primates averaged
25 year generations?), the worst he can make things is to require a population-change of 125 base pairs out of (by his own figures) one billion base pairs in the genome, per generation. In other words, he requires that
each generation’s genome differs from the last in approximately 0.0000125% of
Skewing the numbers, that’s the best he can get.
Getting that rate of change to fix in a population is unlikely – but not impossible. So cheating left and right, skewing things in arbitrary
ways, making things up, limiting the mutation process, slowing down the
rate of reproduction, exaggerating the required amount of change – all that,
and he still can’t produce a scenario where it’s impossible.
The argument just doesn’t work.