I read this article from Richard Dawkins as it seemed to be making a point about politics, language, and the way we think that I find really interesting. (I’ve had a post on language in science in draft for months!) However, what was most interesting was the insight I finally got into mutations in the population.
The most fun paragraph in the piece was this:
If a time machine could serve up to you your 200 million greats grandfather, you would eat him with sauce tartare and a slice of lemon. He was a fish. Yet you are connected to him by an unbroken line of intermediate ancestors, every one of whom belonged to the same species as its parents and its children.
I understood the intended point, that species do not have well-defined boundaries. The figure of 200 million generations struck me as low – it’s only 530 million years since the Cambrian explosion!
Given the source, I tried to understand why I thought it was too low.
Without having any specific numbers in mind, I believe that beneficial mutations are not common in the population . Mutations are common, and every individual will carry some. Many will be harmless, and some will be slightly harmful. But there are more ways for living things to get worse than there are ways to get better. If you took a genetic census of a population in one particular generation, only a tiny number of their members would have a new beneficial gene that was not present in the previous generation. For argument’s sake, imagine that any individual has one chance in ten million of having a beneficial mutation. Let’s call that the background rate.
In a population of 200 million then, I might naively expect only 20 individuals to have had beneficial mutations. This seems like far too low a number to make a difference big enough to count as speciation. So how could we go from fish to human in 20 mutations?
The mistake, of course, is that a population at one point is very different than a set of a current individual’s ancestors. Favourable mutations are more common among ancestors than among a population at a point in time.
The mutations present in a population at time T have not yet been tested by evolution. It may not be possible to know, just by examining the genes, which mutations will lead to a greater number of descendants. Time and competition need to play their parts. The rate of favourable mutations among your ancestors is higher than the background rate because they are your ancestors. They are being selected because they were better at having descendants than the other individuals that were alive when they were.
So the background rate of beneficial mutation should not be expected to be the rate of beneficial mutation in an individual’s ancestors. The two groups are defined very differently. One is the set of all the competitors in a race at a point in time, and the other is all those competitors who we already know won! Imagining that something that is true of one must be true of the other is a kind of selection bias.
 I understand that recombination is a bigger driver of change than mutation, but I want to stick to mutations because they’re easier to think about!