Wednesday, September 19, 2012

Fossil Data and Genetic Data More Concurrent Than Thought

It is now coming out that the “clocks” that are used to calculate mutation rates are badly calibrated. Ewen Callaway of Nature writes:
Geneticists have previously estimated mutation rates by comparing the human genome with the sequences of other primates. On the basis of species-divergence dates gleaned—ironically—from fossil evidence, they concluded that in human DNA, each letter mutates once every billion years. “It’s a suspiciously round number,” says Linda Vigilant, a molecular anthropologist at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany. The suspicion turned out to be justified.
Now it seems that the clock is off by half. This means that a bunch of divergence dates can be recalculated. One is particular is mentioned:
Take the 400,000–600,000-year-old Sima de Los Huesos site in Atapuerca, Spain, which yielded bones attributed to Homo heidelbergensis, the direct ancestors of Neanderthals. Genetic studies have suggested that earlier ancestors of Neanderthals split from the branch leading to modern humans much more recently, just 270,000–435,000 years ago. A slowed molecular clock pushes this back to a more comfortable 600,000 years ago.
It isn't a magic bullet, there are still some kinks to be worked out but it has remarkable potential for getting a handle on when the Denisova genome might have arisen.It might also give us a believable figure for the LCA of apes and humans since the current, 5 million year date is plainly not so.


  1. Anonymous5:10 PM

    Once every billion half that. Did you mean once every million years?

  2. Didn't catch that. Have to go back to the Nature article to see if there is a correction. That is a direct quote.

  3. I wondered about that number, too. I've been looking at some articles on the mutation rate in humans, and I'm seeing 1-2 x 10^-8 point mutations per nucleotide per individual per generation. That's at least 10 times the rate you quoted. It gets to be much less than 10^-9 if you do it on a per cell division basis, as there are several hundred cell division generations in a single human generation in the germ line (several times as many in males as in females. The overall rate averages that out.)

  4. Just realized that my comment didn't make sense - I was referring to the individual mutation rate, not the species level mutation fixation rate. I'm not enough of a population geneticist to predict one from the other.