I've read that one can estimate how long ago it was since two individuals shared a common ancestor by comparing their DNA - the more differences that have accumulated, the longer ago it was.
Chimp and human were once said to be about 98% identical, but a few years ago it was noticed that that number did not take "insertions and deletions" into account and so the current estimate of similarity is about 95%, I think.

In Dawkin's "Ancestor's Tale" it is said that proto-humans and proto-chimps parted ways about 6 million years ago.
Is that number based in 98% or 95% similarity or did that change not affect our estimate of the human-chimp splitting point?

Hi Joakim,

The molecular clock is, as you say, based on accumulated differences in DNA. However, before we can put dates on times of divergence (such as the human-chimp split) we also need a rate. However, any measure of rate must incorporate time (rate = change / time). So our clock must first be 'calibrated' using real dates obtained from fossils.

If we were to use human or chimp fossils this would bias our analysis, so instead other divergences in mammalian history are used to get a handle on the average rate of change for a mammal. Once we have this we can make our estimate.

This is a simplification of the molecular clock, as we know that different parts of our DNA change at different rates and also that the rate of evolution varies between different groups. Having said this the human-chimp split is consistently estimated to have happened 5-7 million years ago, which is broadly consistent with what we know from the fossil evidence.

(The reason we don't just use fossils is because the fossil record is patchy - we can never be sure that the oldest fossil we have for any particular group is really the first member of that group to have existed).