Cooper: Assumptions in mutation rate

[swamidass (S. Joshua Swamidass)]

@purposenation

So mutation rate is really the only assumption used? And is it assumed to be constant, then?

Not exactly. We use a mutation rate that is (at first) assumed to be stable (but not constant). We then can test this, to verify with data if it is stable. It is. So this is not an assumption in the end. The model is established by data, through and through. Perhaps there is another way to explain this data, but I have not yet figured out how to do so. I don’t see anyone else close to figuring it out.

@purposenation

What is the sample size of the verification of mutation rates prior to ~10,000 years ago? Seems like it would be a very small sample size from a limited number of geographic distribution points?

Thousands of genomes at this time. I’m not sure (from memory) how many have been used to calculate mutation rates, but there are several studies. If there are any anomalies, we can be certain they will be discovered soon.

@purposenation

A bottleneck population, Group A, of 100 people exists 20,000 years ago. They have a strict reproductive society, where only a certain, small upper class of people are allowed to reproduce and it is strictly regulated.

After 10,000 years, Group B, a population of 5,000, splits off from A and is isolated. They adopt the extreme opposite of Group A and have a society of “free love,” where everyone has offspring with everyone else.

Perhaps one of these groups, maybe Group B, values diversity, where those who are the most “unique” are more attractive and thus, more likely to reproduce. In Group A, uniformity is more attractive and those with very different features are less likely to reproduce.

In the last 2,500 years, Group A and Group B recombine into Group C, and they adopt the free love society of Group B.

We have genetic samples from Group C and Group B, but none from Group A, so as far as we know, Group A may never have existed?

What bottleneck would we come up with and when?

In this scenario (assuming it is not a trick question), the sharpest bottleneck at 20,000 years ago is 100.

We may or may not be able to detect this in DNA. It would depend on the data. This is an imaginary scenario, we just cannot say for sure. Maybe we could detect it. Maybe not. Maybe, if it didn’t happen, we could rule it out. Maybe not.

In the scenario above, we have samples from Group B and C. We still might be able to infer that Group A exists. The inference of ghosts populations like this have been done several times already, and then confirmed later by subsequent studies.

@purposenation

Again, there are probably better ways to do illustrate my point here, but it seems you could come up with some scenarios like this that would give wildly different bottleneck sizes and ages, based on wildly different genetic mixing rates and mutation rates, right?

No. I do not think so.

@purposenation

ALSO: Perhaps we can break this off into a new thread, such as “Assumptions used in population bottleneck models.” While I might be a dummy, I guess I object to these questions of mine falling under this heading. =)

You are no dummy. These are good questions.