Adam, Eve and Population Genetics: A Reply to Dr. Richard Buggs (Part 1)

[Swamidass (Peaceful Science)]

@Chris_Falter

Perhaps one of our biologist friends like @Swamidass, @glipsnort, @DennisVenema or @RichardBuggs would be able to shed more light.

Thanks for the invitation.

@Chris_Falter

The models are estimating the minimum size of an ancestral population at various points on the past. T

This is not accurate. They are estimating the average population size in a sliding window (that is quite large). They are not estimating the minimum. Just the average. The really interesting question is whether or not there is a way to determine the minimum.

Population size estimate studies, to be clear, are not testing for a brief bottleneck followed by an exponential expansion. The only paper I know that did this is the Ayala paper on MHC (http://www.pnas.org/content/91/15/6787.abstract) from 1994. Eventually we will get around to it, but this is the only published study I know that actually tests the idea. That was 25 years ago though, so the follow up studies are going to be interesting to look at.

The point that @RichardBuggs is making is that the power of population studies past about 500 kya to detect brief bottlenecks are not well studied. The fact that they do not find them, therefore, cannot be taken as evidence they do not exist. At least not yet. This is a question about detectability and statistical power.

@Lynn_Munter

And theologically speaking, just to clarify, would an argument that there was a bottleneck to two in very ancient homo sapiens, only it’s genetically undetectable due to later interbreeding with Neanderthals etc, rely heavily on reinterpretation of the sons of God/daughters of men in Genesis 6:2 to be referring to this interbreeding?

We are not speaking theologically at all. This is about the science.

I imagine those that take this position might take it not for Genesis, but because of Paul’s statements in Acts, Romans, and I Cor. I suspect New Testament theology drives this more than Genesis hermeneutics. Whatever the case, the impact on theology should be sorted out later. Perhaps we start a thread for that?

@Chris_Falter

Question: Why is this an accurate simulation for haplotype distributions? A Gaussian distribution is essentially pure noise except for the mean and the variance. @DennisVenema is claiming that the existence of TMRCA values > mya is a legitimate signal of ancestry, not just noise. If the observed PDF across the genome is the result of a stochastic process that makes some regions seem younger than the ground truth (as measured by TMRCA), then the noise should not be interpreted as nullifying the signal from regions at extrema that point to a very ancient TMRCA.

Does that make sense?

I’m not sure I understand the question. For one, I agree that that TMRCA > 1mya are legitimate signal of ancestry, for DNA. We are not talking about segments of DNA though, but a couple with four genome copies between them both. Its an equivocation to place them at an autosomal TMRCA. Right?

@glipsnort

I still don’t see why the demographic history of one branch of the structured population I descend from has anything other than purely academic interest to anyone. The fact that many (but not all) biologists have slapped different names on the different branches doesn’t strike me as relevant at all.

Well, for one, I am an academic! =)

Also, we were brought here by the claim that “Homo sapiens” never dip below a few thousand. Though, we all know, they eventually go to zero in the past. How can we know then that they do not stop at 2 on the way to zero? Yes, this is all about definitions. No scientific study goes here, because there is not enough traction to get clarity here. Which is why heliocentric certainty is not likely.

As I understand it (and it seems was confirmed by others, there are two claims at question here.

@Swamidass

1. Homo sapiens specifically do not dip down to a single couple in 300 kya to the confidence we have in heliocentrism.
2. Our ancestors as a whole do not dip down to a single couple between 300 kya and 3 mya with very high confidence, but maybe not as high.

@Chris_Falter

AFAIK, it’s basically the same thing, but on a longer time scale and across more of the genome than typical Y-chromosome studies.

That’s right. I’d say this higher confidence than the Y-chromosomal studies too.

@GJDS

Validation of models, as I have practiced (and is commonly understood) requires a result from the model to be similar to an observation/measurement of a system independent of the model data base

That is exactly what these studies do. There is real quality work here.

@GJDS

Perhaps you would elaborate; what predictions can any model make on what seems to be current data used to set up the model itself?

If you read the papers, you will see that quite a bit of validation is going on. There is really good work being done by scientists here. It is a mistake to dismiss it this way. There is a great deal of independent validation. Often they find that specific parameters do not affect the results much (e.g. mutation rate and generation time), except to scale the time. That leaves the debate on exact dates open (and does expand the confidence intervals), but it does not invalidate the whole effort.

The question here is actually far more interesting on a scientific level. We are getting into understanding exactly what these approaches can and cannot tell us.