And isn’t the estimated minimum population dependent on how far back you go? This seems to be a reason for some people to put Adam&Eve or Noah (as progenators of all) much earlier in history than previously. I do not support such an idea, just to be clear.
2 Likes
Yes, very much so.
1 Like
I don’t know the number of individuals offhand – it was whatever was included in the defined superpopulations. Probably somewhere between 500 and 1000 individuals for each. For the SNPs, I included everything with a minor allele frequency > 1%, which amounted to 14,063,876 for the African population and 8,100,227 for the European pop.
Thanks @glipsnort for clarifiying this. If I understand you correctly, the African sample is several hundred individuals from each of the following populations:
YRI Yoruba in Ibadan, Nigera
LWK Luhya in Webuye, Kenya
MAG Mandinka in The Gambia
MSL Mende in Sierra Leone
ESN Esan in Nigera
ASW American’s of African Ancestry in SW USA
ACB African Carribean in Barbados
I think this is the same population set as in the 2015 Nature paper from the 1000 genomes project, with the addition of the MAG population, which I don’t think was in the 2015 paper.
I note that in the supplement for the 1000 genomes paper in 2015, the authors said (p. 5): “The amount of divergence among African populations is much higher; it would require sampling many more populations to capture rare variation adequately.”
Do you have any concerns that such a limited number of populations, (two of which are outside of Africa and perhaps to some extent admixed?) might not accurately portray the true African SFS?
The dataset I used did not include MAG but did include GWD (Gambian Mandinka) as an AFR population.
Note that in that paper we defined ‘rare’ as MAF < 0.5%, which means I discarded all rare variants for my exercise.
I’m not at all concerned about the limited number of populations, since the number sampled here is adequate to capture variants reflecting the shared deep ancestry of the populations – which is what we’re interested in. Rare variants unique to a population are almost always recent additions.
I would exclude the two admixed populations if I were doing a rigorous study, as they will distort the allele frequency spectrum. The distortion should be quite small, however. They might contribute 3 or 4% non-African (mostly European) chromosomal segments to the total set. You can get an idea of the effect of excluding them by increasing the African data points in the plot by 4% of the difference between the African and European data points. For the purposes of the illustration, it doesn’t matter at all.
4 Likes
Thanks Steve. Would it be worth looking at the SFS the H3Africa Consortium just to check that better sampling does not change the shape of the curve, or are you completely confident it would not make a meaningful difference?
Typically when this discussion comes up I always see people trying to argue for a single couple that is all of our ancestors. But I don’t think anyone in the relative scientific fields are saying that. Swamidass says that theoretically it’s possible that all humans can trace their genealogical history alive now back to a man and a woman but that these two would not have been a couple or even alive at the same time.
As soon as you step into even a tiny bottleneck population with two ancestors who are not an actual couple you have changed the genesis story of Adam and Eve completely since it paints them as a couple. So that would be a non literal interpretation and these non literal interpretations span from completely mythological and they are just metaphors for mankind as whole to beliefs thst it’s an ahistorical highly hyperbolic narrative written as a myth.
Theologically speaking, there is some reasons to be able to justify either of those positions. Genetics would obviously play no role in the two if they are just fictional characters snd if they were just a couple out of many being used as the characters in a historical fiction similar to Job then science would probably not be able to detect it.
What would constitute a meaningful difference for you? I’m quite confident that looking at a different dataset would not change the broad conclusion, and if you’re looking to set a firm limit there are better approaches.
2 Likes
I guess I would see a difference in SFS that would lead to a change of 100,000 years or more in the minimum time to a bottleneck of two as “meaningful”. I looked into downloading the H3Africa data, but it seems quite a process to get the right permissions, so I did not pursue it further.
This topic was automatically closed after 365 days. New replies are no longer allowed.