Timing of ancient human Y lineage depends on the mutation rate: A comment on Mendez et al

Timing of ancient human Y lineage depends on the mutation rate: A comment on Mendez et al
Melissa A. Wilson Sayres
(Submitted on 22 Apr 2013)

Mendez et al. recently report the identification of a Y chromosome lineage from an African American that is an outgroup to all other known Y haplotypes, and report a time to most recent common ancestor, TMRCA, for human Y lineages that is substantially longer than any previous estimate. The identification of a novel Y haplotype is always exciting, and this haplotype, in particular, is unique in its basal position on the Y haplotype tree. However, at 338 (237-581) thousand years ago, kya, the extremely ancient TMRCA reported by Mendez et al. is inconsistent with the known human fossil record (which estimate the age of anatomically modern humans at 195 +- 5 kya), with estimates from mtDNA (176.6 +- 11.3 kya, and 204.9 (116.8-295.7) kya) and with population genetic theory. The inflated TMRCA can quite easily be attributed to the extremely low Y chromosome mutation rate used by the authors.

Advertisements

5 thoughts on “Timing of ancient human Y lineage depends on the mutation rate: A comment on Mendez et al

  1. It’s not “extremely low”: in fact it can be even still somewhat high. The pop. genetics section of the Academy is generally stuck in arbitrarily “fast clock” ideas that have the wrong fundamentals, like an OoA c. 60 Ka (when it was surely 80 Ka BP or older, cf. Petraglia 2007 and others) or the Pan-Homo divergence some time between 5 and 7 Ma BP, when it was surely 8-13 Ma BP, i.e. double (Langergraber 2012 and others).

    Using T.D. Robb’s long sequence Y-DNA tree and recalibrating it for the CF being c. 80 Ka BP, I get (by extrapolation, as Robb only used A0) an even older date for A00 than Méndez.

    So A00 and A0 lineages are not really (always in my understanding) H. sapiens senso stricto but must come from related but distinct populations living in Central-West Africa. This is consistent with Hammer 2011, who claimed genetic admixture with pre-Sapiens populations in Pygmies, Khoesans and other Africans.

    In turn all this explains well why Henn 2011 got highest diversity maximums towards Central-West and Southern Africa: if there was some minor admixture with Sapiens-related “archaics” in the expansion to Central-West and Southern Africa, that would significantly increase autosomal diversity in those populations, while possibly adding some lineages (Y-DNA A00 and A0, maybe also A1a, which is borderline in my chronological estimates) to the core H. sapiens population, for which the mtDNA suggests a more northeasternly origin (Ethiopia for Behar 2008, Upper Nile in my own revision of his data).

  2. I am fairly new to the Y-haplotpe tree crowd, so appreciate your references. I see that the recent work you cite does suggest a longer TMRCA for the Y (although this was not discussed in the Mendez et al paper). How do you extrapolate to A00? Have you used the sequence published in Mendez et al?

    You state that it is low to assume “Pan-Homo divergence some time between 5 and 7 Ma BP, when it was surely 8-13 Ma BP”, but, to be fair, I do think the Langergraber 2012 paper does have 7 Mya as a lower bound.

    I am working on a re-write of this paper to more appropriately highlight my major concern, which is that I don’t think it is appropriate to estimate a mutation rate for the Y chromosome using the mutation rate on the autosomes. There have been relatively few studies comparing substitution rates on the X, Y, and autosomes, but those that have (several papers by Pink et al, for example) find that estimates of male mutation bias are not equal when using X/A, Y/A, or X/Y, suggesting that there is not a linear relationship between the substitution rates on X, Y, and autosomes. Granted, substitution rates aren’t mutation rates, but I imagine with ever-increasing data, we’ll get a better hold on the mutation rates soon.

    • In any case it is my opinion and nothing more (nor less). I’ve generally been skeptical of the “molecular clock”, not so much in theory but in practice, especially the tendency by many to consider the estimates as “facts” and not for what they actually are: educated guesses based on a number of assumptions, which often are questionable.

      Recent developments like the ones mentioned seem to fit better in my scheme of things than in that of “scholastic” work within the Academia. But please weight all the evidence on your own with a free mind as I try to do myself.

      … “to be fair, I do think the Langergraber 2012 paper does have 7 Mya as a lower bound”.

      That’s not what I’ve read. Whatever the case the upper bound is much higher of even the 10 Ma I used to consider based on previous works.

      “I don’t think it is appropriate to estimate a mutation rate for the Y chromosome using the mutation rate on the autosomes”.

      Rather than using any pre-determined mutation rate, what the long sequence tree of Robb, for example allows for is for an SNP-based tree (and not anymore a mere STR one, which is probably not the best approach) which is blank in the dates. Filling those blanks actually got Robb (who prefer more conservative short estimates based on the literature, what I dub “scholasticism”) and I at odds, as I prefer to calibrate based on archaeological dates.

      He does not use the autosomes in any case: just the Y chromosome.

      … “but I imagine with ever-increasing data, we’ll get a better hold on the mutation rates soon”.

      I can’t but agree with that: full or at least long sequence Y-DNA study seems to be a more complete approach than studying just a bunch of microsatellites, which, for what I have discussed here and there, seem to behave differently depending on which haplogroup (or maybe other factors too) producing appearance of “low STR diversity” for ill studied clades like C or others, where maybe other STR choice may be more adequate, if anything.

      Thanks for your kind reply, Melissa.

      • Thank you, Maju!

        Regarding 1000 genomes: I am curious about how imputations affect the 1000 genomes data on the Y specifically. It is low coverage, but because of imputation, there is no missing data. I really don’t have a good sense of whether this inflates or deflates estimates of TMRCA.

        It is such an exciting time to do Y chromosomes work!!

  3. Pingback: Most viewed on Haldane’s Sieve: April 2013 | Haldane's Sieve

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s