Archiver > GENEALOGY-DNA > 2009-12 > 1260626924

From: "Anatole Klyosov" <>
Subject: Re: [DNA] R-U152 and R-L21 on the European Continent
Date: Sat, 12 Dec 2009 09:08:44 -0500
References: <>

> From: "Tim Janzen" <>
> I think that most of us who are reasonably knowledgeable about TMRCA
> estimates would agree that in all probability Y Adam didn't live between
> 300,000 and 400,000 years ago as is suggested by the above TMRCA estimates
> using only the slow-mutating markers. This would suggest that either John's
> mutation rates for these markers aren't highly accurate, that these markers
> haven't mutated at a consistent rate since Y Adam lived, or that there is
> some other variable we haven't yet accounted for such as "overcounting" as
> Anatole suggested. I would appreciate any insights into this that anyone
> may have.

> I think that the advantage of Anatole's method of looking at the
> actual haplotype trees is that he notices major aberrations in the numbers
> of mutations and tries to take those into account when generating his TMRCA
> estimates. Anatole, you previously mentioned that your best estimate for
> the age of Y Adam was about 80,000 years ago. Is that still your current
> working estimate for the age of Y Adam?

Dear Tim,

There are (at least) three items following from the above quotation: (1) accuracy
of John Chandler's mutation rate table, (2) significance of that possible inaccuracy,
if anything, for TMRCA estimates, and (3) "age" of Y Adam.

(1) If John has pulled together all, say, DYS426 mutations from, say, YSearch
database, there certainly was an overcounting on the reason I have explained earlier.
To be specific, instead of 23 mutations of this particular marker in J1 haplogroup,
there were only two. Hence, instead of his mutation rate of 0.00009 in his table,
there might be (just a wild guess) 0.00002. Similar thing would be with DYS388,
instead of his 0.00022 there might be, say, 0.00013.

(2) Is it significant for TMCRA estimates? Not at all, if, for example, FTDNA panels
are employed in their entirety, or just slightly abbreviated. Those the most slow
markers do not contribute much when considered in that "blender" of both fast
and slow markers. For example, I use the mutation rate constant of 0.00183
for the 12-, 25- (the FTDNA notation) and 21- and 39-marker haplotypes (the latest
Underhill et al, 2009, paper on R1a1). It turned out that the mutation rates for all
these four haplotype datasets is just the same. "Experimentally proven".
Those the most slow markers contribute almost nothing there. However, I cannot
remove them since I have a principle not to remove anything, otherwise it borders
with manipulations. However, if only slow markers are used, it is a big deal.
It is a huge margin of error, probably no less that 100% or more.

(3) I do not remember my "best estimate" of "age" of Y Adam. I doubt if I have
or ever had any. Furthermore, I do believe that ALL those estimates were a bogus.
It was one of those fables population genetics and DNA genealogy are so rich with.
No doubt that human population passed through a bottleneck some time ago,
and all haplotypes coalesce to that time. However we do not know when was it -
it could have been between, say, 50,000 and 500,000 years back.
I have tried to do my best with haplogroups A and B, but bumped onto a severe
bottleneck for haplogroup A ~28,000 ybp (technically I got 28,200 ybp, with 107
of 37-marker haplotypes from YSearch). A similar story was with haplogroup B,
~26,000 ybp (technically 25,900 ybp, with only 15 of 25-marker haplotypes
from YSearch). THEIR common ancestor, of haplogroups A and B, lived 36,700 ybp.
Then - completely black hole.

Of course, YSearch does not give us a good selection of African haplotypes.
Most likely, this selection represents well educated people, mostly African-Americans,
able to pay for their DNA tests. However, even 36,700 ybp shows, that those
presumably educated people came from variety of African tribes.

Now, modern anthropology is not a great help either. It is itself in a disarray.
They put an estimate for appearance of the anatomically modern Homo sapience
between 195,000 and 35,000 years ago, with a median around 160,000 ybp, which
is highly arguable for very many of anthropologists. This is an area of fierce debates
between anthropologists and paleo-archaeologists. Still, it is not quite related
to "age" of Y Adam. It could be "younger", it could be "older". It could have been
not necessarily "anatomically modern". It could have been an "advanced Neanderthal",
a sort of, but presumably of an African nature, advanced morphologically and mentally,
which is not found yet by anthropologists and whose DNA not tested as yet.

We simply do not know. Yet.

Anatole Klyosov


> However, there is no problem with the John Chandler's estimated marker
> rates, unless you use ONLY slow markers and do not separate branches on a
> haplotype tree.

> However, when one uses ONLY slow markers, it is a big deal what those
> overcountings do. You count many more mutations that supposed to be, and
> make a phantom common ancestor much more ancient. That is what you, Tim,
> typically get with your estimations. As you have undoubtedly noticed, your
> "slow marker" TCA's always higher compared with that by other methods.
> Regards,
> Anatole

This thread: