GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2009-07 > 1247552290


From: "Tim Janzen" <>
Subject: Re: [DNA] TMRCA estimates for haplogroup J nodes
Date: Mon, 13 Jul 2009 23:18:10 -0700
In-Reply-To: <BAY111-DAV4229995994E19B51B0A96B1220@phx.gbl>


Dear Lawrence,
Good questions. I used the J1b haplotypes because they are a true
sister haplogroup to J1e. The J1* haplotypes are part of a paragroup. I
have done interclade TMRCA estimates using paragroups in R1b in the past. I
think that it is reasonable to use paragroups if you have no true sister
subclades to use for calculating TMRCA estimates. From the limited testing
I have done within R1b my impression has been that comparing a paragroup to
a true subclade tends to lead to a somewhat lower TMRCA estimate, but the
estimate is still reasonably close to interclade TMRCA estimates between two
true sister subclades.
I ran a variety of interclade TMRCA estimates for you using the
other 67-marker haplotypes in Bonnie Schrack's haplogroup J FTDNA project.
I used 11 J1*, 31 J1e, 1 J1d and 2 J1b haplotypes. Here are the results:

Haplogroup TMRCA in years
Node 50 markers YHRD 24-slow slow medium-slow medium medium-fast fast
J1e/J1*: 11845 6126 24342 103476 6418 6156 8272 8657
J1e/J1b: 6480 6972 9412 9418 9821 5818 6157 4984
J1e/J1d: 9937 2719 20187 40779 20449 5948 8399 6243
J1d/J1b: 11730 6554 30900 30808 37562 11016 10702 4023

J1d/J1*: 11885 4065 25630 65260 19933 9826 10154 6688

J1b/J1*: 11854 9238 22630 89017 7247 8046 8760 10008

I also ran an intraclade TMRCA for all 45 of these J1 haplotypes:

Haplogroup TMRCA in years
Node 50 markers YHRD 24-slow slow medium-slow medium medium-fast fast
J1 7039 4594 13030 50955 7521 5094 7154 5792

If I had to make a guess, I would say that the J1 node is about
20,000-25,000 years old. Note also that the intraclade estimates are about
10-30% lower on average than the interclade estimates, which has been my
experience when testing haplotypes generated by a program that creates
haplotypes with random mutations at a set generation interval from a known
haplotype.
Another point to be noted is that TMRCA estimates aren't
dramatically affected when a relatively low number of haplotypes are
included in the interclade TMRCA estimates, such as in the case of the
J1d/J1b node calculation where I had only a total of 3 haplotypes available
between the two subclades.
Sincerely,
Tim


-----Original Message-----
From:
[mailto:] On Behalf Of Lawrence Mayka
Sent: Monday, July 13, 2009 8:00 AM
To:
Subject: Re: [DNA] TMRCA estimates for haplogroup J nodes

Is there a reason why you used 2 J1b haplotypes instead of the more numerous
J1*/DYS388=13 haplotypes? I understand that technically, J1b is known to be
a proper clade whereas the DYS388=13 has not yet been proven to have that
status, but surely you don't expect to get a well-balanced calculation from
a bare pair of haplotypes?


This thread: