GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2009-12 > 1260548944


From:
Subject: Re: [DNA] R-U152 and R-L21 on the European Continent
Date: Fri, 11 Dec 2009 16:29:04 +0000 (UTC)
In-Reply-To: <853828256.123041260548823161.JavaMail.root@sz0002a.westchester.pa.mail.comcast.net>


From: "Tim Janzen" < >

> However, we also
> have a problem in that John Chandler's estimated mutation rates for the
> slowest 10 or so mutating markers in the 67-marker FTDNA panel probably
> are a lower than they are in reality.



In fact, they (or some of them) are faster because of overcounting of the most slow markers (DYS388, DYS426, etc.)



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.



I gave an example in a recent article in JGG (Part 1, Fig. 4). They were 23 mutations of DYS426 in haplogroup J1 section in YSearch. In fact, there were only two of them (the "base" allele is DYS426=11 in J1). An allele "10" got to a Jewish community ~550 ybp, and repeated itself in the lineage to the present time. There were no mutations anymore of this marker in the lineage. Another mutation, DYS426=12 also repeats itself in another lineage. Hence, if to count mutations in two lineages, we got 23 mutations of DYS426 instead of 2. 



A similar situation is with DYS388=10 in R1a1 British, Irish and Norwegian haplotypes. This is a separate lineage, which should not be messed up with DYS388=12 in those haplotype series.It would be particularly disastrous when a "quadratic", ASD method is employed. Many more extra mutations will be added, making an ancestor MUCH older.  



However, those markers are so slow that they do not add much to the Chandlers' a veraged mutation rates per a panel, say, per 12-marker haplotypes. With them (DYS426 and DYS388) the mutation rate constant is 0.02243 per the 12-marker panel (per generation), that is 0.00187 mutations per marker per generation, without them 0.02212 and 0.00184, respectively. Not a problem, the same thing.



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: