GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-12 > 1292193581


From: "Anatole Klyosov" <>
Subject: Re: [DNA] NW European R1b from Iberia?
Date: Sun, 12 Dec 2010 17:40:01 -0500
References: <mailman.2473.1292187996.2081.genealogy-dna@rootsweb.com><509DA485D5984C82BCFDBF7A5A2A8EE3@anatoldesktop>


>From: Vincent Vizachero < > Let me clarify. It is the
>number of STRs (or, if you will, the summed mutation rate of such
STRs) that is too small. In other words, accurate phylogenies of R-L11 and
subclades cannot be drawn
using only 67 or 100 STRs. I am fully aware that this fact does not prevent
some people (including you) from producing such trees, but they are still
worthless regardless of the amount of effort put into them.

My response:

It is truly funny, how many empty words are thrown here without any
substance. What is "too small"?? (with respect to "the summed mutation rate
of STRs). "Too small" for what? Why would not you take any specific example
to show what you are talking about? For example, if 655 of 67 marker
haplotypes of subclade L21 (data on January 2010) have 10,468 mutations from
their base haplotype, and the "summed mutation rate of such STRs" is 0.12
mutations per 67 marker haplotype per 25 years, what is "too small" here?
0.12 is too small? I repeat, "too small" for what?

It is not "too small" to divide 10,468 by 655 and by 0.12 in order to obtain
133 "generations" (25 years each, since it the mutation rate constant 0.12
was calibrated for 25 year time span, or a "generation"), or 3850 years
(with a correction for back mutations, the list of corrections was published
in J. Genet. Geneal., 2009). You can figure out a margin of error in this
case, for 655 of 67 marker haplotypes with 10,468 mutations in them. It is
not too wide. Please, calculate the margin of error, if you can (I seriously
doubt it), and I will correct you in a friendly manner.

By the way, among those 655 haplotypes were 34 base (that is, identical)
haplotypes in the 12-marker format. This gives [ln(655/34)]/0.022 = 134
"generations", that is 3875 years to a common ancestor. Compare with 133
"generations" and with 3850 years above, and you understand better the
accuracy of the methodology and strict randomness of mutations in large
haplotype series.

Now you understand, what "accurate" means?. What is "accurate" in your
quotation above?? "Accurate" for what?

The same goes to the rest in your "comments". Just (negative) words, no
data, no substance.

Do you want the same illustration with 464 of 67 marker P312 haplotypes? Ask
me. I will gladly help you.

Anatole Klyosov


This thread: