Archiver > GENEALOGY-DNA > 2012-08 > 1344477731

Subject: Re: [DNA] Dienekes y-dna age estimates
Date: Thu, 9 Aug 2012 02:02:11 +0000 (UTC)
In-Reply-To: <>

>From: argiedude < >
>I obtained identical estimates as you did in all the ages you posted, but I didn't take into account this issue (divide result by 2 to account for mutations accumulating on 2 chromosomes after splitting apart).
>For example, take I2a1. Your graph gives an age estimate of 23 kya, by comparing L158 with L178. I compared ...(L158) with ...(L178) and obtained the following:
>Total SNPs...18692
Valid SNPs (didn't have a no call in either chromosome)...17146
Variant SNPs...240
>...(the rate used by Dienekes for his estimate)...3x10^-8
Generation used by Dienekes...25 years
Total bases in 1000Genomes data (approximately)...9,300,000

240 / {[9,300,000 x (17146 / 18692)] x 3x10^-8} x 25 = 23.4 kya

I didn't divide by 2 at any point and my result was the same.

> Dienekes response: (Tue, 7 Aug 2012 08:55:38 +0300)
> I remembered.




VERY interesting. The "Argiedude" data shed light on a mystery with the SNP-based calculations conducted by Pontikos. If his response "I remembered" was truthful, then he must have obtained 46,000 ybp for I2a1, divided by 2, and finally obtained 23,000 ybp. However, for the same set of SNP Argidude obtained 23,400 ybp, which, after division by 2 should results in 11,700 ybp. 


Since argiedude has clearly showed how he obtained the dta, and Pontikos did not, it is time to know  how he did it.

This can explain the error which Pontikos obtained with the "age" of R1b1a2a1a1a5 of 7,000 ybp, since he apparently did not divide it by 2, to obtain 3,500 ybp, the likely "age" for the subclade, as I have noticed earlier. He did not want to admit it, and never answered the question on the origin of that 7,000 ybp, which Didier and myself repeatedly addressed to Pontikos. 


If so, most or all the figures on the Pontikos list are in error, and represent double figures (100% error). If so, I retract my statement that SNP-calculated numbers by Pontikos give the same numbers as our STR-based calculations, because his numbers are erroneous, either most of them or all of them.


There is one way only for him to explain the situation: present the calculations for both the R1b1a2a1a1a5 subclade, and for some other haplogroups and subclades on his list. I would not have pressed for the matter, however, Pontikos poured a lot of venom on my STR-based calculations, at his blog and at this Forum. Now, it turned out that HIS data are highly questionable.

Anatole Klyosov



This thread: