GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-12 > 1292253704


From:
Subject: Re: [DNA] NW European R1b from Iberia?
Date: Mon, 13 Dec 2010 15:21:44 +0000 (UTC)
In-Reply-To: <1451740972.603020.1292253529072.JavaMail.root@sz0002a.westchester.pa.mail.comcast.net>


>From: "Anatole Klyosov" < >


>> From: "Lawrence Mayka" < >
>>Even for a given region, the calculated estimate may vary
>> depending on the specific mix of samples.

>... Nobody guarantees that R-P312* haplotypes in Poland are ALL
descended from a P312* common ancestor who lived 3600 ybp necessarily in
Poland. This is related to ANY region, whatever it is and whatever subclade
is. All those "regional" estimates should be taken with a grain of salt, and
compared with that of the overall subclade.
(...)
>Instead of using such a general and inaccurate description as in the
quotation above, it would be much better and more productive to take P312*
datasets from Poland and Spain, and illustrate the "statement" in a good
fashion, with data and numbers.






My response (a continuation):



Here are some data on R-P312* from Poland and Spain, and on R-P312* overall. As a follow-up of the discussion, Robert Tarin kindly sent me 41 of R-P312* 67 marker haplotypes from Spain and Portugal (current and with ancestors there). Then, Tim Janzen sent me 277 of 67 marker P312* haplotypes, including 4 from Poland, 15 from Spain and 5 from Portugal. 



Besides, I have data for 464 of 67 marker R-P312 haplotypes and 750 of 19 marker R1b1b2 Iberian haplotypes. So, let's compare our notes. It would be also of interest to see, what as few as 4 Polish haplotypes can give us in that regard.



First, "naked" results:



464 of R-312 haplotypes from all over Europe gave 158 "generations" (25 year each, corrected for back mutations), which is 3950 years to a common ancestor (I do not give here margins of error in order not to divert attention of readers. This can be done later, if needed).      



750 of 19-marker haplotypes (from the paper of Adams et al, 2008) gave 150 generations, that is 3750 years to a common ancestor. These two fisures are within margin of error.



Since the 750-haplotypes contained 16 base haplotypes (that is identical in the dataset), the logarithmic method gives [ln(750/16)]/0.0285 = 135 -->156 generations, that 3900 years to a common ancestor. It is the same figure as those above. It also shows that there is only one common ancestor in the dataset.


 

Robert's 41 R-P312* Iberian haplotypes:



1. A quick  logarithmic method (base haplotypes counting, in the 12-marker format) gave  159 generations, that is 3975 years to a common ancestor.


2. The "linear method" (mutation counting) gave:



12-marker format, 145 generations, that is 3625 years.

25 marker format,  161 generations, that is 4025 years.

37 marker format,  150 generations, that is 3750 years.

67 marker format,  146 generations, that is 3650 years to a common ancestor.

As one can see, they all fit to each other and to the above, "historical data". The average between 25-, 37- and 67-marker haplotypes (12 markers I normally omit, since statistics is not there) give 3810+/-190 years from the common ancestor. The actual variance between those figures is +/-5.0%. Now you see why I object that I use too tight margins of error. Nevertheless, I use much larger margins of error at the calculated figures.


Now, Tim's data calculated for Poland and Spain/Portugal:



Poland, 4 of 67-marker haplotypes, 39 mutations in the 37-marker format and 50 mutations in the 67 marker format. It gives 39/4/0.09 = 108 -->121 "generations", that is 3025 years to a common ancestor, and 50/4/0.12 = 104 --> 116 "generations", that is 2900 years. The same thing. The margins of error would be around +/-20%. 



Spain, 15 haplotypes, for 12-, 25-, 37- and 67-marker haplotypes 3550, 3575, 3325 and 3550 years to a common ancestor, respectively.



Portugal, only 5 haplotypes, 3350, 4050, 3700, and 4075 years, respectively.



Spain+Portugal, 20 haplotypes, 3525, 3675, 3450, and 3675 years, respectively.



If more detailed illustration is needed, how specifically calculations were conducted, I can gladly show it.



Conclusions: R-P312* haplotypes in Spain+Portugal (and separately, albeit with a larger margin of error), around 3600 ybp on average of all haplotype panels (on Tim's dataset) and around 3760 ybp (on Robert's dataset), have the same "age" as R-312* overall in Europe, that is 3750-3950 ybp. A few Polish R-P312* haplotypes are "younger", 2900-3025 ybp.



Regards,




Anatole Klyosov

1. A quick  logarithmic method (base haplotypes counting, in the 12-marker format) gave  159 generations, that is 3975 years to a common ancestor.


2. The "linear method" (mutation counting) gave:



12-marker format, 145 generations, that is 3625 years.

25 marker format,  161 generations, that is 4025 years.

37 marker format,  150 generations, that is 3750 years.

67 marker format,  146 generations, that is 3650 years to a common ancestor.

As one can see, they all fit to each other and to the above, "historical data". The average between 25-, 37- and 67-marker haplotypes (12 markers I normally omit, since statistics is not there) give 3810+/-190 years from the common ancestor. The actual variance between those figures is +/-5.0%. Now you see why I object that I use too tight margins of error. Nevertheless, I use much larger margins of error at the calculated figures.


Now, Tim's data calculated for Poland and Spain/Portugal:



Poland, 4 of 67-marker haplotypes, 39 mutations in the 37-marker format and 50 mutations in the 67 marker format. It gives 39/4/0.09 = 108 -->121 "generations", that is 3025 years to a common ancestor, and 50/4/0.12 = 104 --> 116 "generations", that is 2900 years. The same thing. The margins of error would be around +/-20%. 



Spain, 15 haplotypes, for 12-, 25-, 37- and 67-marker haplotypes 3550, 3575, 3325 and 3550 years to a common ancestor, respectively.



Portugal, only 5 haplotypes, 3350, 4050, 3700, and 4075 years, respectively.



Spain+Portugal, 20 haplotypes, 3525, 3675, 3450, and 3675 years, respectively.



If more detailed illustration is needed, how specifically calculations were conducted, I can gladly show it.



Conclusions: R-P312* haplotypes in Spain+Portugal (and separately, albeit with a larger margin of error), around 3600 ybp on average of all haplotype panels (on Tim's dataset) and around 3760 ybp (on Robert's dataset), have the same "age" as R-312* overall in Europe, that is 3750-3950 ybp. A few Polish R-P312* haplotypes are "younger", 2900-3025 ybp.



Regards,




Anatole Klyosov


This thread: