Archiver > GENEALOGY-DNA > 2010-02 > 1267011475

From: Alan R <>
Subject: Re: [DNA] : low variance MRCA dates for P310 clades in Italy andSEEurope
Date: Wed, 24 Feb 2010 11:37:55 +0000 (GMT)
References: <><009901cab349$8797d6a0$5e82af48@Ken1><><><><011601cab4e0$3fb4adc0$5e82af48@Ken1>
In-Reply-To: <011601cab4e0$3fb4adc0$5e82af48@Ken1>

I fully take on board all the comments about sample issues, confidence intervals etc.  I would like to suggest one last thing to try and get around the sample issue and in the hope of getting results sufficiently distinct to overcome the confidence interval issue.   Could the MRCA calculations for S116* and U152 calculations be regrouped into two simple geographical groupings and re-run.

1. Counties with substantial Mediterranean or Mediterranean-linked coast (I suppose Turkey, Greece, the Adriatic Balkans countries, Italy, Spain and Portugal)

2. Countries without a substantial Mediterranean or Mediterranean linked coastal area (all the rest)

The one problem county would be France as it has Mediterranean, central European and north European aspects.  Unless one wants to divide the Mediterranean French samples off, I think France is best put in the second of the above two categories as only a very small portion of the country is Mediterranean coastal due to the mountains and other geographical aspects. 

I think the justification of this approach is:

1. A simple division is the best way of getting a decent sample for each category
2. By and large east-west movements were divided into those two groups due to the geography of Europe.

I think too that previous divisions had blurred this a little.  For example a category like eastern or SE Europe can include both landlocked central European/Danubian places like Bulgaria and Mediterranean accessible areas like Croatia.  I think that sort of mixing blurs the issue by mixing countries that would have been divided by the main migration routes as noted above.  My suggested simple division is of course a very broad brush division and there are aspect where it is imperfect but I think it has several virtues. 


This thread: