20th cousins? So you're saying it *is* possible to use atDNA to confirm lines into medieval times -----Original Message----- From: Tim Janzen <[email protected]> To: genealogy-dna <[email protected]> Sent: Thu, Sep 28, 2017 8:26 pm Subject: Re: [DNA] Endogamy does exist in USA data Dear Belinda, I have been emphasizing in my presentations for quite a few years that trying to predict the precise genealogical connection between any two people who share say 50 cMs or less with each other is fraught with problems. I try to emphasize that the relationship could be as close as 2nd cousins or could be as distant as 20th cousins. You can improve the predictions if you start using cousin clusters to generate the prediction for the genealogical relationships. However, none of the companies are doing that yet. We need to see the companies automate chromosome mapping, but that hasn't happened yet either. Sincerely, Tim -----Original Message----- From: GENEALOGY-DNA [mailto:[email protected]] On Behalf Of Belinda Dettmann Sent: Thursday, September 28, 2017 5:19 PM To: [email protected] Subject: Re: [DNA] Endogamy does exist in USA data Thanks for that, Tim. The degree of endogamy becomes important in the context of predicting relationships. Examples are in the statistics and guidelines given in references such as ISOGG where values only apply to populations with a degree of endogamy similar to that found in the USA. As a practical guide, the commonly used predictions (at FTDNA, 23andMe and ancestry.com) for known relationships from third cousin outward are not correct for ancestry which comes from mixtures of populations with less inbreeding than in the USA overall. As an Australian with mixed ancestry from 8 different regions of the British Isles, plus Germany and Poland, I have thousands of DNA matches, for which, for third cousins and worse, the predictions are all wrong. The relationships are much more distant than predicted and most links are impossible to find. Ranges usually cover the right value, but the actual predictions are rubbish. Similar situations apply for testers who live in the British isles, or South Africa, or New Zealand, or Poland, to my certain knowledge. We all find it disconcerting when told how close our middle-distance matches should be, when we know they are nothing of the kind. I agree that chromosome mapping is essential in doubtful situations, and I routinely download DNA for my known relatives to Gedmatch for that purpose, if they agree. There again, the guidelines for length of segments need some adjustments, as I find that fewer long and more short segments are the order of the day for known cousins of third degree or more in testers from mixed populations. ------------------------------- To unsubscribe from the list, please send an email to [email protected] with the word 'unsubscribe' without the quotes in the subject and the body of the message