Dear Belinda, You make some valid points. A point I would like to make is that endogamy is population specific. There is a lot of admixture in some populations and little in others. Even in the U. S. you have some areas and populations where there is a lot of endogamy and others where there is little. As a general rule, the Americans who live on the West Coast (outside Utah) have less endogamy than those who live on the East Coast who have early Colonial American roots. If you want to take a look at my mom's tree at https://www.ancestry.com/family-tree/tree/73764861/family/pedigree?cfpid=322 84129043&selnode=1 you can get a perspective about how diverse her ancestral background is. She has ancestors from Wales, Scotland, England, and North Ireland (Scots-Irish), as well as a little Native American ancestry. Some of her ancestral lines are Colonial American, but even those lines as scattered up and down the East Coast. When I need to do quantitative analysis to draw genealogical conclusions about probable genealogical relationships I can do that much more readily and accurately for my mom's relatives than I can with my dad, who comes from a Low German Mennonite background. The importance of chromosome mapping can't be overemphasized since you can't use segments linked to one ancestral branch as part of the quantitative analysis of a genealogical research question for a different ancestral branch. Sincerely, Tim -----Original Message----- From: GENEALOGY-DNA [mailto:[email protected]] On Behalf Of Belinda Dettmann Sent: Thursday, September 28, 2017 3:34 PM To: [email protected] Subject: [DNA] Endogamy does exist in USA data It is important to realize that DNA testers with predominantly US ancestry come from an endogamous background, compared with the rest of the world. That is, there is a degree of inbreeding in the USA, compared with the rest of the world. This may not be apparent to local researchers, as most of the statistics are calculated from USA data, where endogamy has been ASSUMED to be zero. It is small indeed when compared to Ashkenazi or French-Canadian populations, but it is larger than in mixed populations generally. One of the side-effects of this is that, although USA DNA testers get relationships predicted at roughly the right level, outsiders who share ancestral, but not modern, populations with USA testers, tend to get more US cousins than they would expect, and these cousins are predicted to be closer relatives than they really are. This occurs because the degrees of relationship are calibrated from USA data, so that testers from other areas, or from mixed populations, get misleadingly close predictions. This remark was prompted by Tim Janzen's remark earlier today that " My mom fortunately doesn't come from an endogamous ancestral background, which is helpful when working with her triangulated groups. Endogamy definitely complicates things." I have an enormous regard for Tim's knowledge and acumen when it comes to DNA matters, but this statement needs qualification. It can be regarded as correct in an American context. It is NOT appropriate for testers whose ancestors never lived in the USA, who but share older populations with them, such as Australians, Britishers, Irish, Germans, Poles, etc.

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. -----Original Message----- From: GENEALOGY-DNA [mailto:[email protected]] On Behalf Of Tim Janzen Subject: Re: [DNA] Endogamy does exist in USA data Dear Belinda, You make some valid points. A point I would like to make is that endogamy is population specific. There is a lot of admixture in some populations and little in others. Even in the U. S. you have some areas and populations where there is a lot of endogamy and others where there is little. As a general rule, the Americans who live on the West Coast (outside Utah) have less endogamy than those who live on the East Coast who have early Colonial American roots... The importance of chromosome mapping can't be overemphasized since you can't use segments linked to one ancestral branch as part of the quantitative analysis of a genealogical research question for a different ancestral branch. Sincerely, Tim -----Original Message----- From: GENEALOGY-DNA [mailto:[email protected]] On Behalf Of Belinda Dettmann Subject: [DNA] Endogamy does exist in USA data It is important to realize that DNA testers with predominantly US ancestry come from an endogamous background, compared with the rest of the world. That is, there is a degree of inbreeding in the USA, compared with the rest of the world. This may not be apparent to local researchers, as most of the statistics are calculated from USA data, where endogamy has been ASSUMED to be zero. It is small indeed when compared to Ashkenazi or French-Canadian populations, but it is larger than in mixed populations generally.

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.

I agree with Belinda that for us with, like in my case, purely European ancestors (born in Germany) the estimates on relationships are always off from 3rd cousin onwards. As the large majority of customer that were used to build up the models for relationship estimation is coming from the US, these models aren't accurate enough for us. To give a practical example, two verified DNA cousins of mine were able to quickly identify a common ancestor (https://www.wikitree.com/wiki/Adams-22118) who is born relatively early (1799) I was not able so far to make a connection from my nearby ancestors which are also well documented. I even bought the family book of the Hirten church in the hope to identify the common ancestor through it as it goes another 100 years further back (or more for unknown births but documented marriages). There is always the chance of a NPE or that my DNA cousins have more than one common ancestor but so far no evidence. The genetic distance (cM) isn't much different for us. Andreas > On 28 Sep 2017, at 17:19, Belinda Dettmann <[email protected]> wrote: > > from