Dear Andreas, What I mean is that if you run one-to-one comparisons between a cluster of known cousins and someone whose genealogical relationship to you is unknown, you can more accurately predict the genealogical relationship than you can by simply running a single one-to-one comparison. For example, if you take a look at the blog I wrote in 2012 at https://blog.23andme.com/ancestry/who-were-the-parents-of-jacob-youngman/ you will note that I made 5 one-to-one comparisons between people who had a known relationship to each other and Troy McCoy, whose relationship to my cousin cluster was uncertain prior to the start of that project. As you can see by looking at the data from those 5 comparisons, there was a wide range in terms of the amount of autosomal DNA that these cousins shared with Troy. However, the average was consistent with a half third cousin genealogical relationship, which was the degree of relationship I had hypothesized to be true prior to starting the DNA project. This type of analysis is definitely something that you could include in your app. Sincerely, Tim -----Original Message----- From: GENEALOGY-DNA [mailto:[email protected]] On Behalf Of Andreas West Sent: Friday, September 29, 2017 12:57 AM To: [email protected] Subject: Re: [DNA] Endogamy does exist in USA data Hi Tim, Do you mean by your remark that with more DNA cousins in the TG you can make better predictions for their relationship to each other than rather focusing on just one relationship estimate between two people? If you are interested to try your idea out, then let's take the discussion offline and I will give it a try in my web app. BTW, we do automatic Chromosome mapping. Unfortunately can't send a picture on this email list to show you. Andreas