GEDmatch Genesis Multiple Kit Matches

Question

I was using the "People who match both, or 1 of 2 kits" tool on GEDmatch Genesis today for the first time. Out of curiosity, I checked my kit against my son-in-law's kit using the default cutoff levels  of 10 cM.  This produced 20 results that matched both of the kits even though my son-in-law has wildly different ancestors and geographical locations. I checked the matches using the one-to-one tool and got the same results.  (BTW, I'm not talking about triangulated matches, just kits that matched both me and my son-in-law at over 10 cM.)  This puzzled me, so I checked several random kits and always got a dozen or more results that matched both kits.  Does this mean that IBS matches are much more likely than I previously thought, even at segments over 10cM? If I increased the cut-off to 20 CM, then the I got the expected result of no matches with both kits.  I knew under 10 cM was very possibly IBS but I was accepting matches between 15 cM and 20 cM as almost certainly IBD.  Seems like that may not be the case. Thoughts?

Answer 1

Mardon,

I have checked hundreds, (probably closer to several thousand I have over 1500 that I have working on in my Exel file), of matches using Gedmatch.  As you seem to be seeing solid matches it seems that the most likely answer is that there is a connection and you will need to expand both of your family files to find it.  I have several hundred matches where I can not find the common ancestors but I have no reason to believe so far that they do not exist.  I am confident that the matches are for real cousins.

While not appearing to have a common geographic location seems like an issue - people move a lot.  I have one woman in my ancestry that married four time, her husbands were born in Ireland, Austria, Sweden and Australia.  She was born in Georgia, USA.  One of her marriages was in Canada and the rest in the United States.

Comments

Thank you to everyone who replied.  If nothing else, this thread has helped convince me of the truth in the statement:  "it's a small world".

I thought it would be interesting to use the "People who match both, or 1 of 2 kits" tool to determine how many people match both my GEDmatch kit and the kit of one of the people who replied to this thread.  I used Philip Smith for the experiment because his GEDmatch kit number was readily available through his DNA confirmations.  The results showed that there are 24 people in the GEDmatch database who match both me and Philip at a level of 10 cMs or higher.  Interestingly, the first person in the match list is Kathy, a second cousin of mine once removed.  I know that Kathy and I share MRCAs of Elisha Brown 1817-1893 and Marion Eliza Hills 1832-1916.  The match between Kathy and Philip is on chromosome pair 6.  The matches between Kathy and me are on chromosome pairs 3, 11 and 17.  None of this suggests that Philip is related to me, nor does it offer any proof that Philip is related to Elisha and/or Marion. But it does mean that Kathy is related to both me and to Phillip (as are 23 other people in the GEDmatch database).  I'm still having a hard time getting my head around the idea of how small this makes the world seem.

I also wanted to clarify my comment about triangulation in my OP and emphasize that the "matches both tool" is not about triangulation.  As I understand it, triangulation is the process of comparing three segments that are known to match at the same genomic position on the same chromosome pair. The purpose is to find out if all three segments match each other or if two of the segment matches occur on one chromosome in the pair while the other match occurs on the other chromosome in the pair.   This situation is far removed from the "matching both tool" where someone matches both of two people but maybe even on different chromosome pairs or at different genomic positions on the same chromosome pair.

Thanks again for the feedback.  Mardon

Answer 2

Really, this is about a very subtle issue that seems especially hard for people to grasp. I have certain second cousins who I would certainly match if they saw fit to actually get tested. My good friend and former co-worker would ALSO match these people, even though he and I are not related. [Well, he's definitely at LEAST 4th or 5th cousins, if we are related! ;) Plus, he isn't in my matches.]

The thing is, I discovered that one of my dad's 1st cousins (who happens to share our surname) married a 1st cousin of my friend's dad (who happens to share HIS surname). Their kids are MY 2Cs - and have my surname; but they are also my FRIEND'S 2Cs - and their mom's maiden name is his surname.

Just because we're related to a few of the same people, that DOESN'T make the two of us related to EACH OTHER.

At 10cM, though, were talking about the same thing, but at a much more distant level. At LEAST 3C,and most are probably more like 6Cs.

So what it's about is that you might have countless thousands of 6Cs, a fraction of whom of have tested and matched you. But THEY, in turn, have countless thousands of 6Cs, too - including your son-in-law, in a relative handful of cases.

It may seem counterintuitive, but it really makes perfect sense. In fact, THIS is EXACTLY why they demand that you match sufficiently distant relations on the exact same segment, for triangulation!

Answer 3

If your family tree was not reconnecting in the past then at 6th gen you have 2^6 = 64 ancestors, or 32 pairs.  Assuming each of these 32 pairs had an average of 5 surviving kids and non intermarried, these 64 ancestors had 32 * 5^5 = 100,000 6th cousins.  (or 5th or 7th - I get fuzzy with cousins vs. generations).  My home town - and most of its surroundings - never exceeded a population of 100,000 since being founded in 1780's.  

Almost everyone that has similar geographic roots as you is likely to be related at that genetic distance.  Take it to 10 generations and 1024 ancestors and 512 * 5^10 = 50,019,531,250 people or 7 times the world's population.  This is the real reason we all have royal ancestors.