Using GEDMATCH

[LarryBurford]

I have a distant cousin that just tested on Family Finder. He matches my Great Uncle on FTDNA at the mid 60cM level. He doesn't match myself or my mother. I know it's because the default limit on the amount of cM we have in common is larger that our matching segments. My question is that when I go to gedmatch and compare his kit to mine I get matches with him when I drop the minimum segment length to 4 cM and the SNP count minimum threshold to 300. Is this a legitimate match or will I start matching unrelated people by dropping these numbers to this level.

Larry

[John McCoy]

That sounds like a weak match, not suitable for proving a relationship. If it were me, I would look for other kits on GEDmatch that triangulate with any of your known relatives, concentrating on segments of 10 cM or more in length. Of course it helps to have ALL of the related kits on GEDmatch so you can use the tools there and investigate the matches in all combinations.

[ltd-jean-pull]

It's always a bit disappointing when a known relative doesn't match, but 2 generations down from your mutual match these things happen.

4cM and 300SNPs will come up with loads of matches that aren't really matches.

Have you had a look at the mutual matches of the cousin and your gt-uncle and seen if or where they match you?

[LarryBurford]

My Great Uncle and I have several matches that show up on the distant cousins matches also.

[LarryBurford]

So if I go to gedmatch and start adjust the setting how low can I go below the default settings and still get a match that is legitimate.

[MMaddi]

So if I go to gedmatch and start adjust the setting how low can I go below the default settings and still get a match that is legitimate.

That's a decision that you have to make yourself. None of the three commercial companies rely on segments below 5.0 as the basis for declaring a match, although 7.0 cM is regarded as far more reliable.

Actually, in the case of FTDNA, as long as the total shared DNA is equal to 20 cM and at least one of the segments is 7.0 cM, they will declare a match, even if the smaller segments are all lower than 5.0 cM. However, in what you reported in your first post, Family Finder is not finding a match with this distant cousin and you and your mother. So, the distant cousin hasn't met the threshold of a total of 20 cM shared DNA and the largest segment at least 7.0 cM.

The more you lower the threshold below 7.0 or 5.0 cM, the higher the chance that the segments shown do not really represent a common ancestor in the last few hundred years. You have to decide from the circumstances of the paper trail genealogy and how much DNA is shared how legitimate a "match" below 5.0 cM is.

Reading your original post again, you note that the distant cousin shares mid-60s cM DNA with your great-uncle. That's a fairly close relationship. I have a paper trail proven 3rd cousin, once removed with whom I share 62 cM as a Family Finder match. What's the largest segment your great-uncle and the distant cousin share? What is FTDNA estimating as the relationship for your great-uncle and this distant cousin?

If their estimate is something like 2nd-4th cousin and he is actually a 4th cousin of your great-uncle and the common ancestor is also shared with your mother, that would make the distant cousin and your mother 5th cousins. That's the level at which only about 10% of people will share enough DNA to be declared a match.

So, it's possible that due to the distance of the relationship, the distant cousin is actually related to you and your mother, but doesn't share enough DNA to be declared a match. However, you don't clearly state in your original post if you have a paper trail that is clearly identifying who the common ancestors are for the distant cousin and your great-uncle and how distant the common ancestors are. That would certainly help clarify the situation.

[mabrams]

IMHO, the defaults at GedMatch are already a bit on the low side. Personally I dont run into them very often, since my 2000 limit centers around 9 to 10 cM. And I look at them very carefully on the One to One.

Here is an article that is very leary of small segments.

http://thegeneticgenealogist.com/2014/12/02/small-matching-segments-friend-foe/

In a very recent blog, he calls segments less than 5 cM as Poison.

I think you should test close relatives who might get a set of stronger matches that you can work with. There are a lot of strategies that are preferable to looking for the smallest credible segment.

[LarryBurford]

This is what Gedmatch says when I run my great Uncle and the distant cousin.


Chr Start Location End Location Centimorgans (cM) SNPs
20 41,824,387 53,395,927 22.3 3,638
Largest segment = 22.3 cM
Total of segments > 7 cM = 22.3 cM
1 matching segments
Estimated number of generations to MRCA = 4.7

This is what FTDNA says
01/09/2017
2nd Cousin - 4th Cousin
67 shared cM
23 longest block


I know the ancestors in common with the great Uncle and cousin are 5 or 6 generations back but there are 2 or three different ways we are related.

[LarryBurford]

When I drop the to 5 on gedmatch I get this
Minimum threshold size to be included in total = 500 SNPs
Mismatch-bunching Limit = 250 SNPs
Minimum segment cM to be included in total = 5.0 cM


Chr Start Location End Location Centimorgans (cM) SNPs
9 132,315,293 135,637,287 6.5 1,018
20 41,824,387 53,395,927 22.3 3,638
Largest segment = 22.3 cM
Total of segments > 5 cM = 28.8 cM
2 matching segments


Minimum threshold size to be included in total = 500 SNPs
Mismatch-bunching Limit = 250 SNPs
Minimum segment cM to be included in total = 5.0 cM


This is my mothers brother compared to the cousin with the setting at 5
Chr Start Location End Location Centimorgans (cM) SNPs
9 35,556,752 70,401,020 6.8 924
9 132,315,293 135,633,060 6.5 983
Largest segment = 6.8 cM
Total of segments > 5 cM = 13.3 cM
2 matching segments

[mabrams]

If you want, you can quote these "sub 7 cM" matches since they do exist.

But I think you are better off just quoting the 22cM between your great uncle and the distant relative.

At this point I think you have proven a match to your great uncle but not to your mother or you. You can add these sub 7 cM matches as evidence that the distant cousin matches you and your mother but your proof is going to look like Swiss Cheese.

On the plus side, apparently you lowered the GedMatch SNPs to 500, so that is consistent with FTDNA. And 6.83 is almost 7 so there is some respectability here.

Back to my original sentiment. Here is another well respected genealogist and her opinion.

http://www.yourgeneticgenealogist.com/search?updated-min=2014-01-01T00:00:00-08:00&updated-max=2015-01-01T00:00:00-08:00&max-results=18

Okay she is talking about sub-5s and you have near-7s but I still think you are headed in the wrong direction.



"but there are 2 or three different ways we are related."
This also opens up the possibility that the 22 cM is via one lineage and the sub-7s are via another lineage.

I would just focus on the 22 cM block. And if you choose to quote the sub-7s, at least add a disclaimer.

[mabrams]

And another blog, only a few hours old. My last blog reference.

https://dna-explained.com/2017/01/19/concepts-segment-size-legitimate-and-false-matches/

Towards the end, she talks about segment size. 6.5cM (from Chr 9) would fall into the 30% legitimate category.

~~~~
I was thinking a little bit about your segments. If you ran your great uncle's kit, you might see all of the 6.5 cM matches on Chr 9. Not sure, because there are a lot of combinations to consider.

So your great uncle vs Distant cousin, you, your mother, mother's brother. I think you will see a lot of 6.5 cM matches on Chr 9.

I cant find it now, but I really feel FTDNA says you can use the smaller segments as secondary, supporting evidence but not as your primary match.

Even though you can see a match between you and your distant cousin, its not a primary segment. It falls into the secondary, supporting category.

~~~
In my previous post, I mentioned 6.83 which should be 6.5.

Also I mentioned that you lowered the SNPs to 500. Actually you have SNPs of 1018 and 924 so you dont have a SNP problem. Lots of SNPs. But the cM are still shaky.