A cluster of matches in South CArolina

[4b2]

I have a curious cluster of 6 matches on Ancestry from 13 to 28 cM who all share their descent from James Edward BARRINEAU (1822-1885) and Susan McCrea BROWN (1831-1857) with their ancestors also apparently from South Carolina.

Have a read through this: https://www.rootschat.com/forum/index.php?topic=891402.0

What is the largest cM match that is in this cluster? 28cM? Is there a larger match in the cluster you don't know how you are related to.

As outlined in the above post, when you have clusters that don't have any overlaps with larger cM matches, it's likely because they are more distant matches. You're talking more like 7th cousins, so ancestors in the window 1730. At this point the chance of inheriting that specific piece of DNA unbroken is quite small. So it's statistically unlikely to match any, of say, 40 matches that you are related to in the window 1775-.

I have mentioned this before, particularly if the tests are from someone from my grandparents generation, first 1/4 of the 20th C. I've noticed that there are lots clusters, I think all of which don't include matches that are above 30cM - I think 28cM is the biggest. And they all go back to the old US south, mostly to the Carolinas. And I don't think they have any British matches, or very very few. These clusters tend to be huge. I have one that has about 500 shared matches (so far as I can tell). Each match usually has about 100 shared matches.

I've come across numerous of these and don't know what to make of them. One thing I wonder is if they are false positives, where timber has not stripped out some DNA and the result is these huge IBS clusters.

One thing that makes me think that they are real is that I have two tests from mutal 2nd cousins on my grandmother's generational level, and they both have this same massive cluster from the Carolinas. If it was IBS, it would not be likely to show up for both.

I have found, maybe 50-60 matches in that cluster that have common ancestors - not all the same common ancestors. But most of them had the surname Mullinyeux across multiple generations. And the two tests mentioned do have Mollinyeux ancestry from Cambridgeshire. No US ancestry known.

But generally speaking, when I find those clusters it's just a complete mystery, with no obvious surname link.

One thing that makes them suspicious is 1) the size; and 2) that they always seem to be from the old south and not, say, Massachusetts, Maine or New York. The size of them suggests it could be an issue with tinder.

Going back to what I initially mention about these clusters with smaller maximum cM - my rule, assuming the test subject is born in the window 1950, that if the maximum cM in a cluster is X, the MRCAs are in the window Y:

X => Y

< 20cM => 1700
30cM => 1730
35cM => 1760
40cM => 1775-1800

That's a general rule. It could be something different. The reason I have it is I group clusters by location ans assign them an approx window for MRCAs. This lets me guess where they may link in to a tree. See attachment. You can place them where they better look like they fit in and then look for possible marriages between the clusters.

Also worth keeping in mind is that when you find MRCAs in the matches of these clusters with a smaller max. cM, is that there can be multiple generations between the MRCAs in your matches and the overall MRCAs. As it just happens that a few people inherited a big chunk from one ancestor and it stays in a few descendants to a level that can be useful in aDNA.

So in your example, you have MRCAs in the tests in the window 1830, but if the biggest match is 28cM, your MRCA (assuming its not a false positive) is more likely to be in the window 1730.

It's also worth considering that the genealogical record from the old south is not good. I'm not sure what sources they use, but I don't think there are BMDs, or few. Many southern trees are highly suspect. In going through many on Ancestry you can see many variations. So it's just too much of a mess to really be worth putting time into. And how would you know if there was an immigrant ancestor to the Carolinas? Unlikely there is a source. I believe many people were transported to the south up to about 1776. So there would be a lot of possible links. But it's most likely any amateur trees botch in the nearest possible match in the Americas, rather than the immigrant ancestor, for whom there may be no source of the immigration.

I do have one of these clusters where there are MRCAs in the matches with a surname from my ancestry. Maybe that's the link, maybe it's not.

Generally speaking, US matches where the link is prior 1800 are a wild goose chase.

[David Nicoll]

I have found, maybe 50-60 matches in that cluster that have common ancestors - not all the same common ancestors. But most of them had the surname Mullinyeux across multiple generations. And the two tests mentioned do have Mollinyeux ancestry from Cambridgeshire. No US ancestry known.

But generally speaking, when I find those clusters it's just a complete mystery, with no obvious surname link.

One thing that makes them suspicious is 1) the size; and 2) that they always seem to be from the old south and not, say, Massachusetts, Maine or New York. The size of them suggests it could be an issue with tinder.

Hi, 4b2 and OP.

    I think this is exactly what you would expect with pedigree collapse. As far as I understand from there documents, Ancestry assume that no people are related for Tinder. Whilst this may be true for the last 100 years or so in the USA it’s not pretty much everywhere else.
    The south was relatively sparsely populated like much of rural USA so much like Europe if your family was in an area early, the the chances of pedigree collapse are much higher, so you will have 20cM matches who are closer to 10cM matches in time. You can see you are related, but it is before records began or before the 1600’s even. So it is not USA ancestry so much as USA descent, someone emigrated a long time ago and you are seeing that signature in the resulting US population.

Happy Hunting

[4b2]

    I think this is exactly what you would expect with pedigree collapse...
    The south was relatively sparsely populated like much of rural USA so much like Europe if your family was in an area early, the the chances of pedigree collapse are much higher, so you will have 20cM matches who are closer to 10cM matches in time. You can see you are related, but it is before records began or before the 1600’s even. So it is not USA ancestry so much as USA descent, someone emigrated a long time ago and you are seeing that signature in the resulting US population.

Happy Hunting

Your suggestion is that there is a relatively small pool of ancestors in the south/Carolinas, that lead to a greater representation of the footprints of certain ancestors in the gene pool, and thus these matches could be more distant?

What are your thoughts on when false positives creep into Ancestry? At what cM level. a quick Grok give >= 15cM should be about 0% false positives.

[Petros]

Hi 4b2

Thanks for the comments and insights on the issues. It is likely to remain a puzzle to me as I don't know which of the South Carolina couple born in the 1820s is the connection nor do I have any insight on who from a number of my ancestral lines disappeared from England to the US in the 1700s.

I have found a couple of other large clusters of matches in the US but in those I can trace the connections. One goes back to a 4 x GGF whose grandson emigrated to Canada in the 1840s. This yields a number of relatively high matches including a 5C2R who is a 44 cM match! Another cluster also went via CAnada to Salt Lake City

[David Nicoll]

4b2 and Petros,

   In order of the posts!

@4b2

Yes, absolultely. In a simple case I have someone who's match is twice that of his father. He is a similar match through his mothers side. So he looks as if he should be a relatively recent match, the reality is something close to 1740 for the MRCA's.

I am unsure about "false" positives. Looking at the companies claims they are 1%, some of the equipment companies seem to be suggesting nearer 0.5%. even so 1 in a hundred is a pretty small number, and these are going to be singleton matches. Anybody with shared matches are pretty much guaranteed to be real in some way. I think a lot of what people call false matches are simply the background noise of shared ancestry from a common region.

@Petros

I think this is exactly what you are seeing with your Canadian and US relatives, compounded by the fact that at least in my case there was a trailblazer who emigrated and then was followed by siblings, cousins and family friends, and even these were probably relatives 3 to 5 generations back, so there is a background wash of common DNA, in my case NE Scotland.

I think what a lot of people are calling "false" matches are in fact this background noise, with patience you can find matches or at least possible family lines.

Ancestry is even finding me pretty certain common ancestors with as low as 8.3 cM matches so it is possible, and I have a lot where the MRCA is in the 1740-1770 range.

Happy Hunting

[Petros]

David
Thanks for the comments. There are certainly oddities that come through even in England with some 7C showing up. My wife and I share a common ancestor born in 1751 my 4xGGF her 6x GGF and both have some matches to this line. She has one 24 cM match to a descendant of our ancestor who isn't a match to me and I have multiple matches to 5C and 5C1R on the line from her 5 xGGM none of whom are matches to her. And the converse is also true for that line

[David Nicoll]

You’re very welcome, our ancestors did enjoy making life difficult for us !

David

[Biggles50]

To echo what has been written.

I too have very low cM matches that Ancestry has linked via Thruline/Common Ancestors.

Lowest goes back years to when Ancestry gave us 7cM matches, in my case this was correct after due diligence and adding each person with BMD, land records, census returns etc for each.

I also have a batch of DNA Cousins in the trees we manage in the 8 to 10cM range so yes they can and have been validated as accurate.

Conversely two low 8cM and a couple of 9cM Paternal matches completely threw me for a couple of years, they turned out to me false positives.