Too good to be true?

[Biggles50]

I have another tiny match that is about 10cMs and an 8th cousin which I cling to as the only link to a particular line so yes, it happens. The FH software described this latest match as 9c6r and that just seemed so obscure when you read that every generation halves the DNA pool.  And Ancestry says a paternal match and it's definitely not! 

On the one hand there are hard and fast rules about how DNA is shared and I read a warning this week that we're unlikely to get proven links beyond the 4c level.  Clearly that's not true.  And clearly some genes are stronger little blighters than others as I've got several lines without any proven DNA links and others with quite obscure DNA connections.

With very, very distant matches how I view them is that there may well be a pedigree collapse that has occurred with the result of more DNA being inherited down a path.

I have a 364cM match plus her two children have also taken DNA tests, two First Cousins of this 364cM match have also tested & they show as full Cousin DNA matches to my 364cM match.  These two are half Sisters and share 95 & 120 cM with me. Our MRCA has to be a couple who were born in about 1860 as other DNA matches like to them and to their parents.  So I have a 364cM match and her children all sharing more DNA with me than the expected, ergo, there is pedigree collapse in her line but there is no documentation only a hypothesis.

[Steve3180]

Nan,
Referring to your first post, it's perfectly possible to find common ancestors for an 11cM match, I have many at that level and below. However they're not all equally certain, I have a rough and ready scoring system for my matches based on good paper trail and dna matches coming in closer to both me and the match in question. So if you have what I think of as supporting legs to the tree all the way back to the common ancestors then I would consider that as a much more certain route than a single match that has no support.

On the Ancestry 3rd cousin thing I would urge you and anybody else to ignore Ancestry's suggestions for low matches. They have a hard cut off at 4th cousin and have calculated probabilities as if the chances of finding 5th cousins and beyond is zero, which is just nonsense. They suggest my lowest match at 8cM has a 69% chance of being in the 4th cousin group, this is wildly wrong. There is published work showing that a match under 20cM is more likely to be a 10th cousin than a 4th. While the chances of detecting a 10th cousin are very low the number of 10th cousins you are likely to have is enormous (millions) and the expected number that can be detected is one multiplied by the other.

On the 'too far back' thing in reality there is no such thing, all our dna comes from our parents, and all their dna comes from their parents, etc etc. The problem comes in the testing which introduces a lot of noise into the system. We will all have perfectly good dna matches who are our actual cousins and who match at 1 or 2 cM, the problem is distinguishing them from the false matches introduced by the testing process. As better testing becomes cheaper then these false matches will decrease. The annoying thing I find is that we already have better testing available and the companies won't give you the info. If we could combine MyHeritage's triangulation and chromosome browser with Ancestry's paternal phasing and GedMatch's toolset then we would find many more dna matches.
Steve

[Norfolk Nan]

Thanks all.   I don't take these very small and distant connections too seriously, particularly when there are no shared matches with others.  I mentioned it because I thought it was amusing and showed how imprecise the 'details' can be.  In both cases, the paper trail applies and I'm happy with that - it's old school and the best we ever relied upon until DNA was let out of the box.  All in all, it's an interesting additional source of thought and came about quite by accident, which makes it all the more amusing. 

Now if I could crack the significant 1c and 2c mystery DNA matches, I'd be a very happy bunny. 

[Steve3180]

Biggles,
I have come to think that pedigree collapse is much more common in a British Isles population than the literature, which is mostly American, would have us believe. I have two great grandparent lines where nearly all the matches have far more dna than the relationships would suggest. In America nearly everybody's ancestors come from somewhere else, whereas mine come from four or five small areas in Britain, probably going back many hundreds of years.
I realise this is a generality based on just my tree, but I have seen many comments on this forum and others that would suggest it is fairly common. I certainly have cousin marriages both 1st and 2nd, an uncle niece marriage, and a double-cousin marriage. This sort of thing would only happen in a static population such as Britain before the industrial revolution.
Steve

[Glen in Tinsel Kni]

Quite a few of mine were from villages around Lincoln and worked as ag labs then settled in the city when the foundries opened, some strays went to Sheffield but in both places they seemed to marry within their own to some extent, the same surnames keep popping up and it's only the stupidly wide tree I've created that allows me to see the multiple connections. It doesn't seem to inflate the cM to any degree but that may be due to my parents being NPE's so some matches link to me in multiple ways but are only half related to start with.

[brigidmac]

The small matches can be taken seriously when you have several who descend from different siblings . I put  asterisks  in the suffix  box of person profile to show dna connections then add an asterix if there is a dna match to another of their children even if the match doesn't show on. Thru lines .

For example my maternal.grandparents each of 3 asterisks because cousins have tested from 2 of my aunts and I've tested
My paternal grandparents had 2 children and a cousin has tested

on great grandparents side there were children from first marriages
There are Dna matches to descendants of
 1  child from Ggfather s first marriages
2 children from ggmothers first marriage
& 2 children from their mutual marriages

With distant matches i.believe this helps confirm what dna results suggest

[Petros]

It is certainly possible and I have a few matches at the 6C and 7C level.  Two 6C at 20 and 30 cM to the family line mentioned below.However, even the 5C level doesn't give consistent matches.

I have 6 matches to descendants of my GGGGF's daughter at the 9-21 cM level, all though one of her sons. This daughter is also my wife's GGGGGM and she has a similar number of matches but all through different children!
I also have a small cluster of 6 descendants (5C-5C2R) in the 9-11 cM range who are all descended from one individual, who emigrated to the US. This individual being a grandson of my GGGGF's youngest son. Yet I have no confirmed matches to descendants of the 10 other children of the youngest son.

[Biggles50]

Biggles,
I have come to think that pedigree collapse is much more common in a British Isles population than the literature, which is mostly American, would have us believe. I have two great grandparent lines where nearly all the matches have far more dna than the relationships would suggest. In America nearly everybody's ancestors come from somewhere else, whereas mine come from four or five small areas in Britain, probably going back many hundreds of years.
I realise this is a generality based on just my tree, but I have seen many comments on this forum and others that would suggest it is fairly common. I certainly have cousin marriages both 1st and 2nd, an uncle niece marriage, and a double-cousin marriage. This sort of thing would only happen in a static population such as Britain before the industrial revolution.
Steve

Steve

I agree with your conclusions.

Whilst I am a strong advocate of building a tree with DNA validation of the lineage, if I have not found the documentation etc along the line there is always a little doubt. There again just because there is a certificate or baptism record does not mean it is accurate.

I am lucky, Grand Parents, were from Ireland, Italy, Cheshire and Yorkshire.

I do have a tree where my Wife had a verbal conversation one of the people who is now in the tree said that she married her Cousin.  She did, we found that she married one of her 2C’s.  Now the person concerned passed away in 1989 but her GG Grandson has just taken a DNA test and I am working through his matches.  The MRCA between the two 2C’s were both born in 1830 and the DNA that their GG GS has may show interesting “Cousins”.

Edit - pedigree collapse actual DNA.

The DNA we manage with a Pedigree Collapse is a 3C3R to one DNA Match who they share 53cM with and have the MRCA quoted above where DNA Painter gives a mean of 27cM.

Another shares 39cM and is a 4C2R and the DNA Painter mean is 22cM.

Another shares 34cM and is a 4C2R and the DNA Painter mean is 22cM.

[Petros]

certificates often lie.

My wife's GGM was born in 1882 and shown in the 1891 census and her marriage certificate as the daughter of her grandfather. Only the birth certificate revealed that her mother was an unmarried daughter of the said grandfather.