Is it possible to determine precise cousin relationship from DNA alone

Question

I’m trying to work out if person A is the 4th or 5th cousin of person B.

That is to say is person B descended from the daughter or niece of A’s ancestor? (Sister is likely precluded as a possibility due to the age difference)

Is this theoretically possible? What tools/websites would I need to use to work it out? How would I make a start on it?

I would have access to more than just the DNA of persons A and B.

Answer 1

Most DNA sites will tell you at this level 4th and 5th cousins,  you need to have traditional Genealogy to support any theories.  I certainly wish it was easier

Comments

There’s at least one theoretical way I can think of and that’s if A and B share so much DNA that’s it’s mathematically impossible to be 5th cousins (assuming being a cousin twice over can be ruled out).

I’m hoping there’s others because that doesn’t apply to my situation.

---

Others at WikiTree are much more convesant with DNA than I am.

I have quite a few mystery matches that really surprise me to the event that I DO NOT BELIEVE Pocahontas is my husbands' ancestor (considering firstly we have no US relatives).

I have watched countless videos, and read texts and now trying to understand WATO at DNA Painter and still rather lost.   Just a little less lost than before.  

One thing mentioned often is: matches that match you and another usually have the same ancestor somewhere along the line.

---

A neice has a different father than her aunt, so in theory, you need to determine who the father was. This can be done via DNA, however as Mark said, it is very difficult at the 4th/5th cousin level.

---

There is no theoretical upper limit to the amount of shared DNA for 4th or 5th cousins - it just gets lower and lower probability. Only close cousins can be done with a high degree of certainty.

For 100cM the number of segments would be also important. I am sure it would be like 6 or 7 or even more. More likely in those cases that there are other unknown cousin linkages - for example one segment is from a 6th cousin, one from a 10th cousin, etc.

The only way to help you close in on the relationship is to have other DNA matches who are close relatives to this remote cousin.

---

It can't be done. Consider, for example, that there is inherent randomness in the inheritance of DNA (consider that, while siblings share 50% on average, DNA testing shows that there's a wide range of possible sharing percentages) and that a DNA "matching segment" we see in a test does not mean that the two people inherited that segment from a common ancestor (it might include pieces from different origins that coincidentally happen to line up to appear to be a segment).

---

It can be done and it has nothing to do with matching segments. You do it the same way as a DNA adoption case, but you'll need to examine the tests of more than one descendant. I did this to discover who my 4th great-grandfather was, and then I found the one document that proves the relationship.

Answer 2

I think that would be possible if you can get 1st cousin matches with A to establish their branch, some second cousin matches would also be useful.
 If the first cousin matches are at the higher end of the range for first cousins that would be reasonable evidence, using WATO analysis including 3rd and 4th cousin matches would help.

Comments

Hi Gary

Interesting you mention being at the the higher end of the range.

Person A has an identified fourth cousin match with Person C from the ancestor in question, and they have 123cm in common. Which, by my understanding, is very close to the theoretical maximum of 139, based on here. Person B shares DNA with both Person A and C.

Also, Person A is a male line descendant of the ancestor.

Does any of that make the ability to find a solution any more likely do you think?

---

Not that sure I can nut that out at the moment, I'm a bit tired.
 Sounds like you have A and C sussed, the relevant question is does person B share DNA on the same segment as the other two, that is, is there a triangulated match?
 If not the answer is probably no, or at least there are more possibilities.
 The reason for saying high end of the range is that the further you go in 3rd, 4th, 5th cousins, the greater the overlap in ranges appears to be, a 3rd cousin range can start at 0cM (I have one at 10.7 with her cross matches to other 3C in the expected range), if you have matches at the high end of the range that makes more distant relationships unlikely.