Is there a way to prove or disprove if a person is a sister of my 8th great-grandfather by using dna?

Question

Could we be that specific?  How would the match differ if she were his first cousin?  We're trying to settle an 8 year long argument about how many siblings there were, and I am a total newbie when it comes to dna.  Thank you!!

Comments

If you acquired DNA from envelopes sent by ancestors about four generations ago there might be a slim chance. Otherwise, no.

Answer 1

Extremely unlikely.  Autosomal DNA is only reliable back three or four generations.  Only men have Y-DNA which can identify men who are direct-male line descendants, only women have MtDNA (men get it from their mothers, but can’t pass it on) so no joy there since you are trying to connect a brother and sister.  Paper trail is your best hope.

Comments

That is not strictly true; autosomal DNA is 100% reliable for determining relationships 4 generations back (you'll match 100% of your second cousins). Around 90% of third cousins will share autosomal DNA, around 50% of fourth cousins, and the numbers go down from there. There is a chance of shared autosomal DNA with an 8th or 9th cousin, but it's very low (1 in 100 or less); sometimes you get lucky and have what are sometimes called "sticky segments" (relatively large autosomal DNA segments that don't decrease due to recombination for many generations). I have one of these for 37cM shared in common with a group of people whose closest relationship to me is sixth cousin once removed or seventh cousin; my mother has a group of matches for around 22cM who are in the range of seventh cousin or so (the most distant ancestors I've been able to confirm with autosomal DNA are 7th great-grandparents).

---

Yes, I see that phrase occasionally - "Autosomal DNA is only reliable back three or four generations."

At best, it seems to me a misleading statement. Maybe, as "C Handy" suggests, the "reliable" part refers to how likely you are to get a match a particular cousin at a specified level of relationship. If so, that shows a fundamental misunderstanding of how auDNA is used in genealogy.

Matches of at least about 10cM are perfectly "reliable", even if it's a match to a 7th cousin (7C)!

By my count, "4 generations back" gets you to your great-great grandparents; so it's about your 3rd cousins (3C). That's the easy stuff! The further back stuff is where it really gets INTERESTING.

As to "sticky segments", I subscribe to the theory that everybody likely has a handful of these. I've identified four or five so far, for me, and that's just on my father's side! One of them is from too far back for me to even know how I'm related to those dozens of people, but I can tell it's on my gt-gt-gt grandfather's side. If it's not from his maternal grandparents (which would make them 6Cs), then it's even further back.

Answer 2

Possibly; but it depends on a number of factors...

If you have DNA from your Great Grandparents and the Great Grandparents of a relative who are descended from two siblings, then it might be possible to find a common ancestor.  The common ancestor being the parents of the sister and sibling that you are interested in.

I say this because my own father has found a common ancestor from 5th great grand parents.  The 5th great grandparents has at least 2 children.  My father is descended from one child and there is another person within Ancestry.com who is descended from another.  This is from matching only 6cM of DNA.

So, since this is my father's DNA match, then it would be my 6th Great Grandparents.  However, if instead of my father it were my Great Grandparents, then the siblings could in theory be 8th Great Grandparents.

In my father case, his Great Grand children are only toddlers right now.

Comments

Interesting, this idea of using a great-grandparent's DNA to go WAY back. Of course, that's only applies to the very young.

Really, though having a match to a 6C (MRCA = 5th great grandparents) is a positive indicator, but doesn't really prove anything. Something like half of 6cM segments are false positives - you need a triangulation (an additional distant relation), using a bigger segment. Plus, that only CONFIRMS a paper trail. Plus, this question is talking about one generation further back.

Answer 3

"Full sibling" is the one relation that DNA can tell you exactly. Even with a parent/child relation you need to have additional information beyond the DNA test itself (i.e., which one is older) to tell whether our match is a parent, or a child. For the cM value you see for parent/child, "self" and "identical twin" is also a possibility. But for full siblings, the range of values does not overlap other relations, and can't be anything else.

But for an 8th-gt grandparent? Just "No". You're lucky to even HAVE a match to a 9th cousin (my most distant match is an 8C), and without triangulation with a decent-sized segment, and a solid paper trail, you can't even CONFIRM the relation, much less PROVE anything!

At the 7C level and beyond, it's not a matter of how many cMs - you're only going to have one or two small segments that add up to a very small cM level. It's a matter of WHETHER you happen beat the odds to get that segment.

Even for 3C1R through 6C1R, the cMs usually aren't going to help you tell one relation from another. Below 20cM could be anything from 2C1R to 8C+; 20cM to about 40cM might be 2C1R to 6C1R; 40cM to 75cM might be 2C1R to 5C; 75cM to 140cM might be 2C to 3C.

Answer 4

I believe a brother and sister have the same mtdna.  If you can find an all female descendant line from each, they should match.  They could still be cousins through their mothers, if they match.

Comments

I agree.  I would try MtDNA using femaie descendants if you can.  If not you might be able to do hybrid approach.  Do MtDNA to get to a female descendant who is a cousin or within the Autosomal threshhold.  If you match to her and then she has MtDNA that leads to the sister you can intuit a DNA trail and then use the paper trail to support it.

Answer 5

Neat question.  I think if you tried hard you may be able to detect small groupings of au dna patterns going back to the ancestors in question.  For example using the DNA Painter website to keep track of things.  

That would be the first step, narrow down which segments of your au dna go back to that branch of your ancestors.  Then search to see if anyone whom you have dna matches with matches your dna on those segments.  Then try to figure out who your MCRA with that match is.  

It's a real long shot to actually prove siblings relationship to a 10 generations back brother and sister.  Like someone already said if it was two brothers you could use Y-DNA to compare.  

Good luck!