DNA question

Question

This is something Ive never quite understood. Is having DNA confirmed by Wikitree proof of the relationship or just proof that that is the result of the majority of those who’ve researched that family or of my research?

Below is my mother profile # and the men I have fairly reliable proof she has decended from, except for George A. Perry.da If he was in fact my ggreat grandfather, he was a bigamist and went by George A. (of Tennessee)  and  George W. (of Michigan). I grew up long before DNA being told George left his family in Tennessee for long stretches of time and was from up North in Michigan originally. He eventually went North and was never seen again and declared dead in Detroit. But I think I have picked up the second family in Michigan. I approached them years ago when I had less information, but I want to try again. Does the info below show proof of parentage?

My mom Perry-7785   100%

Henry Perry-7997   50%

George A. Perry-7998   25%

Samuel L. Perry-7990   12.5%

Ichabod Perry-7991  6.25%

Samuel Perry-7992  3.12%

Joseph Perry-5123   1.56%

Richard Perry-207  .73%

Comments

Susan, Kathie Forbes explained this really well. The only reason those percentages are displayed on each of those profiles is because they have been connected to one another on WikiTree. WikiTree has no idea if your mother shares that much DNA with them. If you remove one of them, and change him to a completely different person, the same % of shared DNA would then display on the profile of that completely different person, based on what WikiTree sees as her relationship to that person.

Answer 1

Confirmation is not *proof*, and relies upon traditional genealogical evidence as well. If a genealogy is wrong, the DNA confirmation is wrong too.

Think of it as "this DNA evidence is *consistent with* these genealogical conclusions".

The only relationships that are provable with DNA are really really close ones  e.g., when both parent and child have tested.

Comments

Ok, that makes sense.  So these percentages are really useless. I may not have broken through a brick wall at all.

---

I wouldn't say that the percentages are useless -- they are the average, and so are useful for comparison. If you only match 20% with a sibling, then you know they can't be a full sib, but a half (sharing one parent, not two). And so forth. Rather than averages, though, I use the Shared cM Project: https://dnapainter.com/tools/sharedcmv4 to check that what I have found on my tree is consistent with the shared DNA. Of course, the more distant relationships overlap, but DNA is only useful with research within the records, so that's one reason Wikitree is awesome. All evidence can be presented and analyzed so that the relationships can be considered proven.

Answer 2

Wikitree doesn't store any actual DNA information, just what tests people have taken and who their trees say they are related to.  Those % just represent how much DNA two people who are related to each other might share based on the relationship.  You need to use a third party site to see if any of the people listed as possible DNA connections actually share any DNA.  

When a profile is marked "Confirmed with DNA" that means that an actual DNA comparison has taken place elsewhere. It should be described exactly, for example, "Paternal relationship confirmed by Gedmatch match of 3586.9cm between kits T248947 (Joe) and A340796 (Kathie) at a genetic distance of 1."

 If it's more than a few generations back then three people have to match in order to confirm the connection.

Comments

Ok, so in this line I have DNA tests for my mom, her two children, and two of her grandchildren (my brothers). They are all registered on GEDMatch. Is this enough info to confirm my great grandfather? Im assuming it is triangulation I need to do - would it be best to do mother and both kids, or mother and child and grandchild?  In that case better to go with grandmother, son, and his daughter or mix it up, my mother, me, my brothers daughter?  Once I establish him, further back the information is documented really well.  But my great grandfather could be multiple people. He obviously did not give a hoot about family trees!

---

You already know that you, your mother, and your children are related to one another, but that doesn't tell you who earlier generations were.  You need to find some other people outside of your immediate family who believe they are descendants of George and compare your mother's DNA with theirs.  Have any of those possible relatives in Michigan done any DNA tests?  If they are documented descendants of the Michigan man then they are the people you need to match with.

---

To be more precise, if you have appropriate DNA matches to your mother, your brother, and your nephews, that confirms the biological relationship that you undoubtedly always believed you had. If you DON'T happen to match your mother, that would indicate that you're adopted, or something like that. If your match to your brother isn't what it should be, it might mean that you had different biological fathers.

But chances are, they're all turning out to be biologically related exactly the way they're supposed to be. That doesn't mean that it didn't tell you anything - it's just that it's confirming what you already thought you knew.

For the purpose of pursuing the kind of investigation you're talking about, however, if you have access to your mother's DNA then you should only use that one. Everybody else (you, your brother, and his kids) can only have SOME of the DNA she has, so they are all less useful than her test, and none of them tell you anything that hers doesn't.

The answer lies in who you mother matches. Is there somebody within her matches who is a descendant of that suspected "second family"? That's the question. If not, you might try asking them whether they have tested (with the same company). Anybody who is at least a half-second cousin to your mom will almost definitely show up as a DNA match to her. If they haven't tested, see if they'll test (offering to pay for it yourself might help sometimes).