PDA

View Full Version : Using DNA Matches to Find an Ancestor?



markkendall1989
08-08-2015, 06:56 AM
So here's my dilemma:

One brick wall in my genealogy (rather far back, but still) is an illegitimate son, Antoni, who took the last name of his mother, Katherine.

However, I'm wondering if at least theoretically it would be possible to identify the father by the following methods:

1) Start with the DNA of my grandmother who is Antoni's great-great-granddaughter via a second wife

2) using "reverse genealogy," find someone who is a descendent of Antoni via his first wife (they seem to have lived in a different town/area before he was widowed and moved be with his second wife, so probably no other ancestors would overlap and confuse things)

3) having these two, we could be pretty sure any segments they shared were from Antoni

4) if we find other people who match/triangulate (as on GEDmatch, etc) both Antoni matches, and do their genealogy, if we find any who don't have Antoni and Katherine themselves on their tree, we would know that they must have the father on the tree

5) at that point it becomes a matter of what males on their tree were the right age and in the right place for Antoni's conception

6) as a confirmation, or to decide between multiple options, we could trace down to another descendent of the suspected father(s) and get them tested.

I know it's not at all a given, or likely, that we'll be able to find another descendent of Antoni via his first wife, let alone get any other matches after that whose genealogies are fleshed out far enough...but still, is this at least theoretically sound as a method?

My biggest question is about just how much is inherited. One only gets 6.25% (roughly) from a great-great grandparent, and only 3 from a great-great-great-grandparent...so I'm worried that even if both people were in fact descendents, they might just not even show up as matches at that point because they might have inherited entirely different sets of segments from the same ancestors.

Does anyone know the statistics/probabilities on this?

I suppose we could also just go to the original village and test a bunch of people until we find some matches and then trace the genealogy up and hope we don't just find descendents of Antoni or Katherine, but rather of the father...but that seems like more work!

Mac von Frankfurt
08-08-2015, 08:34 PM
I take it you have not identified a living descendant of Antoni Katherinson.

I am not very experienced with autosomal DNA but I don't think the statistics are on your side here.

kjjohnston
08-10-2015, 02:14 PM
So here's my dilemma:

One brick wall in my genealogy (rather far back, but still) is an illegitimate son, Antoni, who took the last name of his mother, Katherine.

However, I'm wondering if at least theoretically it would be possible to identify the father by the following methods:

1) Start with the DNA of my grandmother who is Antoni's great-great-granddaughter via a second wife

2) using "reverse genealogy," find someone who is a descendent of Antoni via his first wife (they seem to have lived in a different town/area before he was widowed and moved be with his second wife, so probably no other ancestors would overlap and confuse things)

3) having these two, we could be pretty sure any segments they shared were from Antoni

4) if we find other people who match/triangulate (as on GEDmatch, etc) both Antoni matches, and do their genealogy, if we find any who don't have Antoni and Katherine themselves on their tree, we would know that they must have the father on the tree

5) at that point it becomes a matter of what males on their tree were the right age and in the right place for Antoni's conception

6) as a confirmation, or to decide between multiple options, we could trace down to another descendent of the suspected father(s) and get them tested.

I know it's not at all a given, or likely, that we'll be able to find another descendent of Antoni via his first wife, let alone get any other matches after that whose genealogies are fleshed out far enough...but still, is this at least theoretically sound as a method?

My biggest question is about just how much is inherited. One only gets 6.25% (roughly) from a great-great grandparent, and only 3 from a great-great-great-grandparent...so I'm worried that even if both people were in fact descendents, they might just not even show up as matches at that point because they might have inherited entirely different sets of segments from the same ancestors.

Does anyone know the statistics/probabilities on this?

I suppose we could also just go to the original village and test a bunch of people until we find some matches and then trace the genealogy up and hope we don't just find descendents of Antoni or Katherine, but rather of the father...but that seems like more work!

4th cousins might be OK >50% but it starts to really fall off after that. Even 3rd cousins will not be a sure bet; >90% will have matching segments.

Family Tree DNA estimates here:
https://www.familytreedna.com/learn/autosomal-ancestry/universal-dna-matching/probability-relative-share-enough-dna-family-finder-detect/

See Blaine Bettinger's shared cM project updated here:
http://www.thegeneticgenealogist.com/2015/05/25/the-shared-cm-project-an-update/

Kathy

Erik
08-10-2015, 02:24 PM
AncestryDNA does something like this automatically. My cousin shows up with 2 people who they say are "New Ancestor Discoveries", based on DNA matches, and their family trees.