RootsChat.Com
General => Ancestral Family Tree DNA Testing => Topic started by: diplodicus on Sunday 20 May 18 17:58 BST (UK)
-
Having far too much time on my hands (although my wife would argue strongly to the contrary) I have just updated my spreadsheet where I cross-match all my 4th to 6th cousin DNA shared matches. This is where I cross-refer a match where Tom matches Dick. Harry matches Dick so Tom must match Harry even if Ancestry's matching application doesn't say so.
I have 196 4th to 6th cousin matches (or better) and only 29 of them don't share a match with someone else in my matches... so far. I expect to find more of them as shared matches with my stronger 5th to 8th cousins.
Ancestry only reports shared matches with 4th to 6th cousins. I won't record any 5th to 8th cousin matches in my spreadsheet that aren't shared. My spreadsheet is already a 250 x 250 matrix!!!
It has been a fascinating but very time-consuming exercise. However, it has enabled me to find a new shared match and be able to identify it as a "Smith" or a "Jones" or to say sometimes that you are from my maternal or paternal line (matches shared with a cousin on one side or other).
There are truffles but you have to hunt for them!
-
A flaw in your reasoning.
"This is where I cross-refer a match where Tom matches Dick. Harry matches Dick so Tom must match Harry even if Ancestry's matching application doesn't say so."
Smith and Jones marry and have children. Both you and Dick descend from this coup!e, Tom and Harry are from a previous generation.
Tom Smith matches you and Dick, Harry Jones matches you and Dick, Tom Smith and Harry Jones are completely unrelated.
Or another way of looking at it -
Tom, Dick and you match on chromosome 6. Harry, Dick and you match on chromosome 10.
Dick and you match them both on 6 and 10. Tom and Harry are not a match with each other.
Sometimes you will be right, because they may only be 5 to 8 cousins to each other, thus not show up as shared, but that won't always be the case.
Regards Margaret
-
Thanks Margaret, I always knew that my statistical reasoning was flawed but was unable to express myself as well as you have.
I think/hope that grouping people in this way provides a "working hypothesis" that can be used as a starting point in the search for a recent common progenitor. I know that in most cases there will be little chance of finding such a person let alone constructing anything other than wholly circumstantial evidence to confirm the connection.
Yes, despite the uncertainties, it has proved useful in a few conversations. Taking your example, and bearing in mind how much less mobility there was for my Ag Lab forefathers, Tom Smith and Harry Jones might not be related but they could well have come from the same part of the world if not the same village. It's better than nothing.
The next stage is to invite Tom, Dick and Harry to upload their results to Gedmatch and/or FTDNA. Then, the connection can be confirmed or denied. Given the general lack of responsiveness to even tentative enquiries, it is something I intend to approach with some caution and not a great deal of optimism.
Si vos can non a iocus, non essemus socii tibi.
Malcolm