RootsChat.Com
General => Ancestral Family Tree DNA Testing => Topic started by: Steve3180 on Saturday 10 May 25 11:47 BST (UK)
-
Does anybody know of any Books, Academic Papers, Websites or anything else that has a methodology for discovering the identity of unknown ancestors (specifically in my case Great-Grandfather)
One of my two main goals when I first did the dna test was to discover the identity of my Grandmothers father which is blank on her birth cert. I was hoping for a bunch of 2nd cousins to pop up and point directly, but no such luck.
Over the years I have collected about 100 matches who I think are related to me through this unknown GGF and built several trees connecting some of them together but I currently feel that although I've got a lot of data I'm still stumbling around trying to get lucky somewhere.
What I need now is a methodology I can apply to this data and testing techniques to somehow validate what is currently just an intuitive feel.
For example if I take two dna matches and find a common ancestor for them that may be on the same line to me or it may not, but what are the odds of each and how are those odds altered if I find three or four or more going to the same common ancestor. Intuitively I feel as though three or four matches connecting to the same ancestor would be on the path to me but I would like some maths to back this up.
Anybody ?
-
This has been covered many times.
Read this thread:-
https://www.rootschat.com/forum/index.php?topic=863488.0
Specifically #28 on Grouping.
Once complete those that are left are the Unknowns and in these may be the link that you seek.
To give an example, in one of the trees for a person whose DNA I manage their Great Great Grandfather is unknown. In the UNKNOWN Group there are about 50 DNA matches in the 20 to 100cM range. The usual applies, no trees, no useful information on their name etc, shared matches for each of them are investigated.
Of the likely candidates that remain a Tree is built around them and if a person is found who could have been in the right location then they are flagged as Potential.
I worked through all the other likely ones building trees.
Going back to the Potentials I am still looking for people with a closer relationship whobcan be approached to take a DNA test.
Another approach is to upload the DNA data to other websites to see if there are closer matches, again the above thread details the process.
Good luck.
-
Biggles, I must have missed that back in January. It's a procedure that I'm familiar with, but you did detail it very nicely.
Two comments, firstly it's not long and arduous. I would recommend that anyone read through that 'comment 28' several times to fully understand it, before applying it
Secondly, people should be aware that one flaw in this system would occur if any of the eight great grandparents were closely related, as quite often might be possible in a small close-knit urban or rural Victorian community. Descendants of such related great-grandparents would then have matches in more than one group.
I really do encourage anyone interested in analysing their DNA results to read the above mentioned posting 28.
Zaph
-
There isn't a method that works for everyone as there are so many variables, grouping and tree building are the foundations but accuracy of records and quality of matches, together with their willingness to collaborate and/or learn about dna and genealogy will always cause difficulties. A lot of my time is spent revisiting my groups, updating them with new matches and expanding branches to the present day looking for target testers. Some matches are more important than others but without the results it's impossible to say who is more significant than another.
-
Alas there is not a “one size fits all” solution.
If the Groups is inconclusive then it could be simply a waiting game.
I have been searching for that GG GP for 15 years, 8 of which has included DNA and whilst I have a possible, there is no paperbtrail to substantiate the hypothesis so it remains a possible.
Yes, Pedigree Collapse can affect results and squew the DNA test cM’s.
The next stage would be to look at DNA Tools to drill down into Segments and Triangulation, alas not possible currently on Ancestry.
Reading Graham Holton’s book will give you an insight into the technicalities, but be warned it is heavy going but it is the defacto standard, Blaine T Bettinger’s book is easier going.
-
Biggles, thanks for the pointer, I've gone through that post and looked at it all including #28. That's pretty much what I've been doing. It's relatively easy to identify dna matches in the Unknown group as geographically they're pretty distinct. A lot of them are via shared matches and don't have trees but for those that do I build trees if I can, I now have 5, maybe 6, trees that are promising and have between 3 and 10 dna matches in them connecting to their common ancestors.
My original idea was to build these trees forming a canopy, if you like, over the target ancestor and join up the trees with the joins focusing down to the target. This hasn't worked at all ! In six years of doing this I've only managed to join two of the trees together. Even the two trees that have common cross matches on the dna are stubbornly refusing to connect.
So now I have too much data, the tree I'm currently looking at which is the conjoined one would probably yield over fifty Potentials. The same attribute that makes them easy to detect also makes a large number of Potentials. They are all within a 12 miles square box on Tyneside and Wearside, which in the year in question, 1892, would probably as easy to travel around as it is now (trains everywhere). Also being the male side the age range is very broad, I'm using 20-60 but if could be outside that even. None of the Unknown group have very close matches (I have very few close matches in general), the closest being 56 cM and all the rest less than 40 cM, so maybe 3rd cousins but mostly fourth and over.
I have pretty much given up trying to join the various trees and so am looking for other ways to focus the data, possibly statistical. Is the book you mention "Tracing Your Ancestors Using DNA" ?
-
Steve - I'm not a dna expert, but I learned enough about it to use dna analysis to SOLVE the mystery of my great-great-grandfather. Not sure if my method can be of any help to you, but I'll give you the short version of my long process. My great-great-grandmother, Caroline Reid, gave birth to my great-grandfather, John Reid, in a poorhouse in Scotland. He was "illegitimate" and no father named or mentioned anywhere. I became pretty obsessed with discovering who had fathered my great-grandfather, who ended up emigrating from Scotland to New Jersey in the 1880's. This is my paternal line, so I had my father do a Y-DNA for me. If he hadn't been alive at the time, I could have gone to one of my brothers or a cousin to obtain that paternal dna line. When the results came in, there were 13 or so people who matched to my Dad whose surname was McLaughlin. When I contacted them and examined their trees, I discovered all these McLaughlins were from Derry in Northern Ireland. So I figured I'm probably looking for someone named McLaughlin from Northern Ireland! Took me 2.5 years of searching, but finally I identified a man named McLaughlin who appeared on several trees of people who were dna matches to Dad. Turned out, as I found more records, it was this man's brother who was my great-great- grandfather, Dominick McLaughlin. The final confirmation was a census record which recorded him as living 2 streets away from where my gr-gr-grandmother lived. Dad was matching to his brother's descendants because Dominick didn't have any surviving children himself. That is, apart from my gr-grandfather who he likely never knew existed. My gr-gr-grandmother was probably working the streets when she got pregnant for a second time, as she already had a toddler, both out of wedlock. I even eventually found the gr-gr-granddaughter of her elder son living in England, and using dna analysis, we discovered that the two sons had to have been from different fathers. So, dna CAN lead you to the right person, but it takes determination, luck, and time.
-
Steve - I'm not a dna expert, but I learned enough about it to use dna analysis to SOLVE the mystery of my great-great-grandfather. Not sure if my method can be of any help to you, but I'll give you the short version of my long process. My great-great-grandmother, Caroline Reid, gave birth to my great-grandfather, John Reid, in a poorhouse in Scotland. He was "illegitimate" and no father named or mentioned anywhere. I became pretty obsessed with discovering who had fathered my great-grandfather, who ended up emigrating from Scotland to New Jersey in the 1880's. This is my paternal line, so I had my father do a Y-DNA for me. If he hadn't been alive at the time, I could have gone to one of my brothers or a cousin to obtain that paternal dna line. When the results came in, there were 13 or so people who matched to my Dad whose surname was McLaughlin. When I contacted them and examined their trees, I discovered all these McLaughlins were from Derry in Northern Ireland. So I figured I'm probably looking for someone named McLaughlin from Northern Ireland! Took me 2.5 years of searching, but finally I identified a man named McLaughlin who appeared on several trees of people who were dna matches to Dad. Turned out, as I found more records, it was this man's brother who was my great-great- grandfather, Dominick McLaughlin. The final confirmation was a census record which recorded him as living 2 streets away from where my gr-gr-grandmother lived. Dad was matching to his brother's descendants because Dominick didn't have any surviving children himself. That is, apart from my gr-grandfather who he likely never knew existed. My gr-gr-grandmother was probably working the streets when she got pregnant for a second time, as she already had a toddler, both out of wedlock. I even eventually found the gr-gr-granddaughter of her elder son living in England, and using dna analysis, we discovered that the two sons had to have been from different fathers. So, dna CAN lead you to the right person, but it takes determination, luck, and time.
Yes, y-DNA can help in solving male line mysteries.
In my own line it has helped me to identify my own Great Great Grandfather, whose child was “adopted” by another family. There will never be any documented proof but this hypothesis is the only one that fits in not just with the y-DNA but with other highish atDNA matches.
Alas in the Ops case the unknown is his Grandmother’s Father so y-DNA will not directly help, unless the man also fathered male children with the Ops Great Grandmother and their male line can lead to a person who will take both y and atDNA tests. A long shot but possibly worth looking into.
-
As you say the y-dna won't help.
There is no evidence that my GGM had any other children after my GM, she was 38 at that time. The earlier children were all from her first husband. A few years later she married her second husband and was living with him and his child (not hers) with the earlier children in orphanages or living with siblings. All of this was investigated at depth several times years ago and again in the light of dna. The first husband is not the father (or grandfather which was an examined if slight possibility), and neither is the second husband. None of the suspected dna matches (the Unknown group) have close enough matches to be able to assume any surnames. The Unknown group, where they have trees, have ancestors in the area outlined above in the late 19th century and beyond, so almost certainly a local rather than a seaman.
So what I am left with is a man between the ages of 20 and 60, with an unknown name, maybe one of the Potentials' names, but maybe not as they may be a couple of generations away, who was in or within traveling distance of Sunderland in mid 1892.
Ideally they had another child whose descendant gets a dna test, they would then be my H2C with about a 120 cM match. There is currently no-one anywhere near that, who is an Unknown, on Ancestry, MyHeritage, Gedmatch or FamilyTreeDNA.
That leaves me with a lot of Potentials who are probably at least one or two generations removed from my GGF so I need to reduce the number down to less than ten, say, and examine them closely.
I am currently looking at BanyanDNA as a way of testing the likelihood of connections but it will be a lot of work as it is really does the inverse of what I need. It tests given Hypotheses and what I need is it to tell me the most likely Hypothesis.
Hence the original question, I'm looking for another angle of attack.
-
I feel your pain!! Despite understanding all the variables involved, I too feel there could be a robust methodology out there that systematically allows the user to either identify unknown DNA Matches or conversely to identify unknown ancestors in your tree.
I completely go along with everything else in this post, especially the Groupings discussion detailed in that classsic RC posting. And yes, multiple relationships/pedigree collapse is a real issue that should never be ignored.
What is frustrating is that for me for the 15 or so DNA accounts I manage, most unexpected GGP identifications have just fallen into my lap fairly quickly based on Groupings analysis and a bit of triangulation and commonsense. However, like the OPer, one unknown GGP for one account and for another account one Match of 136 cM remain stubbornly unresolved after 5 years.
All I can add here is to try the WATO tool. It can help guide analysis, despite some extremely bizarre suggestions, but what never fails to be useful is the schematic tree of shared DNA Matches it presents after entering all your Match data. It shows your inputted data as a pared down tree that ONLY shows your DNA Matches and the connections between them which I find extremely useful indeed.
-
This web article deals with small DNA clusters and small DNA matches in the UK (originally posted by another Rootschatter on another thread) -
https://dna-sci.com/2023/05/04/working-with-junior-clusters-to-solve-unknown-parentage-in-the-uk/
-
Be sure to get the latest edition of Bettinger's book. The first edition, which I've read many times, is already superseded a by later edition.
Zaph
Alas there is not a “one size fits all” solution.
Reading Graham Holton’s book will give you an insight into the technicalities, but be warned it is heavy going but it is the defacto standard, Blaine T Bettinger’s book is easier going.
-
Thank you one and all for your input.
I had a play with the WATO tool and I love the way it generates hypotheses but the results it gave me are wild. The first tree I tried it complained that there was nobody over 40 cM, I Wish !, in total on the maternal side I only have 11 dna matches over 40 cM, 8 of which are placed and only 1 in the unknown group. So I tried the tree with the Unknown at 56 cM, unfortunately sparser than the other, and as its most likely hypothesis it suggested 6 generations between 1874 and 1955, with lesser results even worse. I know a lot of people use this so I need to look at how best to present my data to it so I get sensible results.
The suggestion by Carmella was interesting also and possibly the most applicable as it was UK focused and understood that a lot fewer people test here. The chap who wrote it has a book out which I've just bought on Kindle but it's 500 pages so that might take a while to read.
I have a copy of "Genetic Genealogy in Practice" by Blaine Bettinger, which was useful originally but doesn't offer any help in this regard.
There's a new edition of the Graham Holton book out which I'm planning to get when they produce a Kindle version.
A few people mentioned drilling down to find candidates to test. Am I to understand you just find people who haven't tested but would likely be useful and write them a letter or something ? Does this actually get results ? I can't even get people who have already tested to reply !
I think in reality I'm just going to end up back at the waiting game as Biggles said. I've been at this five years now since the dna test, and fifteen before that and I thought I had enough, but it's not looking good.
-
I notice that DNA Painter states in its WATO Q & A that "Strictly speaking, there is no minimum (for cM), but the statistics underlying WATO do not go below 40 cM, so the majority of your DNA matches should share at least 40 cM. The more they share, the better WATO works."
Not sure whether that means that such low cMs should not be entered in the first place as it may "distort" the hypotheses as experienced by you and me or whether they are simply redundant and are effectively ignored.
In my own frustrating case study I have 12 Shared Matches in my WATO but only 4 are over 40 cM and 2 of them are less than 50 cM. The highest Match is only 136 cM.
-
I notice that DNA Painter states in its WATO Q & A that "Strictly speaking, there is no minimum (for cM), but the statistics underlying WATO do not go below 40 cM, so the majority of your DNA matches should share at least 40 cM. The more they share, the better WATO works."
Not sure whether that means that such low cMs should not be entered in the first place as it may "distort" the hypotheses as experienced by you and me or whether they are simply redundant and are effectively ignored.
In my own frustrating case study I have 12 Shared Matches in my WATO but only 4 are over 40 cM and 2 of them are less than 50 cM. The highest Match is only 136 cM.
Is that not enough to get a good idea? From when I've used it, it seems to give quite good indications with such inputs.
I've also found it's best to just not put sub-40cM matches in - owing to them being to ambiguous to be useful.
-
You know WATO is struggling when the highest hypothesis is weighted at 687,173. Indeed, there are 52 hypotheses in total with no clear winner(s). it is odd as the highest score suggested a Match 3 generations below, and it even strongly suggests that the Match could be 5 generations below. And this is for somebody born in the 1970s!!!!
I will take out the sub 40 cM Matches and try it again.