* Recording Ancestry ThruLines with unknown individuals

[jelv]

I'd appreciate some thoughts on an issue I have:
.

Ancestry ThruLines.png (39.69 KiB) Viewed 2020 times

All the red obscured names are individuals I already have in my project.

I'm recording matches using a small modification of Jane's method with this note template on the DNAMatch fact:

Code: Select all

[[
Tested with:
Name on Test Site:
GedMatch Details
Kit Id:
GedMatch Name:

Links
Shared Ancestor:
Shared Couple:
Primary Link: 
]]

[[
Match details

]]

Normally I'd make the effort to identify the names of the Private individuals, but that isn't always possible.

What do people do in this situation - create dummy individuals (three in the example above)? It would mean that if a query or report displayed the relationship to the Primary Link it would be correct.

Any other suggestions?

[RS3100]

I do nothing in that situation. If you can't identify the two private individuals, how are you going to be able to confirm the identity of "OT" and that the line of descent suggested by the Thrulines algorithm is actually correct?

[jelv]

What I should have added is that looking at the non-private parts of their tree they have the Great-grandaunt and 1st cousin 2x removed. Doesn't that mean that those two are validated as DNA related to me?

I'm using two flags to put icons on diagrams, one flag and icon for people that are DNA matches, the other flag and icon is put on all the people that make the connection. This so I can see which ancestors I've validated by DNA connections. (I don't put the flag on the parents of the siblings that make the connection as it is possible they could be siblings of some other parent(s)). I would put the second flag on the Great-grandaunt and 1st cousin 2x removed but not the unknowns.

[AdrianBruce]

...
What do people do in this situation - create dummy individuals (three in the example above)? It would mean that if a query or report displayed the relationship to the Primary Link it would be correct. ...

I can't claim any experience in that precise case but what I would say is that a lot of my suggested Thrulines are rubbish. It therefore becomes, in my own view, a bit of a moot point to say that "if a query or report displayed the relationship to the Primary Link it would be correct".

Correct in what sense? Genealogically correct? You have no idea whether that's true or not.

Correct according to OT's research? Well, given the way that Ancestry's Thrulines logic patches links together, the links that Ancestry shows, may not match the links that OT has. (I might be making too much of this since a Dummy Record won't have much on it to be wrong but Ancestry has a woman as 2C1R and a man as 3C. What if OT has that order reversed but Ancestry has more data for its version so has ignored OT?)

The only way that Dummies would be "correct" is that they are correctly describing Ancestry's Thrulines suggestions. But who knows, the actual connection might be to another sibling of the 1897 birth (say), so that not even he's a right person in the link. It might even be via a sibling of your 1838 born GG-GF so that the real common ancestor is your 3G-GF. This is something that I've found - the Thrulines suggested come into real people - but the wrong real people, on the wrong generation.

Personally I think I'd record OT as a person in my FH data but omit the two Private people above OT. The note template can be completed with the Shared Ancestor being named in the text but as "John Smith (suggested)" rather than "John Smith". (Assuming that this doesn't mess something else up).

[AdrianBruce]

What I should have added is that looking at the non-private parts of their tree they have the Great-grandaunt and 1st cousin 2x removed. Doesn't that mean that those two are validated as DNA related to me? ...

No, it just means that OT has those two people in their tree. It doesn't mean that those two people have anything to do with OT - worst case scenario is that X weeks ago, Ancestry Thrulines suggested that to them and they swallowed it hook, line and sinker, even though it was wrong. Remember Thrulines has "nothing" to do with DNA - it's just sticking bits of tree together - it just happens to be using DNA to prioritise which bits to stick together.

I have a rule of thumb - I trust people to know their own grandparents, but beyond that... And in this case, the 1C2R is OT's G-GF - one step beyond.

Look, there's every chance that the Thrulines suggestion is correct. There's also every chance it's wrong...

[jelv]

Adrian, I posted again while you were composing your reply. I am aware of the errors that others may make and seek to validate the connections suggested by other research (GRO, parish and census records). Is my conclusion that there is somebody who connects via the names in both our trees which validates the names I do know as being connected correct (even if the unknowns may not be correct).

(Are we heading in to Donald Rumsfeld territory?)

[RS3100]

Unless you can check the trees thoroughly via you own research, to confirm the assertions displayed in them, how can you be certain that they are correct? I have found more incorrect trees on Ancestry and elsewhere when researching DNA matches than good ones. Many of the problem trees seem to have been simply copied en masse from other users trees without any checks on the veracity of sources, or any sources at all quite often.

I am generally selective in recording DNA matches anyway. I don't investigate or record every match that crops up, seeking mainly to record just those that corroborate (or disprove) my paper based and other research.

If you don't know the identity of OT, what is the point of recording them. What does it prove?

Can you find any instances of your known 1C2R born 1897 in other public trees on Ancestry and elsewhere, which might contain information to identify his children and bring that branch forward. Do you have a marriage for him? If you know his wife's maiden name you can search on FMP or the GRO (in 5 year batches for the latter) for any births registered with the same male surname and the mother's maiden name.

I see that Adrian has posted whilst I have been typing, and I agree with everything he has said.

[RS3100]

Adrian, I posted again while you were composing your reply. I am aware of the errors that others may make and seek to validate the connections suggested by other research (GRO, parish and census records). Is my conclusion that there is somebody who connects via the names in both our trees which validates the names I do know as being connected correct (even if the unknowns may not be correct).

No. As Adrian has said, that could be the case. But it could equally be that the Thrulines algorithm has cobbled that link together from software assumptions made by piecing together partial information from several other trees that may not even be connected to your known relatives.

My Heritage have a similar system, but they do at least express probabilities of the correctness or otherwise of the assertions regarding the matching of individuals in the connecting lines, as percentages.

Unless you have to rely on OT to prove that a specific direct line of your tree is correct, by corroborating your other research, I would ask what benefit making an unproven assumption actually gives you, and where it takes you in compiling your tree?

[jelv]

I have a rule of thumb - I trust people to know their own grandparents, but beyond that... And in this case, the 1C2R is OT's G-GF - one step beyond.

I can relate to that. Yesterday I was puzzling over another match for some time. It should have been fairly easy as it suggested a 2nd or 3rd cousin. First I found that his family couldn't decide how their name was spelt with three variations in GRO records in the last 100 years which didn't hep secondly he had the wrong first name for his great-grandmother who was the connection. Eventually worked out I already had his father in my tree from searching using the techniques RS3100 described.

[jelv]

I am generally selective in recording DNA matches anyway. I don't investigate or record every match that crops up, seeking mainly to record just those that corroborate (or disprove) my paper based and other research.

That is my main motive as well, proving the ancestors. If as a result I find more descendants of the ancestors that is a bonus.

If you don't know the identity of OT, what is the point of recording them. What does it prove?

My thinking was that it proved Great grandmother and Great grandaunt were siblings and that 1C2R was her daughter - but nothing further than that.

[David2416]

I use the principle to start from what you know; in in this case JE is a DNA match to OT.

I record this information using a custom Fact (Attribute) DNA Match (the normal time frame is None) with three witness roles Match, Linker and MRCA (Most Recent Common Ancestor).

The text I would use is "OT purported relationship 3rd Cousin 1R 26cM over two segments" (use your data)
I create the individual OT using whatever Information I can glean from Ancestry.
OT will have the role match, the 2nd great grandfather is MRCA, and the others are Linkers.

I use a screenshot of the Thrulines which I attach either to the fact or preferably to a Source. In the Source I could record any investigations into the unknowns I have carried out along with any assessments etc.

[RS3100]

The text I would use is "OT purported relationship 3rd Cousin 1R 26cM over two segments" (use your data)
I create the individual OT using whatever Information I can glean from Ancestry.
OT will have the role match, the 2nd great grandfather is MRCA, and the others are Linkers.

I use a screenshot of the Thrulines which I attach either to the fact or preferably to a Source. In the Source I could record any investigations into the unknowns I have carried out along with any assessments etc.

But Thrulines are simply hints assembled by an algorithm trawling all Ancestry family trees and attempting to create links between DNA matches. The links between the matches are not proven unless and until they can be demonstrated by credible sources. They can have been patched together by Thrulines from several trees, none of which necessarily contain either of the matches themselves.

So without being able to verify the suggestion that Thrulines is presenting, how can you assert that the 2nd GGF is the MRCA between the two individuals?

For suggested matches that I cannot prove, having investigated them, I add relevant information to the note attached to the match on Ancestry, and flag them so I know that I have investigated them, but I see little point in adding them to my tree if they are unsubstantiated, particularly if I don't know who the individual - in this case OT - actually is. It just clutters up my tree with information that doesn't really add anything to it.

I suppose you could maintain a research note or named list of unproven DNA matches, but again unless they are critical to proving a line of your tree that is otherwise not fully evidenced, I can't see the point.

There are other considerations to be aware of with unproven matches. I have investigated two recently who according to Thrulines would evidence a link to one set of my Gx6 grandparents. That information would be useful to me, because their son, my Gx5 GF who is proven by DNA, was one of three men having the same name and born in the same parish at around the same time. Although I am fairly happy that I have the correct individual, a provable DNA link would be irrefutable. Due to the amount of pedigree collapse in rural communities though, I have found that both matches could be linked to me in at least three different ways, and hence linked to the other individuals also, which gets me no further forward. Something I am now very aware of when investigating a Thruline suggestion, and easily missed if the research into intervening generations between the proposed matches is not widened to encompass their wider families.

[jelv]

I create the individual OT using whatever Information I can glean from Ancestry.

Would you also create the intermediate 2C1R and 3C individuals to keep the relationships correct?

In view of the previous comments I'm thinking of modifying this:

I'm using two flags to put icons on diagrams, one flag and icon for people that are DNA matches, the other flag and icon is put on all the people that make the connection. This so I can see which ancestors I've validated by DNA connections. (I don't put the flag on the parents of the siblings that make the connection as it is possible they could be siblings of some other parent(s)). I would put the second flag on the Great-grandaunt and 1st cousin 2x removed but not the unknowns.

Perhaps I need a third flag with a different icon to be used in cases such as this to indicate a tentative DNA connection and put that on the the Great-grandaunt and 1C2R. (in this case the Great Grandmother already has the validated DNA connection because of another DNA match.

I think David is right, the ThruLines match is too important to ignore and not record and notes of the investigations need to be recorded. Who knows what other information may come to light in the future (maybe another DNA match in the same area of the tree) which combined with the information I have gathered now proves the link (in whole or part).

[jelv]

So without being able to verify the suggestion that Thrulines is presenting, how can you assert that the 2nd GGF is the MRCA between the two individuals?

I agree with that. Even if I knew who OT and the other two linking individuals were it doesn't prove the 2xGGF is right. It only proves GGM and GGA are siblings.

For suggested matches that I cannot prove, having investigated them, I add relevant information to the note attached to the match on Ancestry, and flag them so I know that I have investigated them, but I see little point in adding them to my tree if they are unsubstantiated, particularly if I don't know who the individual - in this case OT - actually is. It just clutters up my tree with information that doesn't really add anything to it.

I suppose you could maintain a research note or named list of unproven DNA matches, but again unless they are critical to proving a line of your tree that is otherwise not fully evidenced, I can't see the point.

I'm totally adverse to that. The reason I use FH is that I want all the information I collect in one place and not to be dependant on someone like Ancestry who might change things so I lose it. The advantage of creating an individual in FH is that it's storing the information in the relevant place where I might stumble across it in years to come when I've totally forgotten my recent investigations. Scattering research notes around the internet in different places strikes me as a sure fire way of information being forgotten/lost.

Thinking about this, I'm going to add a note to all Ancestry DNA matches "FH id nnnn" to point to my primary information store.

I use the FH - RootsMagic - Ancestry sync to push changes to my tree to Ancestry. In this case, however I decide to record it (1 or 3 individuals), I will be setting the Private flag on those records so they don't get transferred.

[ColeValleyGirl]

Scattering research notes around the internet in different places strikes me as a sure fire way of information being forgotten/lost.

Why not store it in Research Notes within FH7? And perhaps group them in a Named List called something appropriate to make it easier to find them?

[jelv]

No, because if I am looking at 1C2R at some time in the future when I've forgotten all this, what is there to prompt me to look at the research note for the three possible descendants?

[ColeValleyGirl]

You could link the Research Note to the descendants? It will then show up on their Notes tab in the Property Box.

[David2416]

I create the individual OT using whatever Information I can glean from Ancestry.

Would you also create the intermediate 2C1R and 3C individuals to keep the relationships correct?

No, I would not also create the intermediates, I would as I have no information about them other than they are potential links. The relationship is purported - and Thrulines is showing a possible linkage. The only fact here is the existence of a DNA match. That match might be explained as shown in Thrulines or it may not. Further information is required - chromosome mapping, cluster analysis and so forth.

What I am doing is recording one possible linkage that may or may not stand the test of further information.

[AdrianBruce]

... My thinking was that it proved Great grandmother and Great grandaunt were siblings and that 1C2R was her daughter - but nothing further than that.

I'd agree with all the other cautions. And I'd repeat that your DNA match to OT does not provide DNA evidence that "Great grandmother and Great grandaunt were siblings and that 1C2R was her daughter" - we have no certainty that Great grandaunt and 1C2R have anything to do with OT. Therefore OT's DNA match to you does not provide DNA evidence for Great grandaunt and 1C2R as really being who you think they are. We haven't after all, tested their DNA.

Even if you could see OT's tree at each step, you'd need to validate each (relevant) step of their tree to see if their conclusions were correct. I suspect they are correct but I've seen too many incorrect trees copied too many times. OT may have simply copied an incorrect tree.

Using DNA to back up your paper-based conclusion that your GGM and your (apparent) GGA really are sisters is actually very difficult, not least because there's a huge element of chicken-and-egg here (how does OT know that the 1869 really is their GG-GM? Not thru DNA but through paper....)