DNA link profile to match up people still searching for the paper trail.

+5 votes
207 views
My suggestion is to have a DNA link profile for those whose exact relationship is unknown. I made this suggestion in an answer to another question but thought it should have its own question. We created unknown profiles with family surname to show the relationship. This would be a common problem on my mother's side since my grandparents were from Ukraine. Cousins at the third and fourth level could be linked together without knowing the exact paper trail. It actually happened on my father's side too where I added a possible link as uncertain. I would rather have a way to show DNA links that have not yet resolved a paper trail. The DNA link would be a way to show all those people that are DNA related.
WikiTree profile: Unknown Paduczak
in The Tree House by Sue Hall G2G6 Pilot (146k points)

3 Answers

+7 votes
 
Best answer

WikiTree has a recommended way to show the Y-DNA and mtDNA relationship when the exact genealogical relationship is unknown: https://www.wikitree.com/wiki/Help:DNA_Categorization_for_Projects

by Peter Roberts G2G6 Pilot (559k points)
selected by William Foster
This seems a useful substitute for prohibited YDNA links.  However, it is tied to Ysearch IDs, and Ysearch is going defunct.  Is whoever wrote this going to revise it so it could still be used?

The link can be to a table of matching haplotypes on the Web e.g. http://genealogy.hopetownmuseum.com/bahamasdnaproject/sasserdna.html#Results

I'm also confident other databases will eventually replace MitoSearch and YSearch.

+10 votes

(Before I clicked "Add answer," I re-read this and need to say, in advance, that some of it is a general rant and in no way directed at you, Sue.)

I think the idea would be great in theory, Sue, but I'm not certain it's practical. The whole GDPR mess aside--which, trust me, I very much wish I could toss aside--it's extremely difficult to work backward from DNA only using a "one-tree" like WikiTree. It can and is often done using private trees, or on "mirror" trees on sites like Ancestry, but the problem here--my opinion only--is adding multiple profiles that are only speculative would cause more confusion than it could correct, and that would fly in the face of WikiTree mission statement: "Our mission is to grow an accurate single family tree..."

Neither yDNA nor mtDNA can be used to predict recent-generation relationships, and autosomal DNA without paper-trail corroboration can really only do that at the level of parent/child and possibly full sibling. The average amount of expected shared DNA with a half-sibling, aunt/uncle, or niece/nephew is identical, so DNA as predictive evidence is already beginning to break down at that point...and the variability of the amount of shared DNA becomes more random and more pronounced the more removed is the relationship. This is also all assuming the family tree has no pedigree collapse, which every tree in the world does at some point.

There's seemingly a general assumption that DNA evidence is pretty straightforward to use, that in every case it's binary, either option A or option B. But it isn't. In truth--and I'm not going out on much of a limb here, scientifically speaking--there's a good chance that a casually constructed atDNA triangulation confirming 4th cousins to shared 3g-grandparents that includes a paper-trail is going to be indefensible if closely examined. It's very difficult to be accurate with DNA evidence even in conjunction with solid traditional paper-trail as far out as 4th cousins. Third cousins aren't always the slam-dunk that we accept here on WikiTree, but a line had to be drawn somewhere. To use be able to use DNA evidence accurately out to 5th cousins is, in fact, highly unlikely and I would go so far as to say that vast majority of "confirmed with DNA" settings you see out that far would be erroneous if put under a microscope.

I get slammed--sometimes playfully, sometimes not--for posts that are way too wordy. Guilty as charged. But the fundamental problem is that accurately using DNA in genealogy is not simple. And trying to squish explanations of very complex subjects down to a 100-word flashcard is a recipe for disaster.

Kind of like needing to perform an emergency appendectomy out in the field and telling the doctor on the phone to quit talking so much and just tell you where to cut.

Heck, I know people who self-style themselves as professional genetic genealogists who can't, without cheating, define ROH or recLOH or heteroplasmy, much less explain why they're important to consider in genealogy. The bottom line is, if you want to use atDNA out to 2nd and probably 3rd cousins, you're working with the first-aid kit and are probably good to go. But working past 3rd cousins is performing surgery without the training.

Ahem. Sorry. Big soapbox issue that comes up for me all the time. What we've been seeing the past few years with DNA is not dissimilar to what we saw when the World Wide Web opened up online family trees to the wholesale merging and subsequent propagation of false information. Today, someone posts "DNA proof" of a set of relationships, and the next person comes along and accepts it as fact, no questions asked. But much of the time it is, just like that unexamined online family tree, false.

Soapbox packed away for another day. Cough, cough.

I think the heart of the question is that hypothetical DNA matches can be tremendously valuable and need to be researched and analyzed...exactly what our Adoption Angels do every day. But that's probably best done away from the one-tree itself. Someone pops up as a 15cM match, I can guarantee that's going to take some work to determine if it's genealogically valid, much less how it might accurately fit into a family tree. It isn't enough to simply do a broad brush on DNA Painter and believe the line of segment inheritance is determined because it looks like it fits (or doesn't fit) somewhere among the grandparents. If the goal is accuracy, that is.

Maybe a working FreeSpace page, reference to which can be included in a "=== Research Notes ===" section on the various profiles in that particular relationship chain. That way cousins looking at the profile know what's up, and might even be able to contribute to the collaborative analysis happening on the Space page.

I'll close by apologizing again, Sue. Folks who read this far are the ones truly interested in accuracy in their research. The ones who bailed early just want to know where to cut to get at the stupid appendix.  :-)

by Edison Williams G2G6 Pilot (309k points)

Largest segment = 21.4 cM
Total of segments > 7 cM = 31.8 cM
2 matching segments
Estimated number of generations to MRCA = 4.4

This is the case in point between the cousins. 

Also the Surname matches so there are only 2 generations of unknowns.

Well, and that's sorta the problem. In its current state, that's information, not evidence. And with our standard autosomal DNA tests for genealogy, going any farther than full-siblings without other forms of corroboration, like a researched paper trail, is only assumption. If there are two generations of unknowns in the mix, the DNA is definitely not yet evidence, it's just circumstantial anecdote.

Even without worrying about factors like DNA pile-up areas, both population and haplotype, and considerations of the difference between the female and male genomic map for the segments in questions--what GEDmatch and other services gives us are Kosambi sex-averaged centiMorgan values, and at smaller segment sizes the difference could completely erase a supposed match--I glanced at the relationship finder for you and Paduczak-4.

It shows that you would be the 3g-granddaughter. If we assume the other cousin is a contemporary generation, that would make you 4th cousins. At 31.8cM shared, that means the two of you share 240% more autosomal genetic material than should be the case theoretically. It isn't definite, but if you are believed to be 4th cousins that amount makes a very good case for pedigree collapse in the line...biological relatives reproducing with biological relatives, scrambling the predictive amounts of DNA for relationships.

And this is one place where, I believe, simplistic, formulaic autosomal triangulation breaks down. If the actual sharing is considerably more or less than the theoretical average, accuracy demands that you now aren't merely working to find other cousins who share the same segment. The analysis now needs to include evidence to rule out all other possible channels of inheritance for those segments. Since you're missing any paper-trail at all for half the hypothesized generations between you and Paduczak-4, the line of DNA inheritance may well not run through him at all. The shared surname is information, but does nothing to affect the potential DNA sharing channels.

That's kinda my point. Dealing with autosomal DNA past 2nd or maybe 3rd cousins isn't as complex as performing major surgery, but it isn't nearly as simple as using a first-aid kit to patch up a cut. The amount of data and research required is exponentially greater the deeper you go generationally.

I should have made it very clear that I don't speak for WikiTree or its policies or guidelines. It's my personal opinion only that creating generations of profiles based on unresearched DNA information, not evidence, is probably more confusing to the tree overall than it is helpful.

So we may have too many levels of unknowns? There could be only one? Our earliest known ancestors may be siblings instead of cousins?
+2 votes
I understand what you are trying to do Sue,  You want a way to see the people you match and work back to you.  

That is kind of what we do as Adoption Angels.   Sometimes we have the name of of a person who supposedly matches you at the 4th cousin level with a tree that shows a 4 x Great Grandparents as Joe Blow and Millie Maybe.   So Joe Blow and Millie Maybe had say 4 children Joe Jr, Marie, John, and Jane.  Now that tree traces the rest of the generations down to themself, the match but not so of the other 3 sibling lines.  

So you want to see something like a blank chart you can fill in and look at.  My recommendation is to create a free space page and put your chart there.  

If you use excel you can copy the grid and paste it into  

https://www.wikitree.com/wiki/Space:DNA_Bright_Ideas#Converting_Spread_Sheets_to_WikiTables    then click do it and copy the code onto the free space page.   It will display your chart

The copy the free space page link to your profile page under a header like

===DNA Possibles===
by Laura Bozzay G2G6 Pilot (670k points)

Related questions

+5 votes
2 answers
268 views asked Apr 25, 2018 in Policy and Style by Lance Martin G2G6 Pilot (101k points)
+9 votes
2 answers
+17 votes
8 answers
691 views asked Jul 18, 2020 in The Tree House by Mike Wells G2G6 Mach 4 (46.7k points)
+1 vote
4 answers
337 views asked Sep 28, 2018 in The Tree House by Anonymous Farrar G2G6 Mach 1 (14.9k points)

WikiTree  ~  About  ~  Help Help  ~  Search Person Search  ~  Surname:

disclaimer - terms - copyright

...