Can you help draft instructions on confirming relationships with autosomal and xDNA?

+20 votes
760 views
DNA Project Members and other WikiTreers,
 
We're planning to add the ability to mark parent-child relationships on WikiTree as "Confirmed with DNA". [Old G2G discussion]
 
There would probably be these three status options for each mother and father:
o Uncertain
o Genealogically certain
o Confirmed with DNA
 
This will enable us to show branches of a tree that are confirmed and still need to be confirmed. I consider it a very big thing for the future of WikiTree and our mission.
 
When a member marks a relationship as genealogically certain, of course, they're expected to cite sources.
 
When a member marks a relationship as DNA-confirmed, they should cite their sources as well.
 
We have a system worked out for citing yDNA and mtDNA conclusions: http://www.wikitree.com/wiki/DNA_Confirmations
It's not especially user-friendly but some advanced members are doing it and we can build on it.
 
We don't have anything parallel to this for autosomal tests. We need to start drafting instructions for how to draw a conclusion that a relationship has been confirmed through autosomal DNA, and how to show that conclusion on a WikiTree profile.
 
I am not qualified to write this. At the same time, I don't think this should be written by and for advanced genetic genealogists only. They need to be instructions that any genealogist who has taken a 23andMe, AncestryDNA, or FTDNA Family Finder could follow. It's over 12,000 WikiTreers already.
 
The instructions will need to lay out different confirmation methods, with different levels of certainty. The highest level would be something that third parties could verify. But using a third party tool like GEDMatch.com, DNAGedcom, etc., shouldn't be required for every conclusion.
 
For example, my father and I have both taken 23andMe tests. I think it's clear from this that DNA has confirmed he is my father. If I want to mark my father as "Confirmed with DNA" what should I put on my profile?
 
Maybe I should be instructed to put something like this:
 
== Sources ==
 
* DNA confirmation of paternal relationship based on 23andMe tests taken by Chris and [[Whitten-2|Ted]] showing "47.6% DNA shared across 22 segments" according to 23andMe.
 
My second cousin has also taken a 23andMe test. What relationships could I confirm through her and how do I cite 23andMe as a source on the appropriate profiles?
 
I think we need separate instructions for all three testing companies when there are triangulated groups reported through them.
 
I know a lot of people will worry about the number of mistakes that will be made. I'm not worried about this. I think mistakes are good. It's all part of a process. We need to get that process started!
 
Can you help? Could you draft instructions for using the website of one of the testing companies to draw conclusions? They don't have to be perfect. Just a rough start.
 
Thanks!
 
Chris
asked Feb 27, 2015 in Policy and Style by Chris Whitten G2G6 Pilot (977,560 points)
Hello Chris

I know I'm late to the party, but am not well at the moment.  I am however interested in helping to draft instructions if you still need people.  I'm certaing not an advanced genetic genealogist but am keen on this topic and can certainly see value in the great suggestions so far by Bryant and others.  Haven't had the mental strength to look at the issue too critically but can see this is progressing nicely.  Like Peters idea of an alternate DNA tree in the longer term.  The trick is going to be making it simple for people to understand when they can confirm or not.  Triangulation is the key, but I agree with Ellen there will be some relationships that could be confirmed without it, but where do you draw the line?

As I mentioned on an earlier discussion on this topic if we do go to the extent of documenting the segment matches as Lenny and others suggest (and I tend to agree that this is needed) we will need a standard input template (by chromosome) otherwise it will be difficult to ensure the right data is being analysed and documented.  Gedmatch ids are clearly essential for checking but ideally I think there should be a summary note somewhere on the Wikitree dna page to identify the basis for the confirmation.  It also needs to have tracking attached as to who entered the confirmation info.  You don't want it like a bio which doesn't really identify who made specific changes.  Over time as more and more data gets added it could become a bit messy and hard to follow.  Peter has shown the info in the template form which looks good if there was only one template per profile.  I suppose I'm thinking that the info needs to populate some type of ongoing table perhaps.  Some active testers may build up lots of info and templates if each identified relationship has a template.  Hope this makes sense.

Veronica
I will throw up an idea in case it helps somone else see a way forward.  

If you share autosomal DNA segments with a known 3rd cousin then (because it is not triangulated) it partly confirms each of your ancestral lines back to your shared great-great-grandparents.  If another person also matches BOTH OF YOU on one of those segments and they also descend from that ancestral couple then that is confirmation of that segment coming from that ancestral couple.  It is triangulation.

We do not (may not) need to keep adding additional DNA confirmation templates confirming ancestral lines which have already been confirmed.  For example confirming back to your paternal great-great grandparent already confirms your relationship with your father.
Chris,

I'm happy to help and triangulation is key--atDNA can be used to confirm kinships with triangulation as at least one other person on here has laid it out.  Most folks do not understand this at this point.  HOWEVER, I'm a HUGE advocate of education on this because the more folks who understand how to use the data the more it helps me personally! LOL  If I contact you and you understand then you will know how to reply to help us cement how we connect.  Someone said they use ancestry (no tools-best utilized by uploading data to gedmatch to validate/verify) and FTDNA - also no way to confirm triangulations at present, although we project admins are pushing for it hard! -but they didn't have any experience with 23andMe.  I have done presentations on using 23andMe and am happy to contribute.

This is my original presentation--it has been added to in order to provide more advanced info for beginners on triangulation, and is in need of a cleanup, but it's  out there for anyone wanting to use it for 23andMe specifically.

http://trackingyourroots.com/DNA/Tracking%20Your%20Roots%20With%20DNA.pdf

I'll help in any way I can.

Lisa
 
Veronica, it would be great to have your help developing the standards and instructions on http://www.wikitree.com/wiki/Confirmed_with_DNA
 
Maybe Peter, Lisa, Veronica, me and others who want to work on this page should start an e-mail conversation. We can do some of this on G2G but it's not perfect for coordination. Though, I do think we should answer this first:
 

 

7 Answers

+7 votes
I think it important to note where a DNA match has been made, where it has been confirmed through triangulation, and where it has been further verified through multiple triangulation.

 

To do this, we would need to be able to input the individuals tested whose trees converged and what specific piece of DNA it matched on.

Imagine something like the ADSA report on DNAGEDCOM
answered Feb 27, 2015 by Lenny Darnell G2G Crew (960 points)
Thanks, Lenny. Once we get this page started I hope you'll help work on the instructions.
+6 votes

DNA is something near and dear to my heart, and I'd really like to try and help with this.  That said, I'm not sure how much time I have right now for it...

You mention "...need separate instructions for all three testing companies when there are triangulated groups reported through them."  I haven't tested with 23andMe, only through ancestry and FTDNA.  But FTDNA doesn't report triangulated groups, and ancestry is only in beta testing on this (they call it  'DNA circles').

For certainty on matches, you either need a high number of cM that you share or else confirmation through triangulation, which really needs to occur through gedmatch or DNAGedcom.  From what I've heard, there are too many 'false positives' on ancestry.  Another problem is that you may share more than one family connection with a person with whom you share DNA, so without confirming the chromosome and DNA segment (and actually having a triangulated group), you can't be sure which of your ancestors that DNA pertains to.  Another issue can arise when one of the people doesn't have a confirmed line back to an ancestor, i.e. they show a person in their lineage but down the road it will show up that they connected to a wrong family.  I've seen this happen, but I was only able to confirm it because I had a match with three other people on a specific chromosome/segment, and this other person was matching me on a different chromosome & segment.  Anyway, food for thought.

If I'm not able to write these, I'd be happy to read them over and offer suggestions.

 

answered Feb 27, 2015 by Darlene Athey-Hill G2G6 Pilot (219,680 points)
+5 votes

I like the idea being able to mark parent-child relationships as 'Confrimed with DNA': however, this should be reserved for test results of two individuals where the individuals are half identical on all autosomes. This can be observed in the comparison of autosomal test results in 23andMe, FTDNA, and GEDmatch. Two tested siblings can also be 'Confirmed with DNA' through direct comparison of test results; 50% segments shared with about 25% of genome being completely identical.

I think other relationships should be marked with 'DNA support' and this may be applied in several different ways. I have recently been working with different ways to communicate this in wikiTree. I have been summarizing simple matching results from AncestryDNA, and here are some categories I think are relevant.

DNA Ancestor(s)

For example, this DNA test page has a list of most recent common ancestors supported by DNA matches to a test result in Ancestry. This is the profile of one of the common ancestors, summarizing the information in the DNA Support section of the page. It would be great to have a template for this. Additional information on triangulated segments could also be added in cases where it is available; or that a common ancestor forms a hub of a 'DNA Circle' on AncestryDNA.

DNA Siblings and/or DNA Connection

These would be cases of siblings of a common ancestor(s) where DNA tests of their respective descendants match. For example, Thomas and Patrick are supported as brothers due to DNA matches between each of their respective descendants. DNA Connections are all the intermediate indvidiuals between these brothers and the tested individuals. One son of Patrick is both a DNA Ancestor and a DNA Connection.

In addition to being able to easily summarize supporting DNA results on a profile, I would like to be able to see each DNA supported lineage represented when looking at the family tree. This could be done by changing the appearance of individuals and of the lines connecting them.

I'm willing to help with these features.

answered Feb 27, 2015 by Bryant McAllister G2G1 (1,810 points)
edited Feb 28, 2015 by Bryant McAllister
Hello Bryant,

It is great to have someone with your background on board.

One of my original wishes remains being able to see the DNA supported/confirmed ancestral lines in one's family tree.  That is a "DNA ancestral view"  and a "DNA descendants view" with confirmed lines and/or persons highlighted/bolded or with small symbolic icons.  That will allow everyone to see what has not been confirmed.

My primary wish is that it will be relatively easy for others to be able to independantly verify another person's claim of DNA confirmation.

Sincerely, Peter
Bryant, I'm so glad you're willing to help here. It looks like you're miles ahead of many WikiTreers on this.

The main goal for adding a "confirmed with DNA" status indicator for mothers and fathers is to enable displaying it on trees and descendants lists, like you and Peter are talking about.

I don't think we'll add more than three status indicators. Possibily these three:

o uncertain

o confident

o confirmed with DNA

Exactly what each one means will be argued about for years to come. :-)

Many people are suggesting more levels, e.g. "supported with documentation," "confirmed by an expert," etc. But adding more levels adds more complication.

The "uncertain" and the "confirmed with DNA" indicators are the ones we want, so that we can visually represent this on tree views.
Hello Chris,

I feel VERY strongly that there is vital need to indicate a few levels of certainty for "Confirmed with DNA".   When talking about more then immediate relatives there is a whole lot of uncertainty if people don't share GEDmatch IDs or if more than one ancestral couple is shared by a match.  Triangulation and phasing also greatly increase the level of certainty.

I would not like to see people selecting confirmed with DNA when there is little evidence provided for confirmation.

Perhaps the citation/template can clearly indicate the level of certainty.  But if so, the "radio button" status indicator should not say "Confirmed with DNA" but something like "DNA Evidence Level" or "Level of DNA Support" or "DNA Support (Level)"

WikiTreers should be able to select both confident (from traditional genealogical resourses) and "DNA Evidence Level."

Most sincerely, Peter
Hi Peter,
 
Check out http://www.wikitree.com/wiki/Relationship_Status -- it's my very early start of documentation to describe the database indicator and how it should be used. I started this a half hour ago so it's very rough.
 
The database indicators do not need to represent the full range of genealogical certainty and DNA confirmation. The decision on the number of indicators is based on how we plan to use them, e.g. in tree views.
 
In general, we have very little structured data on WikiTree. For example, we don't have fields for burial date and location, christening, etc. That doesn't mean that this information isn't significant. It just means we didn't think tracking it in separate fields was the right decision.
 
Chris
Chris and Peter,

I agree with Peter, and would have difficulty myself clicking on a radio button indicating "Confirmed with DNA" even if I had lots of support for the relationship. Although I can be extremely confident that I share an ancestor(s) with a set of individuals based on DNA matching, the association of this connection with a couple or an individual is an inference that is prone to error or to bias beyond the matching. I think a radio button with "DNA Support" or "DNA Evidence" or "Supported with DNA" would more generally represent how DNA informs the genealogy.

Some of the posts in this thread also indicate a dependency of only representing DNA matches between individuals each having profiles on wikiTree. This would greatly restrict the utility of a template. I would rather see a template be more widely applicable while still respecting the privacy of DNA matches. This could be accomplished by referencing the matches' first generation ancestor of the inferred common ancestor(s).

Bryant
Thanks Bryant,

When "confirmation" (using Y haplotypes and mtDNA HVR1&2 differences) was being hashed out, we tried to be clear that conformation was not proof, and that it was important to compare results in a publicly accessable database (or table) such as YSearch, Mitosearch (or url in the note field linking to a comparison table) to be able to independently verify a confirmation.  The purpose was to use matching Y or mtDNA to confirm that the ancestry as shown in WikiTree is accuate.

With auDNA I believe we can call it confirmation only when testers have GEDmatch IDs or DNAgedcom IDs or link to publicly accessible source explaining the shared segments, and it is a triangulated group (with segments above 7cM) with a known shared ancestor in WikiTree, or we are working with phased data (or two males who match each other on an X chromosome segment).  If it is just two people who are 3?rd cousins or greater who are a sufficient auDNA match and they have a known ancestry in WikiTree - that I would want to describe as "partial confirmation"

DNA Support or DNA Evidence might better describe examples where independent verification not easily possible.

Having said that - as a side note - I can see that auDNA could be used to "confirm" a parent/child relationship even if no traditional paper genealogical evidence existed.

Most sincerely, Peter
+4 votes

My suggestions follow. 

The confirmation process using autosomal DNA (auDNA) and the X chromosome (X-DNA) can for now use a DNA confirmation template on profiles similar to what is currently being done with the Y chromosome and mitochondrial DNA. 

It is important to point out a lower level of certainty for autosomal DNA confirmation when just two persons share a segment (unphased) and a higher level of certainty when three or more persons all match each other (overlap) on the same segment thus forming a triangulated group.  I suggest the accuracy of the relationship back to the common ancestor be designated “Partly Confirmed” when just two people share a segment 7 cM or greater,   I suggest the accuracy of the relationship back to the common ancestor be designated “Confirmed” when three or more people all match each other on a segment 7 cM or greater (i.e. they form a triangulated group (TG)).

I don’t have experience with using AncestryDNA Circles.  I assume using AncestryDNA matching alone (without being able to verify the match using GEDmatch) places an AncestryDNA match in the Partly Confirmed category. 

The lower level of certainty would show:

WikiTree ID of person1’s relationship to WikiTree ID of common ancestor (and spouse) is partly confirmed because the auDNA tests of WikiTree ID of person1 and WikiTree ID of person2 share a Z cM segment on chromosome Q from start position A to stop position B.  See person1 and person2’s ancestry back to their most recent common ancestor.

For example the Autosomal DNA Partly Confirmed template would read: 

Roberts-7085’s relationship to Pinder-48 (or spouse) is partly confirmed because the auDNA tests of Roberts-7085 and Kitchen-311 share a 8.1 cM segment on chromosome 6 from 166029345 to 168443695.  See Roberts-7085 and Kitchen-311’s ancestry back to their most recent common ancestor Pinder-48.

I believe the input template for this would be something like:

{{auDNA Partial Confirmation

| person1 = Roberts-7085

| person2 = Kitchen-311

| common = Pinder-48

| testURL1 = http://www.wikitree.com/index.php?title=Special:DNATests&u=6035205&id=6

| testURL2 = http://www.wikitree.com/index.php?title=Special:DNATests&u=6250473&id=6

| SegmentSize = 8.1

| Chromo# = 6

| StartPosition = 166029345

| StopPosition = 168443695

|comments =  }}

This template would go on the profile of person1 and person2 and each of their ancestors back to but not including the common ancestor. 

answered Feb 28, 2015 by Peter Roberts G2G6 Pilot (344,820 points)
edited Feb 28, 2015 by Peter Roberts

The higher level of certainty would show:

WikiTree ID of person1’s relationship to WikiTree ID of common ancestor (and spouse) is confirmed because the auDNA tests of WikiTree ID of person1 and WikiTree ID of person2 and the WikiTree ID of person3, etc. all match each other on a Z cM segment on chromosome Q from start position A to stop position B. 

[note: It is not currently possible for WikiTree to link to a url showing the shared ancestry of more than two people]

For example the Autosomal DNA Confirmed template would read:  Roberts-7085’s relationship to Pinder-56 (or spouse) is confirmed because the auDNA tests of Roberts-7085 and Sjostrom-39 and Collins-5366 all match each other on a 10.8 cM segment on chromosome 1 from 163621974 to 173712569.

I believe the input template for this would be something like:

{{auDNA Confirmation

| person1 = Roberts-7085

| person2 = Sjostrom-39

| person3 = Collins-5366

| common = Pinder-56

| testURL1 = http://www.wikitree.com/index.php?title=Special:DNATests&u=6035205&id=6

| testURL2 = http://www.wikitree.com/index.php?title=Special:DNATests&u=6250473&id=6

| testURL2 = http://www.wikitree.com/index.php?title=Special:DNATests&u=6250473&id=6

| SegmentSize = 10.8

| Chromo# = 1

| StartPosition = 163621974

| StopPosition = 173712569

|comments =  }}

This template would go on the profile of person1 and person2 and each of their ancestors back to but not including the common ancestor.

Each auDNA or X-DNA tester must have a WikiTree ID.  Those in the comparison all need to be in the same system (e.g. all in Family Finder or all in GEDmatch or all in 23andMe, etc).

The link to the person's auDNA test would show the testing lab and if the results are in GEDmatch (because their DNA Tests page would show their GEDmatch ID (or DNAgedcom ID when Wikitree has an input field for it)).  As of  Feb 27, 2015 DNAgedcom only accepts Family Tree DNA results for its ADSA tool. 

Results should be in GEDmatch and GEDmatch IDs should be on the tester's DNA Tests page in WikiTree so others can independently verify the auDNA or X-DNA confirmation.

It will sometimes be difficult to know the exact cM in a triangulated group. But you will know the start and stop positions.  

Family Finder’s In Common With (ICW) feature is not the same as sharing the same segment.   ICW should not be used alone for more than partial confirmation.

There can be a problem when a person is in a triangulated group and that person has more than one ancestral path back to the shared ancestral couple of the TG.  WikiTree may mistakenly point out a relationship path other than what is known to exist for the triangulated group.

Later for X-DNA confirmation.

After rereading Chris' original question I clearly see I did not provide instructions or various levels of certainty for confirming ancestry.  That is to say there are many DNA test takers who will not have the ability to or want to use a DNA confirmed template as I described. Yet their DNA test does still provide some basic information useful for their genealogy.  More later.

Sincerely, Peter

I have tried using this suggested template, which gives me:-

Template:AuDNA Partial Confirmation

Only problem is when I click on it in 'Public Profile' it gives me the following message:-

The action you have requested is limited to users in the group Supervisors.

 

This is what my Template looks like:-

 {{auDNA Partial Confirmation

| person1 = Wright-15908
| person2 = Irvine-911
| common = Irvine-900

| testURL1 = http://www.wikitree.com/index.php?title=Special:DNATests&u=10552632
| testURL2 = http://www.wikitree.com/index.php?title=Special:DNATests&u=10628479

| SegmentSize = 23.29
| Chromo# = 4
| StartPosition = 23317655
| StopPosition = 40820851

| SegmentSize = 18.21
| Chromo# = 9
| StartPosition = 98621737
| StopPosition = 111991826

| SegmentSize = 26.9
| Chromo# = 12
| StartPosition = 61880
| StopPosition = 14011756

| SegmentSize =12.82
| Chromo# = 15
| StartPosition = 52274201
| StopPosition = 60817690

| SegmentSize = 21.5
| Chromo# = 16
| StartPosition = 780770
| StopPosition = 10311911

| SegmentSize = 9.94
| Chromo# = 22
| StartPosition = 19221259
| StopPosition = 26148766

|comments =  }}

 

What do the Group Supervisors see?

BTW, I am comparing with a known 2nd cousin once removed.

Regards

Dennis Wright (Wright-15908)

Hello Dennis,

I do not understand templates well enough.  My examples are not functional and I think something needs to happen behind the scenes.   I can imagine they would take up a lot of space on pofiles as currently designed.
+4 votes

Some thoughts...

Autosomal DNA can prove close relationships, but when looking beyond the first few degrees of relatedness, "confirmation" can be elusive. We are often going to be dealing with probabilities, not solid "proof." I say this because the average fraction of our DNA that we share becomes very small after the first few degrees of relatedness and because of the random way that DNA segments get transmitted.

There's a table at http://www.isogg.org/wiki/Autosomal_DNA_statistics that shows theoretical average percentages of total autosomal DNA shared at various degrees of relationship. The actual percentage shares of identical-by-descent (IBD) DNA segments are distributed around those averages. Second cousins may share only 1% of their DNA. Even at relationships as close as about 3rd cousin once removed, the lower end of the reported range is about 0.1%, where it gets difficult to distinguish IBD DNA from the DNA that two people might share by coincidence.

With such small amounts of DNA shared, the goal of "triangulation" (i.e. finding at least three people with the same common ancestor who have DNA matches on the same segment) requires a good bit of luck (what's the chance that three people who each received 1% of their DNA from their common great-great grandparents all received the same bit?) -- although chances of matching are much higher when X-chromosome inheritance is involved. (Also, false matches are a real possibility.)

In the absence of triangulation, shared DNA segments can be an informative clue. I have a known second cousin with whom I share 3.38% of my autosomal DNA, and I believe that provides strong evidence that our reported family history is correct. That second cousin and I also have an identified 4th cousin with whom we each share two DNA segments, totaling 0.4% to 0.7% of our autosomal DNA. It happens that one of those shared segments is a 3-way match on the X chromosome, so we have triangulated evidence of our common descent from our shared great-great-great grandparents. However, in the absence of that 3-way match, the fact that we both have strong matches with this 4th cousin would have confirmed the relationship for me.

Because of experience like this, I'd like WikiTree to document DNA matches found by a pair of people on WikiTree who trace their ancestry to the same person(s), regardless of whether those two people have found a 3-way match with a third person.

answered Mar 1, 2015 by Ellen Smith G2G6 Pilot (613,790 points)
edited Mar 1, 2015 by Ellen Smith
I have also seen a couple of 3-way DNA matches at 23andMe that may or may not actually confirm a common ancestor. These are matches with people with whom I share documented ancestry at somewhere between 7 and 11 generations, on ancestral lines related to the Puritan Great Migration. I interpret these matches as indicating a high probability that we come from the same gene pool, but I don't believe that these 3-way matches could truly confirm that a particular DNA segment came from any particular ancestor or married pair -- particularly when you consider that there were several generations of "inbreeding" (due to marriages between limited sets of families) in the PGM population. If I had Ashkenazi ancestry, I'd be particularly skeptical of DNA matches, as that population is identified as having many generations of marriage within a relatively small group.
Thank you, Ellen. I'm glad you're participating.

We'll work on getting the status indicators enabled in the database, and start a very rough page to coordinate instructions in different scenarios.
+3 votes
I'm not sure  to what extent this is relevant but here's how the mistress of citation recommends citing DNA test results:

https://www.evidenceexplained.com/content/citing-autosomal-dna-test-results-online-database
answered Mar 1, 2015 by Jillaine Smith G2G6 Pilot (549,040 points)
edited Mar 2, 2015 by Peter Roberts
Thanks, Jillaine!

My reading of Evidence Explained would be

"Family Finder," database, FamilyTreeDNA (http://www.FamilyTreeDNA.com : accessed 17 July 2012), using the "Matches" and "Chromosome Browser" options to query for [NAME OF INDIVIDUAL]; results from this dynamic database require the private passcode and kit number of this individual.

- where accessed is used instead of downloaded.  WikiTree encourages the tester to post their kit number when known.

To make it easier for others to verify a match between two persons who are not immediate realtives, - it is important to also identify on which chromosme the shared segment is found and the start and stop position of that segment.

Sincerely, Peter

+1 vote

I know I am a little late to this party also…

The first step in is usually to define the need which may not always be obvious.  In other words, how does someone benefit from its inclusion?

There are two types of users who are interested in this information. 

(1)    Those that have trees use the DNA results as evidence, just as a birth or death certificate does.  We form an opinion as to the degree of confidence in those connections based on the current evidence at hand.  Maybe we have something like a DNA Report; we create, and cite the DNA Report as the DNA Evidence.

(2)    Those that do not have trees, and use the DNA results to provide direction in finding a way to connect their part of the tree to others on Wikitree.

The prior presumes they have a Wikitree connection to other DNA Testers, while the latter presumes there is no connection in Wikitree… Yet!

You do not need triangulation to have confidence in a connection. You use triangulation as a measure of confidence.  For example, 23andme, FTDNA, and AncestryDNA will predict a relationship, and a range.  Gedmatch will predict a genetic distance. If these predict that a parent/child relationship is a parent/child relationship that should be sufficient. If two people have a paper trail showing they are 4th cousins and there is a prediction they are in the range of 3rd to distant cousins, this is supporting evidence of the relationship.  The confidence level may increase as more DNA Testers get tested. 

The real value of triangulation is to prioritize and focus a DNA Tester’s attention to finding a common ancestor with another DNA Tester.

The first step is to identify those who are related by blood (DNA Matches), using the results of the different DNA services. The next step is to prioritize by ordering the list by genetic distance with the intent of locating their tree. Let’s presume this is outside the scope of Wikitree.

Let’s presume now the Adoptee has access to one of those DNA Testers profile.  The Adoptee now compares his list with a Wikitree list of those DNA Testers identified as blood relatives who could possibly share any DNA for Wikitree DNA Tester identified.  Let presume that Wikitree identifies those Testers. In other words, You know you want to concertate on those you know have a tree and you begin with the closest matches first.

Then the adoptee compares his DNA with those who are listed on Wikitree as blood relatives. This comparison is done outside the scope of Wikitree.  The Adoptee is motivated to extend the Wikitree by adding to the decedents of the common ancestors in the hopes one of the paths will lead to him.  He is trying to complete the triangles that include him.  In other words, since he knows he is related to two people, a good starting point is to focus on their common ancestors.

Understanding Segment Triangulation…

A segment is identified by a starting and ending location on a Chromosome.  The distance between them are measured in cM's. Let’s say that 3 adoptees (A, B, C) state they have segments of at least 7 cM’s that overlap among the three of them… There are 3 segments A/B, B/C and A/C.

The overlap of these three segments infers a 4th segment. The Triangulated Segment (TS).  The location of this segment begins at the highest starting position and the lowest ending position of the 3 segments. In other words, the Triangulated Segment can be from 1cM to 7cM in length. 

A Triangulated Segment cannot be used to IDENTIFY a PARTICULAR Common Ancestor. 

For example, three first cousins to each other share 2 common grandparents.  The Triangulated segment can Identify the cousin parents they inherited this segment from, but there is no way to determine which grandparent passed this Segment onto their children.

Another example, aTriangulated Segment indicates Adoptee A, B, and C share “A” Common Ancestor but Adoptee A may share a different Most Recent Common Ancestor (MRCA) with Adoptee B, than with Adoptee C.

answered May 28, 2015 by Ken Sargent G2G6 Mach 4 (46,360 points)
edited May 28, 2015 by Ken Sargent
Hello Ken,

In your view, can a triangulated segment be used to identify a particular ancestral couple?

Thanks and sincerely, Peter

Hi Peter,

The simple answer is No. A triangulated segment cannot be used to identify ancestral couples. Let me explain how I would make use of triangulated segments. I am not sure how good a job Iof explainig I did.

 First, DNA Matching is a two-step process. 

(1)    Identify DNA Relatives in which a relationship can be predicted to a certain degree of accuracy. This excludes those we know are DNA Relatives but cannot predict the relationship to any degree of certainty. The prior group has segments 7cM or greater(IBD), the later group has less than 7cM(IBS).

On 23andme, DNA Relatives uses IBD and Countries of Ancestry also includes (IBS).  

(2)    Determine the genetic distance for each match which requires that both IBD and IBS segments are used in the calculation.

Let me explain how I would precede.

Let’s say that you and I share 1 Segment (5cM in length).  This 5cM Segment indicates we share a common ancestor but we cannot determine the genetic Distance of this Ancestor (IBD).

I then see on Wiktree that you and I are 15th cousins.  It’s important that we have a KNOWN Common Ancestor because this allows us to replace step 1 and step 2 with this information. We begin with the premis our trees are correct.

This narrows down the possible common ancestor to the two people identified in Wikitree as our Most Recent Common Ancestors. We really don’t at this point care which one is associated with the 5cm Segment. We do  care about the Wikitree relationship tree and how we are related.

Next, using the DNA Services, we find matches that have  least one 7cM segment  with a Segment overlaps our 5cM Segment. This can be the same 7cm Segment or any other segment. 

For each match, we will use the genetic distance predictions of the DNA Services to isolate a location or locations on our tree to focus on. 

For example: Let us say that one of these matches is predicted to be a 2nd Cousin. You are approximately the same age, so you probably share 2 of your 8 great grandparents but Wikitree narrows this down to a single specific great grandparent. This would be the great grandparent that connects you to me on the 5cM Segment.  The confidence level of this connection is directly related to the confidence level of the tree and of the DNA Genetic Distance prediction. 

Let’s go one step further.  Since we know that the 2nd Cousin of you is not a close relative to me, we can presume that the 2nd Cousins grandparent is a sibling of your Grandparent who connects to me.

Adding yDNA, mtDNA, and checking for matches/no matches on the X chromosome will help to include and exclude possibilities and support or refute the proposed tree connection. The more of these scenarios that can be performed, the more confident and accurate the proposed tree connections will become.

I can’t think of another way to make use of IBS segments and I believe this approach works in all cases.  The result will not always narrow down to a single person, but it should significantly narrow down the possibilities. Given enough examples, we may be able to narrow down to single person.

Best Wishes

Ken

Thank you. I will continue to believe that a shared segment for a triangulated group is from a shared ancestor of that group.  Often it is only possible to narrow it down to an ancestral couple.

Related questions

+10 votes
2 answers
159 views asked Sep 25, 2015 in The Tree House by Peter Roberts G2G6 Pilot (344,820 points)
+9 votes
1 answer
108 views asked Jan 19, 2016 in Genealogy Help by Michael Stills G2G6 Pilot (327,680 points)
+21 votes
4 answers
338 views asked May 1, 2016 in The Tree House by Vic Watt G2G6 Pilot (281,200 points)
+42 votes
4 answers
+37 votes
5 answers
+14 votes
4 answers
425 views asked Jun 23, 2016 in The Tree House by Peter Roberts G2G6 Pilot (344,820 points)

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

disclaimer - terms - copyright

...