How do we make WikiTree more useful for yDNA surname projects and one-name studies?

Question

I'm working on how to expand our DNA features so they're more useful for those participating in one-name studies and surname projects.

 

I was thinking we would do surname index pages, parallel to http://www.wikitree.com/genealogy/SMITH for example, that would list everyone named Smith who is connected through a yDNA test.

 

For example, the Smith list could consolidate the list of names from the Y-chromosome DNA test of Orville Smith, with the list of Phillip Smith, with the list of Hal Smith, etc.

 

Now I'm wondering if all you'd really need is just a list that includes the test-takers, e.g. Orville Smith, Phillip Smith, Hal Smith, etc., not all their paternal ancestors. You could click over to the members' list pages (linked in the previous paragraph) to see details.

 

That would just be a list of a handful of names on the current Smith surname index page and it would be easy to do. (I think we could still allow for the possibility of spelling evolution, e.g. link to Roland Arsenault from the Arceneaux and Arseneau indexes because his yDNA line includes all three, not just Arsenault.)

 

How important would having the whole list be for you, depending on its capabilities?

 

For example, if you have the whole list, we might enable sorting by birth date and first name. This would enable you to see, for example, all the Smiths connected through DNA tests born in the 18th century, or all the ones named Edward, etc. This might lead to some interesting connections. How valuable do you think would be?

 

Almost the same question ...

 

What will make leaders of one-name studies and surname projects want to participate on WikiTree? Is there something here that could really solve a problem for them?

 

Input or suggestions? Thanks!

Comments

You may want to check with the haplogroup project admins. Some of them are fairly knowledgeable and could be invaluable.

---

Hi Paul. Do you know anyone you could introduce?

---

Chris,
I am discussing this with Mike Walsh (L21+ Admin).  His project has a combined 11150+ R1b-L21+(and subclade) members.  However, with the flood of upcoming next gen sequencing results beginning to come in, he is somewhat swamped in data right now.

Using my little project as a for instance it is not only surname based but geographic as well.  Our progenitor "Uidhir" was first mentioned in the annals in 956AD however, I know of no other details regarding his time on the planet.

Fast forward into the 12th century and Donn Mor Mag UIdhir enters to "establish the family" in Co Fermanaugh.  Then in 1264 Donn Carrach Maguire steps up as "the first of his name to be a king of fermanaugh".

The surnames that make up the various septs of the line of descendents involves around 8-10 family off shoots, all in the Loch Erne/Fermanaugh/Cavan/Monaghan region.

We comprise some 167 Y DNA signatures of Mike's 11,000+, mostly at Y67, several at Y111 but many with lower STR count/inconclusive testing.

Having said all that...I do know Mike is interested in combining the Genetic Gen with Std Gen to find the cross over points.  So we're plugging away at it.

 

edit: to add a cautionary note...we are just NOW beginning to attain SNPs of a genealogical value (I think).  Once those are sorted and verified....well....we'll see.

---

After a year ... these are now coming soon. They could be ready for release by Monday!

A more recent related discussion: http://www.wikitree.com/g2g/116779/would-you-like-gedmatch-for-those-who-share-the-same-surname

---

These are now live. For example: http://www.wikitree.com/index.php?title=Special:DNATests&s=Smith

For more, look under "Surnames" in the "My WikiTree" pull-down menu.

If anybody knows a DNA surname project admin please let them know. These should be really useful as recruiting tools at the least.

---

Good tool Chris.  Popped up some new possible matches.  If we test in different vendors though, comparing is best done via third party data sources.  Such as can be found in the L21+ Yahoo Group and or the FTDNA L21+ project.

Once similarity in Hts is matched or compared, then family tree matching and branching; is a positive indicator in success.

Answer 1

I think I would prefer to have a full list of names with date ranges, but have a link directly to each person's family tree.  For my Anonymous DNA  Smiths, the reader could then easily see the test participant's family tree.  I don't know how this would work with Peter's private profiles, however.  

How will this work for multiple tests for the same Earliest Known Ancestor?  I have 18 tests that trace their ancestry back to Lieutenant Samuel Smith.

Comments

Thanks, Kitty!

This will come out sounding wrong, but ... I don't care what you "prefer." :-) What do you really want? What would be really valuable for you as a surname project administrator?

Depending on how we do this, it could be tough to implement. I want to judge what would be worth the time and effort and how it should fit into WikiTree's priorities.

Regarding seeing the test participant's family tree, that's always available with a few clicks. And the links to all connected profiles' yDNA ancestors and descendants is available with one click on the test connections list, e.g. http://www.wikitree.com/index.php?title=Special:DNATests&u=5940206&id=8

That's how we'd do the consolidated test connections list too. With privacy controls, of course.

Regarding multiples of the same ancestor, everybody would only be listed once. For example, this Samuel Smith would only appear on the Smith yDNA index once even though he's connected through eight test-takers. We could highlight on the index that he's connected eight times.

But how useful would any of that be? That's the real question? Is any of this going to make a surname project admin want to use WikiTree?

---

Hello Chris,

My reply relates to autosomal DNA and not as much to Y-DNA.  This is from slide 83 of Dr. Tim Janzen's June 6, 2013 PowerPoint presentation to the Southern California Genealogical Society.

"What is Needed Now:  A large independent worldwide genealogy database controlled by the genetic genealogy community linked to phased autosomal haplotypes."

This is not as ambitious as it may sound.  WikiTree does not have to give control to the genetic genealogy community to make this work.  The genetic genealogy community simply needs to see who the shared ancestors are for persons who share the same autosomal DNA segments.  Phased autosomal haplotypes (e.g. those created by GEDmatch's Phasing Utility) simply need to be linkable to an improved version of WikiTree's Relationship Finder.

Where GEDmatch, Family Tree DNA, and 23andMe are weak is in the genealogy database realm.

---

A few weeks after RootsTech where I saw Dr Janzen speak as well, I'm still thinking that we have a unique oppurtunity to combine phased auDNA data with our shared tree to help "reconstruct" the DNA of our ancestors and help make auDNA data even more useful.

With the amount of distant relatives contacting me via 23andme, I suspect that it should be possible to partially recreate the auDNA of some common ancestors, and that could become very useful in resolving disputes, busting brick walls and adoptees finding biological parents. The key is to combine the auDNA data with the paper trail in a smart way.

So, how do we make this happen?

---

You rock, Roland.

My problem is that I'm overwhelmed by the ongoing responsibilities of WikiTree. It's hard to keep up with bug fixes and feature improvements that we're already committed to doing. I hardly have time to understand the autosomal phasing tools that are out there, let alone work to integrate them. Still, I'm as committed as ever to the importance of this.

Peter, I know you have some outstanding recommendations on integrating with GEDmatch and DNAgedcom tools. I need to examine those.

Thanks for all you're doing, Roland and Peter.

---

Here are my suggestions regarding improving autosomal DNA features in WIkiTree.

Triangulated groups  (where three or more people all match/overlap each other on the same autosomal DNA (atDNA) segment) need see (as easily as possible) ANY shared ancestry they have in WikiTree. That is to say we need to see the ancestral lines for each of them back to their shared ancestors.   I'm guessing that WikiTree should look back at least 20 generations for shared ancestry.

We need to be able to enter two or more WikiTree IDs (or select two or more atDNA tested persons in WikiTree) and have WikiTree automatically display the lines back to (and including) all the most recent common ancestorS found in WikiTree.  This will usually be an ancestral couple or couples, but in cases where an ancestor had more than one spouse or the spouse is unknown/unlisted it would be individual ancestors.    The current WikiTree Relationship Finder only displays the most recent common ancestor between two people, - when in fact the shared atDNA could be due to other shared ancestry.  So we need to be able to see all the ancestral branches which could be shared by persons in a triangulated group.  WikiTree should identify (via a matrix?) who in the triangulated group have the same ancestors.  For example if a triangulated group of 5 includes 3 people WikiTree identifies has having a shared ancestral couple, then this is a strong indication the other two also have this ancestral couple in their ancestry. (Or the other two may descend from an ancestor of of the ancestral couple).

A further improvement would be to able to enter or select the GEDmatch ID of atDNA tested persons WikiTree and automatically see any shared ancestry..

This feature would also be helpful in finding all the recent shared ancestries between JUST TWO people (not only triangulated groups).

Questions:

If GEDmatch or DNAgedcom's ADSA allowed the entry of a url related to the ancestral tree in WIkiTree of each tested person, could WikiTree automatically identify and display the shared ancestry for those tested person?

Ancestry has various levels of uncertainty.  If John Doe suspects he descends from Adam and Eva Bond and John is an atDNA match with Sally Smith who has strong evidence she descends from Adam and Eva Bond.  Then can there be a clear and easy way to indicate in WikiTree that John Doe's ancestral line back to Adam and Eva Bond is supported (confirmed?) via an atDNA match (on chromsome "5" on a shared segment starting at position "72.4" million and ending at position "98.7" million) with Sally Smith?

---

Chris, given your workload, this may be a good candidate wor an apps project. The API can be used to find related people and could easily be extended I presume to also provide the DNA info the users have entered, such as gedmeatch ID.

Peter, I just stareted pokeing a bit at gedmatch and dnagedcom and I'm still learning the details. My first impression is that they both could use some work improving the user interface. I guess the important question is how easily can we interface with them? I don't think they provide APIs for external programs.

Speaking of APIs, 23andme does provide one. It's just not clear yet it allows access to all the usefull stuff such as phased version of the data.

So I guess my question is should a connection be made with an existing DNA site or should an alternative be developed that is taylor made to interface with WikiTree?

---

Hello Roland,

I recommend you, Chris, or someone suggest to GEDmatch and ADSA at DNAgedcom that they provide APIs for external programs.  I believe the key to using atDNA is easily and clearly showing all shared ancestry which can be found for persons belonging to triangulated groups.

Sincerely,

---

Ok, I really need to learn this stuff, so if my questions are strange, please understand I am still rudimentary in understanding.

I have now recieved two emails trying to make connections to me through auDNA.  Both of them are asking about a specfic spot on a specific chromosome.  As if everyone related is related through a specific chromosome.  Is this correct?  Can we ID these ancestors and attach them to a specific chromosome?  This implies that if someone has a match on a chromosome but has a bad tree, they potentially could work out where the problem on the tree exist.  They would be focused on find the connection to the specific ancestor/Chromosome match.

---

Persons who all match each other on the same segment* of a chromosome inherited that segment from a shared ancestor.

*That segment usually has to measure more than about 7 cM if the comparison is using unphased data (which is usually the case).

---

So is that something Wikitree can help facilitate?  Finding a common ancestor and identifying the associated segment on the Chromosome.  I ask because, based my two emails, this seems to be a way for finding the common ancestor.

---

Yes, this is something WikiTree can facilitate.  Sites such as DNAgedcom and / or GEDmatch can determine who share (have overlapping) segments of sufficient size and form Trangulated Groups (TGs).  If those people have WikiTree IDs then WikiTree can automatically show any shared ancestry of that TG.  

We need more genetic genealogists encouraging sites such as DNAgedcom and GEDmatch to collaborate on making this happen.

Answer 2

Chris,

As I've already set up special project pages for my One Name Studies, and probably several for my assorted surname DNA studies, I'd appreciate a widget for those pages that would enable me to list  Patriarchs (aka all of that surname on WikiTree that are end of line), with an indication on the list as to whether or not any on the direct line had a Y-DNA test, preferably also how many of them had, as that would indicate to me whether a confirmed DNA signature may have been derived. Such a list would need a date range to quickly highlight who is modern and needs connecting to an older tree.

Don't forget the women though for ONS, yes they changed their names and their Y-DNA doesn't exist, but they are still name holders.

I don't want to have to spend a lot of time on my special page, but I do want it to be useful to those interested, with links to where all the rest of the data is to be found. And useful to me, which the Patriarch list/dna link above woud make it.

(As an aside, I found only being able to list 3 surname variants for my special ONS/DNA pages a bit of a challenge to pick the most common variants)

In another dna thread somewhere I had a wishlist item that for any of the dna lists of test participants that the Managed by x also be included so you can quickly see who else is interested.

As for the general surname lists that already exist, I'd love to see the DNA symbol next to the test taker, not sure about whether I'd also want it on all those impacted, unless the type of test (Y, mt, at) was indicated somehow, perhaps different colours like the privacy button, tst taker, derived, y, mt, at (grin).

Comments

I second Lorna's suggestion of having a DNA symbol next to the test taker.  I would like to see that symbol / icon on all those impacted.

Currently each person in trees etc, have clickable icons below their name and birth/death info.  I would like to see a DNA symbol there if that direct line is tested.  Have variations of symbols for Y and mtDNA.  Those symbols should link to a DNA page for that person or persons direct line tested.

As a project admin of several (Y-DNA) surname projects and a major geographical DNA project using mtDNA, aDNA, Y-DNA, and soon X-DNA.  I need to know who took the test (be allowed the option to link to those results - e.g. in YSearch, MitoSearch), I need to clearly see (without digging/clicking down) the direct line impacted by those results.  With one click I need to be able to easily view the descendant chart (ancestral line with Y line descendants, or ancestral line with mtDNA descendants - depending on test) showing both tested and untested lines impacted by that DNA test.

Very important but not something I've found discussed here is the ability to associate one line with another line on the account of the DNA results.  The fact that John "Doe" in  New Zealand is a 35/37 match with Joseph "Doe" in Cornwall, England is highly significant.  No paper conncection has been found (yet) or may ever be found.  But WikiTree needs to allow for a clear visual association for those two DOE lines which distinguishes them from the DOE lines which are not a Y chromosome match.  I suggest having a DNA link symbol for  a Y-DNA indicated connection -- this line and this other line likely share a direct paternal line ancestor in a genealogical timeframe.    (a genealogical timeframe is the time since surnames generally became adopted and vital records have been recorded).

Sincerely, Peter

---

Making a DNA project page show up on the surname list page (eg http://www.wikitree.com/genealogy/wight) and including any surname related special pages on the genealogy page for the surname

(eg http://www.wikitree.com/index.php?title=Special:Surname&s=WIGHT&order=dobup) would be useful.

For some reason, probably my ineptitude, my newly created WIGHT DNA project page doesn't show up on the former, (although a runciman one I did a while ago does) and I've not figured out how to make something show up on the latter.

The latter (http://www.wikitree.com/index.php?title=Special:Surname&s=WIGHT&order=dobup) page would be where it would be good to see the dna symbol next to test takers, and end of line patriarchs as you can see dates AND places at a glance.

Answer 3

Having the ability to see which surname groups in a specific location are related to other surname groups in different locations and perhaps more importantly which are not related, would be of great help when it comes to pointing us in the right direction of which records we need to be looking at.

 

But hold on, maybe I’ve been doing family history too long!

Remember the saying “It’s a wise man that knows there own father”

 

Are we not being a bit presumptuous to assume everyone on a male researched surname line has the same Y-DNA

 

I know within the Cobbett’s alive today many decent from a male of the name Sherwood, (2nd marriage was Cobbett and all the children’s names were changed).

Not to mention the hundreds of Cobbett’s in the USA that now call themselves Corbett, just because of one ancestor that it appears gave up saying to people “No my name is Cobbett not Corbett”.

 

So to be able to differentiate between proven DNA results, where the results match the family tree, confirmed by two sets of DNA results having a common ancestor on the family tree, and the long lines of ancestors where we just presume they are the fathers.

 

Also instead of just linking to a surname maybe we need the flexibility to identify different lines of the same DNA results, 

 

Peterson

 

Comments

I agree that we can't assume everyone on a male researched surname line will be a Y chromosome match.  That is why it is important for WikiTree to make/provide a clear visual differentation between confirmed lines and unconfirmed lines.

As a surname DNA project admin I'm aware of three different types of paternal lines:

1) Y-DNA tested and confirmed.

2) Y-DNA tested but not confirmed.

3) Untested

Here is a ficticious example of the three types:  Adam Smith is b. 1700 in Jamestown, VA.  Genealogical records show he had two sons and each son had sons each generation to the present day ending in Alexander Smith and Andrew Smith who are "8th" cousins to each other.  Alexander and Andrew each had a Y-DNA37 test and were found to be a 35/37 match with each other.  This confirms each of their direct paternal lines back to their most recent direct paternal line ancestor Adam Smith.

Genealogy records also show Adam Smith had a brother Benjamin Smith b. 1705 Jamestown.  He had at least one son every generation to the present day ending in Bobby Smith who has NOT had a Y-DNA test.  Bobby belongs to a Y-DNA tested paternal line but it has not been confirmed that Bobby's Y-DNA would match his distant cousins Alexander and Andrew.

There is also another living Smith who traces his direct paternal line back to Zachariah Smith b. 1707 Jamestown.  This Smith line is untested.

---

Hi Peter,

Where do tests with surprise results go?  I refer to my "twins" born 1806 that have proven to come from two different genetic DNA lines.  I assume one of the "twins" was adopted.  I think your list is ok, but for example #2.  I use the following designations when I feel a need.

1 yDNA tested and confirmed common forefather:  Two or more matching tests from different branches that share a common yDNA forefather (triangulated tests.)

2 yDNA tested and unconfirmed:  Multiple tests from the same branch, unexpected test results, or tested lines with no matches yet.

3 Untested branches.

So in your example, Bobby Smith's branch would be untested.  We have this situation with the John Bland Smith family.  Some branches are tested and confirmed, and some are untested.

Personally, I don't really like the confirmed-unconfimed designation because I think it has a cooling effect on further or deeper testing, but I know I may be in the minority on that opinion.  

---

Hello Kitty,

As most know, surprise results are not uncommon.  Where genealogical records are supported by Y-DNA or mtDNA results then you have a confirmed direct line.  Non Paternity Events (NPEs) - - adoptions, etc. should, when possible, be linked to their biological parents.  There should be suggested wording (good examples of wording) for their biographical notes which explains who the family is who raised them.

Regarding "2 yDNA tested and unconfirmed:  Multiple tests from the same branch, unexpected test results, or tested lines with no matches yet."   I agree.  I believe that it is important for people to clearly see that the line is unconfirmed. 

Sincerely, Peter

---

Answer 4

I am the administrator a well-established Surname DNA project (viz. www.dnastudy.clanirwin.org ) and a WikiTree rookie.  I see considerable potential for better interfaces between Wikitree members and DNA surname projects.  For example our DNA project has over 300 participants, few of whom will be WikiTree members, while I suspect few of WikiTree's 4000 members with relevant surnames will be aware of how yDNA tests are now able to help them: what was thought to be a single-origin surname has turned out to have about 30 unrelated branches! 

I suggest that (1) each Wikitree surname list has a box into which anyone can enter the web address of the relevant surname study, so that those interested can have their DNA-related questions answered by the relevant project administrator; & (2), as Lorna has suggested, each Wikitree surname member list has a column for a marker indicating if each member has undertaken a y-, mt-, or at-DNA test, so others may contact them. 

But to avoid confidentiality problems and unnecessary duplication I recommend WikiTree takes one step at a time.  When I get some Wikitree recruits I will be happy to recommend WikiTree to other surname DNA project administrators.