[DNA Confirmation] Ancestry ThruLines

+26 votes
4.0k views

I'm not sure how new this is, but AncestryDNA has added something called ThruLines. This seems to change the triangulation game (and Ancestry Circles) completely.

Here's an example of how it looks...

For the DNA Project, does this change how we handle triangulation at all? Or the DNA Confirmation guidelines?

in The Tree House by G. Borrero G2G6 Pilot (124k points)
I like it, but like any Beta, still needs some bugs worked out.  One of my thrulines, actually had a line through the stepfather of one of my ancestors.

Another one, ignored the ancestry I had for person X, which was backed up my research and DNA, and substituted a different ancestry based on a faulty line that I had since disproven.

So, yes very potentially useful, but like anything, be skeptical and if you can, confirm.
OK, OK! I get it that this is potentially exciting and new, but nobody really expects a new policy on triangulation to be adopted within 48 hours of some brand new thing becoming available, do they?

Honestly, I'd be more excited and optimistic, except my experience with the software on Ancestry it often behaves badly, and the past behavior with DNA Circles has often been enigmatic. It seems like changes on AncestryDNA are sometimes actually useful, but are more often just part of some new sales gimmick. I guess I'll cross my fingers and actually go read about this thing!

Thanks for posting this, BTW!

I just found a sentence within Ancestry's own documentation of this feature that appears to answer the question of what this means, as far as WikiTree DNA Confirmation:

"ThruLines does not validate your relationships with DNA matches."

In other words, this is 100% meaningless, in terms of DNA Confirmation here. Maybe it could help you identify matches that might be useful for DNA Confirmation, but that's about it.

If someone expects ThruLines to write the confirmation statement, do the triangulation for you, well - that's not going to happen.  What ThruLines *does* do though is set the triangulations up for you, and you can then make contact with the matches, identify the actual segments, and write your own confirmation segment.  For me, it popped up a completely unknown set of about 20 matches, all converging on a single and previously unknown ancestor.  That's incredible!

The method will be to grab 2 or more screen shots of the lineages from MRCA to matches, then add info about the identified segments.  As Frank noted below, the lineages even include the 'Private' individuals in those paths, quite amazing.  They are marked as private and aren't identified, but are just as valid as identified people (in my opinion).  That's not something that was possible before!
Rob, I think what I'm seeing, as far as the private trees go, is that if I have any of their ancestors on MY tree, a person who would show up as "private" if you just looked at THEIR tree shows up as having their name filled in.

In some cases, I have people in my tree that go right down to the parents of the other match (so that private tree might as well not be private, at least private from me), or at least there's only one or two "private" rectangles left in the presented lineage. Generally, I just want to know where they fall on my tree - I don't necessarily care about names, when you get down to distant living cousins.

I added some people to my tree try to clarify about my one "private" match, but it didn't have any effect. Probably, it needs time - maybe even days.

Anyway, it's really more like a souped-up version of "Shared Ancestor Hints", than like DNA Circles, if you ask me. Like you're saying- raw material for attempting DNA Confirmation, but certainly not the whole ball game.
Not sure exactly how it works and it is in beta. I have a situation where someone I am working with used their married name in their tree instead of maiden name (because they don't have a known biological father for the proper maiden name) and ended up erroneous bio father match.
It's a good tool, but it doesn't eliminate the need to do your own research on newly reported links.  On one set of connections reported by ThruLines I share 12 cM with a 5th cousin and 20 cM with one of their children.  While that's not completely impossible it seems improbable and warrants additional research.
So, the this works based on a combination of paper genealogy and dna genealogy. So even though my half sister matches me and my maternal grandfather I knew all my life, it shows up as a thruline with one match of my half sister (but not my cousin whose mom and a different dad. I would be more sure if some of my other cousins matched as well. But I have no other comparison except the one half sister so that does not give me confidence. My grandmother married him after all the kids were born, so....

18 Answers

+16 votes
 
Best answer

My initial impression - Truly groundbreaking!  Tremendous potential for some users!  This may very well demolish some brick walls!

I was as skeptical as others when I first heard about it, wasn't even interested in checking it out yet.  I've seen too many software updates that consisted of moving features around the page, renaming and re-hyping them, and calling it an update.  But late last night I needed a distraction, so took a look, not expecting anything.

I was quickly impressed with the way it organizes your lines and adds each of the people that match you, on each line.  That in itself is amazingly helpful, but not something to fall out of your chair for.  But as I was browsing around, I found one of my brick walls, Mary McConnell - with a suggested set of King parents and siblings - ALL with numerous people with matching DNA!!!  And some of them are at least 15 to 20 cM!  People have suggested various surnames for her, and I've removed them on FamilySearch, because NONE of them had a shred of evidence.  Someone somewhere had suggested her surname as King, but also without any evidence, but there was a line item in a distant church that said Robert McConnel married Mary King on a certain date - but no evidence at all that Robert or Mary had ever been near there, or involved with that church, and there were other Robert McConnells and lots of Mary's.  What's even more amazing, I have never found a McConnell tree with Mary King, *and* I cannot so far find any King trees with Mary!!!  So ThruLines came up with this on its own, because of the family of matching DNA testers!  Next step of course was to see if Mary was plausible in this King family.  The first child in the King trees appears to be born several years after the parents were married, and we all know there's a very high likelihood of a child within a year after marriage.  Where does Mary fit - a year after marriage and about a year before the known first child!  And since she was first and married young and left, and there aren't yet any family records during this period, that could explain why others did not know about her.

Obviously I cannot confirm her, but this DNA evidence is compelling.  Could she be a child of one of the uncles instead, it's possible, and will need checking.  But this still was an amazing feat for ThruLines!  It suggested something without any paper evidence, or anyone's Ancestry tree ideas!  In fact, the King trees I looked at are awful!  (To be fair, only a cursory look.)

The one sad thing - for each match they display the total cM and number of matching segments, but not where they are located!  You still need them to upload to GEDmatch.

by Rob Jacobson G2G6 Pilot (136k points)
selected by Andrew Payzant
That's great!  I was hoping for something similar for a couple of my own brick walls, but nothing so far.  Looks like I have some more support for my Lavender line, though.  I'm still lacking absolute paper trail confirmation that my great great grandfather Daniel Lavender is the same Daniel who's the son of David Lavender and (probably) Nancy Goins, but there are enough cousins matching on that line that I'm feeling pretty good about it.  Need to look at all these newly connected cousins more carefully to see if it all really makes sense.
Oooooh... I'm feeling even better about this now.  I have a whole pile of matches descended from Nancy Goin's siblings that I had never noticed before, mostly in the 15-25 cM range.  I think that pretty much confirms that you and I are half 4th cousins once removed.  I'll need to verify their paper trails and triangulate on Gedmatch to really lock it down, but this looks really good.
+15 votes
It was announced today at the RootsTech. It's still in Beta.
by Randolph Maynard G2G2 (2.7k points)
Oh! I'm not there and I'm not watching. That explains it arrival, but not how the DNA Project will view it here on WikiTree.
My internet decided to stutter and crash just as she was talking about ThruLines, so I have no idea what they are or what they do.  And the crash came because I opened up Ancestry to follow along with her!
+19 votes
I saw this, and as an ancestry.com user, I was curious. In the few days since it has rolled out, I have been in touch with two cousins whose link has been verified via DNA analysis. (They've both been Ancestry users for years, so why I didn't see them before Thrulines is a bit of a mystery.)

One of them is related to a family line that I have been researching in vain for the last 9 years. Her husband is my cousin and getting in contact with them has helped knock over a major brick wall for me.

Pretty impressive of only a few days on-line... !
by Alex Stronach G2G6 Pilot (363k points)
+20 votes

Hi G,

I've tested at FTDNA but not at AncestryDNA, so I'm not familiar with the specifics of the new ThruLines tool/feature there.

But my understanding is that the WikiTree guidelines for triangulation for DNA confirmation require (among many other things), the ability to confirm that 3 or more distant cousins all match each other on the same 12cM segment or greater. My understanding is that FTDNA and AncestryDNA haven't had tools available to show that (so far). GEDmatch provides such triangulation tools (other sites may as well).

While FTDNA provides tools that show me that I match multiple other individuals on a specific segment, the tools there don't disclose whether the other individuals all match each other for that same segment. If the new features at AncestryDNA do provide such capabilities, I would assume that the DNA Project (or whomever is responsible for setting the DNA confirmation guidelines here at WikiTree) would consider updating the WikiTree DNA confirmation guidelines.

by Rick Peterson G2G6 Pilot (187k points)
Well I have read most of the comments on this.  I KNOW what my heritage is on my mother's side.  The Thru-lines confirms what I know.  But as I understand it Wikitree doesn't want to accept that.  Getting some of those folks to get set up with Gedmatch is going to be impossible.  So DNA confirmation is still a pipe dream for most.
Confirmation is a spectrum, not a switch (as is any form of genealogical proof). From what you say you absolutely have some DNA confirmation (or call it corroboration, if you like). But Wikitree is set up with the confirmation “switch” on connections, so we have to draw an arbitrary line at what constitutes evidence enough to flip the switch. The standard currently adopted is that Ancestry DNA matches alone cannot constitute a basis for flipping the switch when the relationship between the testers is farther than 3rd cousins.
Triangulation is too strict to be useful. I like the term corroboration. Why would it be necessary to inherit the same DNA strand? We can have different bits in common with different cousins. We inherit different bits from our parents. I found that I don't share DNA with my cousin 6 generations back but we both share DNA from the same cousin 4 generations back, that cousin is in both thrulines. We should make use of what it provides, not complain about what it isn't.
+12 votes
In my humble opinion, DNA Circles, which showed interrelatedness among the group rather than just relatedness to me, using even more theoretical lines, was better proof than ThruLines. WikiTree will reject ThruLines for the same reasons it rejected Circles.

In reality, DNA confirmation as presented in WikiTree is not accurate period.  Every DNA connection is, just like a census record or birth certificate, is one more piece of evidence that you must evaluate as part of your proof.

Ultimately, if I have several genetic cousins who all descend from a common ancestor from many children, I might regard that as proof. But are these lines correct? Are any of them also related to each other? Could any be descended from another common ancestor? ThruLines doesn’t answer that without a lot of extra analysis. Of course, neither does Triangulation except for relatedness to each other.
by Davis Simpson G2G6 Mach 2 (26.9k points)
From what little I've seen so far your commentary about ThruLines seems correct. From the images above, it doesn't appear to provide anything that Circles wasn't already providing. You still can't say what part of which Chromosome you match anybody on.

As to DNA Confirmation here, it's not really intended to be absolutely 100% foolproof, but I expect it works extremely well the overwhelming majority of the time. I'm not sure if I've ever heard of a case on here where a properly-performed DNA Confirmation was later proved to be confirming the wrong ancestor.

The odds of COINCIDENTALLY matching someone on a specific part of a specific segment AND that person also having a bogus tree with that specific ancestor on it must be absurdly astronomical.
Blaine Bettinger discussed what you're talking about here earlier today at RootsTech.  Frank, it's not that someone's tree is bogus, it can be that the other people that you are matching don't have very fully developed trees.  So while you may come across a common ancestor, that isn't going to mean that's the ancestor from which you all got the shared segment.  There is a distinct possibility that you may share another ancestor from whom you got the DNA.  The less developed the tree, the higher the likelihood that this is the case.  With smaller segments, this becomes more likely.
Honestly, Darlene, that has a bit of a "snake oil" smell to it, if you don't mind me saying so. Where does he get this stuff, and why does anybody listen to him?

When you inherit a certain segment from a common gt-gt-gt grandparent, for example, the odds of even getting a match at all with someone in your own generation (a 4th cousin) is something like a 50/50 proposition. Then you manage to get a match from yet another 4th cousin - through sheer "force of numbers" - because it was just a very prolific family, or just from dumb luck.

And so this guy is trying to tell us that it doesn't mean anything, because the three of us might also have some OTHER unknown common ancestor, for whom the odds of getting that kind of match is even smaller - MUCH, MUCH smaller?

Doesn't it strike you as convenient, that this theory is completely unprovable - that practically nobody has a complete tree going back to all their 8th-gt grandparents, or wherever?

What is this guy trying to accomplish? Establishing a name for himself by throwing stones at accepted practices - the old "see how smart I am" scam - when what he's saying can only be idle speculation?

"Distinct possibility"? What, exactly, is THAT supposed to mean? Well, I'll tell you what it means. It means "I don't know what the probability is, and if somebody tried to calculate it, and gave you a number, you'd all laugh me off of this stage."

There's going to be cases where endogamy is a real issue, and the tools of genetic genealogy might be rendered somewhat useless - when your ancestry includes multiple ancestors who were from families that lived in close proximity where there was a small population, and other special cases. At least, useless for that part of your lineage. These are exceptional cases - and not hard to spot.

But what does he imagine anybody is going to do with this "profound wisdom"? Throw the proverbial baby out with the bath water, out of extreme paranoia? Is he even making some sort of specific recommendation?
I’ve was trying to identify a 5x great grandfather back in Cornwall, whose son migrated to Canada about 1830. Found a 10 cM match to man wo lived not far from the Cornish homeland who shared the same man in his tree. This was on Ancestry.com, so it probably wasn’t a endogamous match.

After analysis, I found that he also had some Irish ancestry and it was there that we matched. I know the family line he matched on but not the specific individual because we have brick walls in Ireland.

So, yes, unless we’re talking about a endogamous community, these are likely problems a bit further back when you both know your family tree.

But ultimately, you should treat DNA as any other piece of evidence. A match needs to be analyzed in context of the paper trail and with all your other DNA matches. Discrepancies need to be explained. If you’re dealing with 2 good trees and a 4th cousin, that should be easy.
Davis-

Indeed! That's case study in why triangulation is required beyond a certain point, and you need an additional person to match on the same segment.

Your case reminds me a bit of one of my higher matches on GEDmatch. A tree was attached, and I saw that the person was descended from a 5th-gt grandfather who lived in Germany, on my maternal grandfather's side (very cool!)

I assumed that that was where the relation was, and as I delved deeper into analyzing my GEDmatch matches, I eventually ran into a befuddling situation that I felt I needed to write down and sort out - I called the file "madness", or some such thing.

What I eventually figured out (after a pretty long time) was that my shared DNA with this person wasn't from that 5th-gt grandfather - we apparently share NONE from him. Heck, it wasn't even from Germany - it was Irish! And on my DAD's side!!

Now everything makes sense (the "madness" is gone). It's too far back to tell the exact relation, but I can still tell. By some OTHER bizarre coincidence, his wife is also some kind of distant Irish relation - to my MOM. My parents aren't even from the same state, and their ancestors haven't lived in the same place for at least 150 years.
I would be careful using Ancestry Thrulines to find potential ancestors ...it replaced one of my known ancestors with a person with a similar spelled name with someone from another tree . The owner of that tree didn't even have matching DNA .
Agreed. I have the same problem. A few family lines on other trees were wrong - likely just originally guesses but now widespread - but I found the read ancestors. However, this tool keeps suggesting the wrong people as potential ancestors. My genetic cousins will be tricked into believing DNA proved this incorrect ancestor.

Frank, I think you're misinterpreting what I said.  And the reason that we 'listen to' Blaine is that he's an authority on the subject.  You (and others) might like to watch Blaine's presentation (free to view):  https://www.rootstech.org/video/essential-considerations-for-dna-evidence-blaine-bettinger

True that you "can't say what part of which Chromosome you match," but you can drill down to # of shared cMs.

Here's one random example. Note that in virtually all cases, it's possible to "do the math" on what the relationship should be, assuming that the relationship is represented correctly.

You and NN(removed for privacy) share 10 cM. This table shows the percentage of the time people sharing 10 cM have the following relationships:

Percent Relationship
33%
5th cousin
4th cousin 2x removed
half 4th cousin 1x removed
half 3rd cousin 3x removed
27%
4th cousin 1x removed
half 4th cousin
half 3rd cousin 2x removed
3rd cousin 3x removed
20%
4th cousin
3rd cousin 2x removed
half 3rd cousin 1x removed
half 2nd cousin 3x removed
12%
3rd cousin 1x removed
half 3rd cousin
half 2nd cousin 2x removed
2nd cousin 3x removed
5%
3rd cousin
2nd cousin 2x removed
half 2nd cousin 1x removed
half 1st cousin 3x removed
<1%
2nd cousin 1x removed
half 2nd cousin
1st cousin 3x removed
half 1st cousin 2x removed

Fann,

That's an excellent example of the problems we're having with "authority figures". Just because somebody write a computer program, and numbers come out, it doesn't mean they know what they're doing.

I think of this output as an example of "backwards statistics". It's easier to do the "forward statistics" - by which I mean, for example, "What distribution of cM values do you see between 4th cousins?" To go the OTHER way, and infer a relation level from centimorgans is harder. My biggest concern is in taking into account how likely it is to see a 4th cousin vs a 5th cousin, or whatever - the "a priori" probabilities. I'm not sure those are very well-known.

10cM is a particularly extreme example - it's really scaping the bottom of the barrel. I've seen plenty of 6th cousins in my matches, yet 6th cousins isn't even listed - basically, they're saying a 0% probability.

So - obviously - these results are somewhat bogus. Beyond that, even if you BELIEVE these numbers, what it tells you is that the relation can be practically ANYTHING, as long as it's 2C1R or more distant. At this level, it tells you almost nothing about hoe closely related you are. To be useful for THAT you'd probably need at least about 100cM!
Can anyone point me to a source like an article, video, blog or forum post that explains how one gets in one of the DNA Circles. Or feel free to message me if you can explain. Yes, I am referring to the feature that preceded ThruLines.
I just goggled your question and the answer is.

You need at least 3 people no closer then 2nd cousin, all related thru the same ancestor to start a circle . And it can take months to happens .
Thank you, I am aware. I've read the same thing. Still don't understand it. I believe I meet the criteria on at least one set of common ancestors for example. (ThruLines suggests at least 34 DNA connections to this pair, and I do believe it's correct. Of course not everyone has those ancestors in their tree. I've not counted but I presume at least 3 of those do.) Been at it over a year without being part of any circles at all. Does it require you to also have your match in your tree and have them accept an invite via email? (That option you get when adding close relatives?) Something I am missing?
Sorry I don't have an answer for you ...it was a long time after I submitted my DNA till I was in any circles too ...you might try just calling Ancestry and get the answers you're looking for..... that's what I'd do ...

And even if you have a lot of people in those circles doesn't mean that they will all match your DNA but that they match some of the other members and the common ancestor ...so figure that out ...I sure haven't
+11 votes
It isn't using triangulation, or identical segments, and my experience of ThruLines thus far is that it's very very bad and shouldn't be implemented in its current state. It gave me "potential ancestor" hints for the third husband of my 3rd great-grandmother (by whom she had no children--I descend from her second marriage), and for a woman who was born in England and died in Australia who happened to have the same name as a lifelong-Virginian 4th great-grandmother, and also for people I've disproven.
by C Handy G2G6 Pilot (207k points)
My experience was the opposite.  I never expect much from other's trees on Ancestry, so false entries don't bother me there, I just look for the good and verifiable entries.  They do use identical segments though, and I'm curious why you think they don't?  Better yet, they validate them better than others do, throw out the 'by chance' ones, so even smaller segments are useful.

May I suggest you try it again?  Certainly you will want to ignore the trash suggestions, the result of other's poor trees, but if they have a DNA match, they do match you *somehow*.  The trick will be to figure out how those matches *do* match you.  Plus, I wouldn't be surprised if you find that ThruLines will surprise you elsewhere.  It has already helped a number of us find very interesting new discoveries.
I had a similar issue where it picked up the 2nd wife of my great great grandfather and show hints for her ancestry instead of my actual great great grandmother.  I also found a 5x great grandfather that wasn't on my tree and matched with 10 members from four of his offspring.  I think it has potential but I hope they work out the bugs.
I would like a button to "ignore" suggestions in Trulines that I know to be invalid due to errors in someone else's tree.  I have contacted the tree owners to provide records to illustrate the incorrect relationships in their trees, but have not heard back.  Consequently, 90% of the "potential ancestors" are invalid.  Though the 10% is gold.  

Alternately, a button to ignore specific trees would work as well.
+9 votes
Yes, it's VERY new. I've been afraid to click on stuff about it, for fear that it might do something irreversible that I won't like, but I suppose that's silly of me. At any rate, I thought I'd come in here and see what others said about it before "jumping in".

Maybe this explains what seemed like odd behavior of late, with the DNA Circles. I've had new matches show up who clearly ought to have been added into my DNA Circles, but my Circles stayed completely stagnant - no new people in weeks, if not months.
by Living Stanley G2G6 Mach 8 (90.0k points)
Yes no new circles and no shared ancestor hints either..a common way to tell when Ancestry is up to something. For weeks!
Try it Frank!  Yes, it's new and beta, and yes it depends on 'Garbage In Garbage Out', but already it has tremendous potential!
OK, I "dove in" and checked it out some.

Basically, it appears to be kind of the same thing as the "Shared Ancestor Hints" list, except that you're viewing a subset of that list that falls under a selected ancestor. It seems like a somewhat expanded list though, beyond the "Shared Ancestor Hints", so maybe it's a bit more sophisticated in what it does there.

Besides that re-formatting, I see two main benefits:

1) Ironically, while they officially and explicitly claim that the privacy of private trees shall be protected - and that this is very important - I've been able to finally see how some of my matches with private trees are related to me!

2) Additional matches from way down in the "Distant Cousins" list have been identified. HOWEVER, it's important to remember that it's MORE LIKE "SHARED HINTS", and fundamentally NOT really like "DNA CIRCLES" at all. Yes, matches are associated with ancestors, but here they're just matching trees - there's no Mystical Analogy to Triangulation.

I've come to find hints that my way-backist people on my Shared Hints list likely actually share DNA with me from that ancestor that has been found on both our trees, but recognize that this can just be a coincidence sometimes.
+11 votes
I played with this for a while yesterday. Initial impressions: positive. I've been critical of ancestry's DNA features for a while, given that they've been basically nonexistent, but I think this represents a step in the right direction.

Personally, I found several people and good hints for lines I hadn't previously had any information on, including one that's been a bit of a brick wall. I did run into a couple odd connections, but no blatant false positives, and even then that's preferable to me than circles which haven't done anything for me whatsoever.

The ancestryDNA product still needs *a lot* of work, but I'm hopeful this is the start of something good. I'd love to see direct matching for multiple people, including more detailed information on what segments specifically are shared - if they do ever put that in, they'll be pretty close to providing equal functionality to some other alternatives and might push me over the edge to recommending people test with them over others (they already have the advantage of biggest test pool).
by John Trotter G2G6 Mach 4 (42.4k points)
+14 votes

MyHeritage also just rolled out some new toys . . . I mean tools!  I stayed up late last night reviewing the information.

One is an automatic tool that organizes your DNA Matches into clusters that likely descended from common ancestors.  It's been taking a few hours to receive the email with the file, but it's great.

Another is their 'Theory of family relativity', which provides you with a few theories of how some of your DNA matches are related to you (i.e. it gives you one or more lineage charts for how you and your DNA match might be related up to a theoretical common ancestor).

I definitely recommend checking them out if you're on MyHeritage!

by Darlene Athey-Hill G2G6 Pilot (534k points)
I am also on My Heritage. The Autoclustering tool is AWESOME IMO!!

 The Theory of Family relativity, not so much. It is pretty much the same as the Thrulines from Ancestry!!  It relies heavily on genealogy and NOT on DNA. Whereas the autoclustyering tool is all DNA and not genealogy.

The Theory of family relativity tries to connect you to a DNA match via a common ancestor and also through other peoples trees and records.

If those other peoples trees have incorrect details, then the connection path will be wrong. The common ancestor may also be wrong.

I have a few theory lines like that and those that are beyond my second cousins (for which I have genealogical proof and evidence) are just plain wrong!!

Hopefully genealogical trees and records will improve, but for now, the family relativity tool is not very helpful.
Has anyone tried to download the autoclustering to an iPad? I tried and just got a table full of numbers next to names. I am so confused as to why I can not get it to work. Any tips?
Missy, I have had no luck trying to download autocluster information.  I think you either need to review it while on-line, or take screen shots.

Edit:  I was using the autocluster options from GEDmatch and Genetic Affairs.  I haven't used MyHeritage.
On MyHeritage, when I clicked the Generate button on the AutoCluster page, the results were emailed to me in a zip file. I extracted the contents of the zip file (using 7-Zip), resulting in HTML and CSV versions of the AutoCluster results as well as a ReadME PDF. I was able to successfully open the HTML version in my web browser (visually shows the clusters of matches). If you want to load the resulting data into a spreadsheet, the CSV version is likely what you want.
But Rick, if you are using your web browser to access the visual display, then as I understand it, you haven't downloaded the visual display because you have to be on-line to use it.  Is that right?

Missy's asking about an iPad.  As Rick said, it works fine on a desktop.  The problem may be that you need to be sure to save the file when you get it as opposed to opening it from the email that they send.  Once you save the file, then you can unzip it and will have three files.  Here's information on zip files for iPhones or iPads:  https://www.howtogeek.com/fyi/how-to-unzip-a-zip-file-with-your-iphone-or-ipads-files-app/

Hi Julie, MyHeritage doesn't appear to have an online version of the AutoCluster results available on their web site. They email the results to us, when we request the AutoCluster info to be generated. Once I've extracted the attached files, they're on my device, so I don't need to be online to access them. I'm using my web browser to access an HTML file that's physically located on my device.

Hopefully Darlene's comment can help with accessing the emailed files on an iPad (I have no experience with that particular device).
Thank you.
Thank you, Darlene. I will check out your link.
I just tried this out on my iPhone - running IOS-12  (my iPad is not connected to email).  The apple mail app gives you an option to "preview" the attached zip file and selecting that actually runs the html and embedded script to display the cluster.  It requires a lot of zooming and panning on a phone, but it does work.
+16 votes
There is no way this can be considered triangulation.  It's really a new way to see 'potential ancestor hints', it's not taking the DNA into account at all, except that someone is a match.  

The main benefit I can see is that it can help bridge the gap for you for with some of your matches where it suggests the missing people in your trees, between you and your matches.  I found two distant matches this way yesterday.  People I hadn't really looked at because the match was so small.  You still need to do the research so that you can confirm their suggested ancestor, as with any automated system - garbage in, garbage out!  Whilst I've identified a 'genealogical' connection with these two people they can still only be called tentative 'genetic' confirmations without the segment data to back it up.
by Veronica Williams G2G6 Pilot (213k points)
Exactly!

I'm looking at it as a sort of expansion of the already-existing "Shared Ancestor Hint" list. Some differences come to mind:

1) The presentation - it's inherently organized by ancestor, instead of everybody all mixed together

2) That presentation also exposes, to a certain degree how people with private trees are related. Living people are still blanked out, but it shows where they fit on your tree.

3) It seems like I'm seeing more matches than on that Shared Ancestor Hint list, so they must have an enhanced algorithm for doing that stuff.

Like the already-existing Shared Ancestor Hints list, you wouldn't liken THAT stuff to DNA Circles, or talk about using them "as is" for DNA confirmation, but you CAN use those results to try to get a real confirmation going.

For what it's worth, my own limited experience is that the matches these thing come up with really DO turn out to be related the way it suggests. I basically laughed at a super-distant cousin (7th cousin) match on there at first, but after I discovered that I also match her 2nd cousin, with one of their MCRAs being in the line that leads to me, it really got my attention. I even have an 8th cousin on that list, which I also wouldn't take too seriously, but it's the SAME ancestral line as the 7th cousin, plus one more generation back!
Yes, I can see lots of potential for things like you have described, will keep us busy!!  

Whilst AncestryDNA circles weren't triangulation either they were a nice thing to look at and help you hone in on other matches who might help your research even if you didn't match them yourself.

I'd still like to see both in the future!
Mostly useless because they are mainly based upon Ancestry Trees which are frequently wrong, misleading, and unsourced. Pretty useless, for competent researchers anyhow.
+15 votes
ThruLines has provided me with a clue to verify an oral tradition about an ancestor who changed his name. Until today I had not found any DNA matches to his putative original family. But I now have a DNA match from the putative family for both my mother and her second cousin, descended from the ancestor who changed his name. At 9.5cM, it's small and not one I would have investigated manually. And in the end, it may not triangulate, I'll have to see if I can persuade the parties involved to upload to GedMatch. But after 25 years of searching for evidence of the oral tradition, this is a welcome lead.
by Nic Donnelly G2G6 Mach 8 (80.1k points)
Wow!

Really, there are a lot of great matches buried down within all those "Distant Cousin" matches. Anything labelled a "Good Confidence" match basically has to be a distant relation at SOME level, and even cousins as close as 3C can show up even lower than that (I have one as low as 7.0cM). Plus, AncestryDNA throws away certain DNA, so your match on GEDmatch will undoubtedly be stronger.

Sometimes you can pick up on those low cM matches by doing surname or geography searches. Another way is to have a sibling tested, and look at their results. They could have a really good match to someone you match only very weakly (or not at all).
Yes, there are a lot of gold nuggets in those low cM matches. In addition to your suggestions, I've also found using the auto-clustering tool at Genetic Affairs (which MyHeritage has also just introduced) a good way to identify distant matches worthy of further investigation. Because it looks at all the Ancestry shared matches for those distant cousins and makes the link where possible. Sometimes a low cM match to me may be a higher cM match to a known second or third cousin, so will show up as a shared match, but manually checking this for thousands of distant Ancestry matches is time-consuming. The auto-clustering tool makes this so much easier!
+11 votes

I for one like it overall, but am worried about those that just take any 'hint' as being totally true. What Thrulines does is takes information from those that match you DNA-wise, and tries to extrapolate where your most common resent ancestor (MRCA) could be based upon the trees built by those matches.

On the positive side, it finds "potential" common ancestors (with your DNA matches) regardless of spellings, typos or differing names, I never would have found some of these … (Herman, Herrman, Hermann, etc). It also helps identify DNA relatives that you may have overlooked otherwise.

On the negative side, some of these matched ancestors have no sources or other information, and are based on other Ancestry trees with no sources (circular reporting). Some of those spelling variations or typos could be misleading (Schoenemann, Schuneman, Shoneman, Schoeneman, Schainemann, etc.).

So while this is still in Beta, just keep in mind that it is just another tool, and it is NOT a magic bullet for DNA triangulation. You still have to do the homework to make sure the information is correct just like any other tool you use.

by Ken Parman G2G6 Pilot (117k points)
+7 votes

Thus far, I have found ThruLines to be useful. It's relative utility is sort of demonstrated on a different G2G: apparent extra-marital daughter Martha(2) Strickland as shown by DNA.

With yet another line in my ancestry, I had struggled with a gr-grandparent who never married, but obviously had children. Because of the CM-cousin matching data, the relationship that I had found to be supported by circumstantial (e.g., census) data was in fact supported by DNA match reports on (once again) cousins.

So for me, that's 2/2 where the utility was "performs as expected," and at least "meets expectations." I would say that I actually got more value from it than I had anticipated and that I was pleasantly surprised. 

The interface is somewhat cumbersome, I will say.

As you may note by my G2G query, linked above, exactly what can and should be said on a profile that is "proven" by the DNA results, the jury's still out. Hello? wink

by Porter Fann G2G6 Mach 9 (92.7k points)
+9 votes
My take on the new Ancestry Thru-Lines: I have played with it and I can see some positives. My great wish for the final version is that Ancestry would give us the ability to delete a "Potential Ancestor" from our view when we know he or she is the wrong person for our tree. Right now there are several showing up for me that are wrong. They are there because other users have them in their trees (wrong or no research). Important reminder - we have a DNA match to the tree owner and that does not mean that tree is 100% correct!! Trust your instincts and your own research!!!
by Virginia Fields G2G Astronaut (1.2m points)
Virginia, I see some good things coming out of Ancestry Thru Lines and yes one drawback is indeed the fact that  they do not allow us to delete a wrong person. I have  GG Grandparents that have never left England but still I have several  that they insist my GG Grandparents are from Virginia USA No matter how much proof they have or no matter how many people do accept this clue without reading it, My GG Grandparents were never in Virginia, USA. A delete would sure come in handy.
+7 votes
I think ThruLines is a major improvement and will be a very useful tool.  Ancestry's focus on family trees is both a strength (because Ancestry's family trees are easily accessible and fairly complete for some of the people tested) and a weakness (because of difficulty differentiating poorly sourced trees).  I found ThruLines provided a number of promising clues I would not have found by manually sorting through family trees of my DNA matches, but I still need to independently verify the data on the linked family trees.  In a few instances the AncestryDNA tree merging algorithm appeared to have a bug (or perhaps it was errors in the trees it was accessing) which either skipped or duplicated a generation; but I feel confident AncestryDNA can debug this problem, and this ThruLines feature is worth the effort to do so.
by AL Wellman G2G6 Mach 1 (19.6k points)
+4 votes
I don't care about triangulation, it is too strict. When many paper lines and DNA lines converge that is exciting and suggests that the paper lines make sense. I would like a way to show it other than a long description on my profile. We had a way to show DNA circles. This is too interesting to ignore.
by Sue Hall G2G6 Pilot (167k points)
+7 votes

I have also found new cousins that I never found before. I will often use ThruLines to start my DNA confirmation trail. Then I go look for them on GEDMatch and/or MyHeritage and analyse the DNA. If they are not there, I send a message to them explaining what I am doing, and ask if they mind helping out. I have all but one response and several have uploaded to GEDMatch. So I would say I have had a good experience with ThruLines and the cousins I have found (because of it) on Ancestry.

Missy smiley
by Missy Berryann G2G6 Pilot (216k points)
+8 votes
I also recently discovered this, and think it is a very good way to confirm relationships, but you have to realize that the names in the family tree showing descend are only the ones people post on Ancestry, and not all of them are accurate. So the DNA match can be accurate, but the line showing how you are related may not always be correct. It could be off a generation or two or someone entered a wrong sibling or cousin as an ancestor in their family tree. Also if your family was in the same area for more than three or four generations, there may be other families lines you share with a particular individual who is a match. That is very true in my case and I have to figure out which line is closest!
by Daniel Bly G2G6 Mach 8 (83.5k points)
I agree, David. It is just as likely that a Through Lines DNA match has the incorrect genealogy as not. My example:  Jean Pinel b 1731 and his cousin  Jean Pinel b 1738, different parents, in  Quebec. The old transcripts had both men married to the same woman, and dying in the same place. On Ancestry, I first selected Jean b 1731 and had a lot of Through Line matches. Was this the correct Jean, based on my matches?  To double check I selected Jean b 1738, as the proper ancestor, and waited a couple of days, Again, I  had a lot of Through  Line matches show  up. Through lines results Didn’t care who the correct man was. The results depended upon which man the Through Lines matches had attached as the husband.  After subscribing to Genealogies Quebec, and viewing original records, we found a death record, husband of the proper wife, in the right town, for Jean Pinel, d 1810, age 79. That’s as close as I can come to proving the correct ancestor was Jean b 1731.

And this shows up a possible  fault in Thrulines, because the DNA could just as likely  have come through the mother, and so the same people should have been connected through Thrulines whoever the father was listed as.

Related questions

+3 votes
1 answer
+40 votes
9 answers
+15 votes
10 answers
960 views asked Mar 14, 2020 in The Tree House by Bob Scrivens G2G6 Mach 2 (21.1k points)
+8 votes
9 answers
318 views asked Feb 14, 2020 in The Tree House by Craig Albrechtson G2G6 Pilot (102k points)
+9 votes
3 answers
1.2k views asked Mar 8, 2019 in The Tree House by Tim Campbell G2G4 (4.0k points)
+8 votes
1 answer
+8 votes
3 answers
484 views asked Nov 10, 2020 in Genealogy Help by Shirlea Smith G2G6 Pilot (280k points)
+27 votes
24 answers

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

disclaimer - terms - copyright

...