Ancestry’s Better Mousetrap – DNA Circles

This is it…the big day.

Ancestry’s better mousetrap is called DNA Circles and it launched today.

DNA Circles is a result of three things.

  1. Phased data
  2. Improved genetic Matching
  3. Pairing DNA matches with submitted trees

Yesterday I wrote about my matches in the old version.  So, let’s take a look at the new version, available now.

All three of the autosomal DNA genetic genealogy testing companies have the same issue and that’s how to provide us with quality matches, eliminate false IBS matches while preserving real ones, and making the consumer experience both productive and easy to use. All three of the companies approach this challenge in different ways.

23andMe has an arbitrary cutoff on the number of matches you can have, at 1000, unless you’re in contact with your matches and then you are allowed more. Family Tree DNA has both a cumulative match threshold of about 20cM and then an individual segment threshold of about 7.7cM.  The word “about” appears in that last sentence because the matching algorithm contains some situational variables.  Until today, Ancestry really didn’t have a good tool to eliminate low confidence, spurious or IBS (identical by state) matches.

At 23andMe, I have just over 1000 matches, which is to be expected based on their 1000 cutoff. At Family Tree DNA, I have about 1875 matches and at Ancestry, until today, I had over 13,000 matches.  Clearly, Ancestry needed to refine their matching process, and they have.

Ancestry has implemented population based phasing to help reduce false positive matches. Blaine Bettinger wrote an excellent article about how Ancestry is accomplishing this task, why it works, and how, in his article, Finding Genetic Cousins – Separating Fact From Fiction.

As I described in my article, DNA Day with Ancestry, Ancestry has discovered that we all have what they describe as pileup areas where many people from the same population will match.  This means that those matches, while they do come from specific ancestors, aren’t actually genealogical in the way we might think.

genome pileups

Here’s an example of my own genome and my pileup areas, as provided by Ancestry.

You can see that in one region I have almost 800 matches – and clearly that’s not from one ancestor, especially given that most of my match numbers are under 200, and most are significantly under 200.

genome pileups2

Here’s my same chart AFTER they ran the phasing algorithm on my matches and removed those pileup areas. Please note that the scale is different.  Now my highest number of matches is about 25.

Are some of those phased regions probably valid matches? Sure.  Are some of them occurring in people whom I match in other regions too?  Of course.  And those people will remain as matches, where people I only match on pileup regions will be removed.  In other words, any match to me in a pileup region won’t be considered a match, regardless of how many other places we match.

Ancestry did not provide us with a list of regions by chromosome that were removed in the experiment above. I wish they had, because I have a couple of chromosomal areas that I’ve been finding confusing because I have multiple matches with proven connections to specific different families from the same parental line that match me on the same segments.  Let me say that again, another way.  On Mom’s side, two different families match me on the same chromosomal segment region.

Now, unless those separate families are interrelated, that is impossible.  Those families being interrelated certainly isn’t impossible, but given one line is French (Acadian) pre-1600 and one is Swiss Brethren from the mid-1600s, an interrelationship between these families had to have occurred before 1600 which is more than 12 generations ago – and probably many more generations before that, given their strong religious leanings and lack of geographic proximity.

So, I’m presuming here that these confusing segments are an example of pileups and that explains why the multiple family lines match to the same segments.

Ancestry’s Updated Product

So how has this new technology changed your Ancestry results?

  • New Home Page
  • Updated Match List
  • DNA Circles
  • Updated Help Page and White Papers

circles

Your home page now has a new category, DNA Circles.

But first, before we look at the circles, let’s look the matches.

Matches

Yesterday, I reported on my matches and how they were distributed. I had 262 pages of matches, or about 13,100.  Today, I have 67 pages, or about 3,350 matches.  My matches were reduced by about 75%.

Yesterday Today Shakey Leaves Yesterday Shakey Leaves Today
Total Matches 13,100 3,350
 
2nd Cousins 1 – 99% confidence 0 – shifted to third cousin 0 0
3rd Cousins 10 8 – shifted to fourth cousins 2 1 (shifted to 4th cousin)
4th Cousins 243 161 10 14
Distant Cousins 12,846 3,181 36 18

Of the fourth cousin shakey leaf people, three that were distant cousins are now shifted up into the fourth cousin range, my third cousin is shifted down to fourth cousin range, and one prior fourth cousin shakey leaf match is gone entirely.

However, the numbers aren’t the entire story. I compared my list of shakey leaf people from yesterday to today, and I discovered that some were missing, but I also have 6 new shakey leaf matches in the distant cousin category that I didn’t have yesterday.

And one of those shakey leaf matches, if it is correct – meaning that if the DNA does point to the genealogy – would shatter a very long-standing brick wall.

Now, before I share this with you, I want to be very, VERY clear – just because we share DNA and a common genealogy line does NOT MEAN that we are genetically connected via this genealogy path. However, having said that, it’s a very good hint and a wonderful place to start.

In my case, Elijah Vannoy was born in1784 to one of 4 Vannoy men in Wilkes County, NC. The question is, which one?  Based on census, tax, Bible and other records, I’ve positively eliminated one candidate and probably eliminated a second.  But that leaves two and possibly a third.  I decided a long time ago that this quandry would and could only be solved via a DNA connection to the wife’s line of the men involved.

  • Nathaniel Vannoy married Elizabeth Ray (Rey) – Eliminated as Elijah’s possible father via Nathaniel’s Bible record
  • Andrew Vannoy married Susannah Sheppard (I am related to Susannah’s father through a different family line.)
  • Francis Vannoy married Millicent Henderson
  • Daniel Vannoy married Sarah Hickerson.  Her parents were Charles Hickerson and Mary Lytle.

Based on tax lists that include males of specific ages, my “best choice” is Daniel Vannoy and Sarah Hickerson. That’s who I have in my tree at Ancestry, even though I strongly debated entering that couple since it was so tentative.  Am I EVER glad that I did.

Here’s my new match.

circles2

I can tell you, when I saw this, it took my breath away!  Lordy, lordy, I’ve caught a mouse.  But now what do I do with it???

Now, for the frustrating-makes-me-screaming-insane part – I have NO WAY TO VERIFY THIS without a chromosome browser. So, what am I going to do?  I’m going to contact this person, and pray, PRAY, that they reply to me.  I’ll be glad to pay for them to transfer to Family Tree DNA where I have a chromosome browser to work with and can prove that this individual indeed does match other descendants of Elijah Vannoy and not just me.

If this is just true….

But wait, maybe there is more evidence at Ancestry. Let’s look at their new DNA Circles.

DNA Circles

DNA Circles is a composite tool that links people who are genetically connected with people who have the same ancestors in their trees, and puts them together in a circle.

In other words, all of these people genetically match at least one other person in the circle, but they don’t all match each other. The only matches you can see are people that match you.  The common link, is, of course, that in addition to genetically matching someone in the circle, they all share a common ancestor in their tree.  Now, yes, it does go without saying that if everyone has the same wrong ancestor – the circle will show that ancestor. Conversely, if you are the only one with the right ancestor’s name, and everyone else has the wrong name, then you won’t be shown in that circle.

Now, for the caveats.

You must be an Ancestry subscriber to see Circles.

If you have a private tree, Ancestry is respecting your request to remain private and you will not be included in Circles.  If you make your tree public, you may or may not have circles.  Not everyone does.  Ancestry updates their data base every 3-4 hours, so if you make your tree public, it won’t take effect immediately.

Of course, if you have no tree, there is no way to include you in any circles.  Ancestry is looking back 7 generations for circles, so if you’re entering a tree, enter at least 7 generations.

Having said that, both private trees and no tree matches are still included in match lists, if they pass the new matching criteria, but they won’t be included in the new Circles feature.

So, let’s take a look. Please note that the new Circles feature is in Beta.

Here are my 12 DNA Circles.  I was actually surprised that there weren’t more.  However, one person in our blogger group had no circles.  How disappointing.

circles3

Sadly, the Hickerson ancestor I was hoping to see is not identified as a circle. Maybe someday.

Let’s look at my smallest circle, Jacob Lentz.

circles4

Ancestry refers to this as an emerging circle. I match one individual genetically, but not the second individual, which I would presume (how I hate that word) means that H.C. and pawruby match each other genetically.  How I would love to see the three of us in a chromosome browser.

I can click on “View Details” to see how they both connect to Jacob.

circles5

The tree above is from my DNA match. The tree below is from the other member of the circle who I don’t match genetically, but who presumably matches H.C.

circles6-2

Jacob Lentz’s wife is Frederica Moselman or Musselman. The spelling of the name varies in documents.  I was curious as to why there is no circle for Frederica, so I looked to see if perhaps her name is absent from the trees.  As it turns out, two trees show her as Moselman and one as Musselman, so the disparate spelling has defeated the creation of her circle.  During the discussions with Ancestry about this product, I specifically asked about situations like this and they indicated that they have soundex and other matching tools and they felt that this would not be a problem.  Obviously, in this case, and others, those tools didn’t work.

If you want to learn more about how DNA Circles works, and you are a member of a DNA Circle, click on the “Learn More” button at the bottom of the DNA Circles information box.

circles7

Learn more takes you to this page where you can read about how the circles are created, grouped and the white paper which describes the technology behind the circles.

circles8

My larger Nancy Mann circle shows that I have 12 members in this circle, of which I match 4 by DNA and the rest have a DNA connection with other member(s) of the group. We all have a common ancestor in our trees – Nancy Mann.

To clear up any misconceptions here, ancestry has very specifically stated that they are NOT using trees to do DNA matches, but only after DNA matching is completed, they are searching for common ancestors in trees of matches.

circles9

Of the Nancy Mann circle members, I match 4 people utilizing DNA. Three of those show on my match list, but one, C.M. doesn’t show on my match list today nor on my old list.  This is a strong match, so I find this confusing.

One of my non-DNA tree matches used to be a DNA match, but isn’t anymore. This would be one example of where a legitimate match was removed by the new matching routines, but I can still see that there is a circle connection to a common ancestor.  While Circles don’t confirm a genetic connection, they are another tool that is certainly suggestive that the DNA connections between these individuals lead to a common ancestor.

Nancy Mann’s husband was Henry Bolton. She was his second wife, so there will be people who connect to Henry, via his first wife, but not to Nancy Mann.  What this means is that everyone in Nancy’s circle should also be in Henry’s circle, but some people in Henry’s circle won’t be in Nancy’s circle.

When looking at why someone in my Nancy Mann circle wasn’t in my Henry Bolton circle, I noticed that Williamlowe94 does list Henry Bolton, but has spelled his name “Henry Bolton (Boulton)” and apparently the parenthesis name was considered a non-match. C. M. has spelled Henry’s name Boulton, so that’s why C.M. is in the Nancy Mann group, but not the Henry Bolton group.

Another circle, Joseph Preston Bolton, was Henry Bolton’s son. There are 4 members of that circle, one of which I match via DNA.  There is one new member of this group that is not in the Henry Bolton group, and who is not on my DNA match list.  I wondered why they aren’t on Henry’s list, so I looked at their pedigree chart and their chart stops at Joseph Preston Bolton.  This would seem to be a good opportunity for Ancestry to utilize the power of their software to see if she actually DOES fit into the Henry Bolton and Nancy Mann circles and suggest to her that in fact, she does.  For her, this might indeed tear down a brick wall.  Most people aren’t looking for confirmation of what they have, they are looking for that next step – that elusive ancestor who isn’t identified.

That is why we do DNA, and genealogy.

The John Campbell tree only has 3 members and both of the other Circle members are a DNA match to me. Of course, that doesn’t mean they are a DNA match to each other.  All 3 of us show John’s wife to be exactly the same person, spelled exactly the same way Jane “Jenny” Dobkins, but there is no circle for her.  I wonder if somehow the quotes interfered with the circle creation.  Given that all 3 of us form a circle for John, we should also form that exact same circle for Jane.

Fairwick Claxton and Agnes Muncy hold another odd match. One charlenecarlson0126 shows to be both a DNA match and a tree match, but she does not appear on my DNA match list, nor does her tree include any Claxton or Clarkson at all.  This has to be a bug of some sort, but it seems odd that it would pass both criteria, DNA matching and the tree.

circles10

Match above, tree below.

circles11

What I was actually searching for is why Fairwick’s father, James Lee Clarkson/Clarkston/Claxton is not listed as a circle. My suspicion is that the name is not spelled consistently.  Of the 5 Circle members, one is spelled, Claxton, 2 Clarkson and 2 Clarkston.  This looks like another miss that could be a hit.

My John Hill circle is actually quite interesting. There are only 3 people and I match one via DNA.  I’ve been working with my non-DNA match on this genealogy line.  It’s nice to see him in the Circle, even though our DNA doesn’t match directly.

The John Hill group, again, begs the question of why there is no wife’s group. She was Catherine Mitchell and all 3 of us list her as such.

In Summary

Ancestry has certainly improved their methodology and utilized their new tools to add the DNA Circles feature.

Certainly, we had too many matches to deal with before and now we have a much more reasonable number. Ancestry’s shakey leaf remains one of the best tools they have ever implemented and their user interface remains clean, crisp and easy to use.  There are a few bugs, but this is a beta version and with feedback, I’m sure they will resolve those in short order.

In order to get a handle on what was really occurring, I created a spreadsheet of my pre-Circles shakey-leaf matches as compared with my matches in the new Circles version. The individuals in bold are the ones that appear in both versions, the pre and post Circles.  Non-bolded were in one or the other versions, but not both.  In some cases, like with the first 4 matches in this group, I wonder why they don’t form a James Lee Claxton group.  Me plus two more would be enough for an emerging group, and we have that for sure.

Shakey Leaf Matches and Ancestor Previous Current Circle Members
Rodneybranch1 – James Lee Claxton and Sarah “Sary” Cook distant gone
urbadntx – James Lee Claxton and Sary Cook absent distant
Ctkatherine – Fairwick Claxton and Agnes Muncy 4 4 Fairwick Claxton, Agnes Muncy
Dbreeding63 – Fairwix Claxton and Agnes Muncy 4 4 Fairwick Claxton, Agnes Muncy
charlenecarlson0126 Fairwick Claxton, Agnes Muncy
Petwin73 – John Hill and Catherine Mitchell distant gone John Hill
Greatpyr616 – Henry Bolton and Nancy Mann distant distant Nancy Mann, Henry Bolton
Marsha Bolton – Henry Bolton and Nancy Mann distant gone Nancy Mann
Ctlynch01 – Henry Bolton and Nancy Mann distant gone
C.L.M. – Henry Bolton and Nancy Mann distant distant
Tjfhorn1 – Henry Bolton and Nancy Mann distant gone
johnryder42 – Nancy Mann absent distant Nancy Mann, Henry Bolton
Dblrich – Honore Lore and Marie Lafaille distant distant
Rkoelpin – Francois Lafaille distant gone
William Lowe94 – Joseph Preston Bolton distant distant Nancy Mann, Joseph Bolton
E.J.H. – John Francis Vannoy and Susannah Anderson distant gone
Rheainhatton – Francis Vannoy and Catherine Anderson distant gone
Viero111777 – John Francis Vannoy and Susannah Anderson distant gone
Maggiejames113 – John Francis Vannoy and Susannah Anderson distant gone
J.M. – John Vanoy distant gone
annelynnward1 – Jothan Brown absent distant
RWECIII – Jotham Brown distant gone
Raymond Brown – Jotham Brown distant distant
Tgbils917 – Jotham Brown distant gone
Skyrider3277 – Jotham Brown distant gone
Browndavid239 – Jotham Brown distant distant
R.G. – John R. Estes and Nancy Ann Moore distant gone
Chuck2810 – John R. Estes and Nancy Ann Moore distant distant
Lodikid – Andrew McKee distant distant
C.A.W. – Daniel Miller and Elizabeth Ulrich distant distant
Ostate4454 – John Campbell and Jane “Jenny” Dobkins distant distant John Campbell
melby01 – John Campbell and Jane Dobkins absent distant John Campbell
A.F.B. – Nicholas Speaks and Sarah Faires distant gone
nellf_1 – Nicholas Speaks and Sarah Faires absent distant Nicholas Speaks, Sarah Faires
Razzanozoo1 – Lois McNiel distant gone
EHVannoy – Joel Vannoy and Phoebe Crumley 3 3 Joel Vannoy, Phoebe Crumley
D.V. – Joel Vannoy and Phoebe Crumley 3 4 Joel Vannoy, Phoebe Crumley
Spklegirl- Francois LaFaille 4 gone
H.C. – Jacob Lentz and Frederica Moselman 4 distant Jacob Lentz
Alyssa- Lazarus Estes and Elizabeth Vannoy 4 4 Joel Vannoy, Phoebe Crumley
J.L.B. – Daniel Miller and Elizabeth Ulrich 3 4
drjcox51 – Charles Hickerson and Mary Lytle absent distant
M.S. – private tree distant gone Not in circles
Christine414 – private tree distant gone Not in circles
DDicksson – private tree (Jane Dobkins and John Campbell) distant distant Not in circles
FruitofVine – private tree distant gone Not in circles
Lisa36ang – private tree distant distant Not in circles
J.M.F. – private tree distant gone Not in circles
1_perry22 – private tree distant gone Not in circles
Jcarolynbh – private tree distant gone Not in circles
Nanbowjack – private tree 4 4 Not in circles
L.W. – private tree (John R. Estes) 4 4 Not in circles
P.B. – private tree 4 4 Not in circles
1_cmarse – private tree 4 4 Not in circles
MDgenealogy20 – private tree 4 4 Not in circles
Susanharmon – private tree 4 4 Not in circles

Obviously, several people are in multiple circles.  There are a total of 15 DNA matches distributed between 12 circles.  That leaves 3,335 matches that aren’t helping me or correlated in any way.  While I do like the circles, I’m disappointed that so few of my matches sync up with pedigree charts.  It looks like there would be a lot more if Ancestry would review the matching routine, and perhaps more yet if they would reach beyond 7 generations.  But first steps first.

Some circles contain only DNA matches.  Others have more non-DNA matches (to me) but have a pedigree match to everyone in the DNA Circle. That’s really what these are, DNA circles that happen to have a common ancestor in their family tree.

Does a circle confirm that the connection to that ancestor is via DNA? Nope.  Does it confirm that your DNA connection to your match is from that ancestor?  Nope.  You still need a chromosome browser to do that – but this certainly helps.  It’s a step in the right direction.  It gives us another tool.  And, in some cases, like my Elijah Vannoy, changing the suspected parents periodically from one possibility to the other might be viewed as a new method of fishing.  So might changing the surname spelling.

And regarding that chromosome browser from Ancestry, well, all I can say is don’t hold your breath…

Truthfully, I’ll tell you exactly when we’ll get a chromosome browser.

Tim Sullivan, Ancestry’s CEO, is a genealogist, just like the rest of us. The day he has to transfer his autosomal file to a competitor to use their chromosome browser to confirm an ancestral match…well…I’m betting that’s the day a chromosome browser will become a priority for Ancestry.

So Tim, my friend, I wish for you a lot of new circles – including one just like my Hickerson match – one that you have been desperately seeking for say, about 30 years. Wouldn’t that be a great Christmas gift?  But, you see, I know that having a hint but not knowing, i.e., no proof, is going to just about kill you.  It will break your genealogist’s heart.  It will make you beat-your-head-against-the-wall insane.  Screaming yellow zonkers nuts.  I don’t want that to happen to you, or anyone else, for that matter.

So, while you’re waiting for Ancestry’s chromosome browser to be developed, here’s the link to Family Tree DNA so you can confirm your genetic ancestral match…assuming of course that you can also convince the other people to download their results from Ancestry to Family Tree DNA as well:)

90 thoughts on “Ancestry’s Better Mousetrap – DNA Circles

  1. Interesting stuff. Looks like they are making some progress.
    However I wish my experience was even a little bit like yours. I have zero shaky leaves and no DNA circles, and yes I do have a tree. I only have 33 matches, of which 11 are 4th cousins – nothing closer.
    What is my problem?

    • Unlucky? Recently immigrate from another country? Your ancestors carry a gene that makes them genetically resistant to DNA testing? (Just kidding, although I’ve often wondered if my family line has that.)

      • No circles, but I am not a paying member. After being unable to access my paid account for 3 1/2 months due to problems on their end and getting no response or credit for the lost time, I gave up my paid subscription last year. Took another 3 1/2 months before they fixed the problem.

        I have 7 shaky leaf hints, 106 4th cousins with nothing closer than 3rd cousin, once removed. Beginning to wonder if I have an NPE a closer than expected. A big plus was seeing how one 3C 1R might match through my great-grandfather and his sister.

      • Actually your first reason may be the one, although I emigrated from England in 1966. Interestingly I also have a John Hill ancestor although with such a common name he is probably no relation. Mine was born in Redditch, England in 1827.

  2. Huh huh. Fairwick Claxton is the best name ever. Fantastic post, Roberta. My number of matchpages fell from c320 to 59, so the Ashkenazic endogamy issue has been addressed. I want to make time to read their white paper. The circle idea is sort of ingenious since I think it will induce a more open nature in the severely reserved amongst us, especially once the implications of Ancestry having no chromosome browser are successfully explained to them. Anyway, thanks for all you do to teach me! I depend on your analyses.

  3. I didn’t see this mentioned, but for anyone wondering, the option to download your previous match list is under Settings and says it is available only for a limited time.

  4. Personally, in the grand scheme of things, ancestry’s tools are still “mickey mouse”, too much sifting and sorting at FTDNA,and I never use 23’s tools, but my home away from home is still Gedmatch’s Tier 1. Just my personal favorite, and I am actually doing all my segment matching there. I just take a snip of the segment and our numberical grouping, send it to each person, asking to see their tree……….

    At 6 am, this morning I sent out 11 emails and by 8 am, I had 3 responses.

    Roberta, thank you for presenting this.

  5. I have five circles. One of the circles has three people in it and we all spell her name differently (Cathryn Smyth, Catherine Smyth, Catherine Smith). Nevertheless, we have a circle. We all spell her husband’s name the same (Elisha McNew), but not circle for him.

  6. This is a direct message to Tim Sullivan and the rest of the Ancestry.com Senior Management Idiots.

    Tim, before you roll things out how about a real life beta test of 500 or so of your customers … and not just a handful of Genetic Genealogy bloggers like Roberta …. who I agree with her this time on the criticisms of Ancestry.com ?

    Tim, you are relying WAY TOO MUCH on a band of Algorithmic – White Paper writing PhDs to make key customer facing decisions.

    Tim, you say: “Less is More”. I say hogwash! What happened to all of my Starred Relations? In you new version you wiped out these “PROVEN STARS” even thou I and others have a Paper Genealogical trail proving relationship as well as triangulate DNA proof also.

    Tim, I will not be recommending your company to others … as it is a PIA and a non-customer friendly company.

    Tim, I would also recommend to Henry Louis Gates Jr that he advice the Finding Your Roots organization & PBS to no longer affiliate with your company for the above stated reasons and also your lack of progressive Genetic Genealogy movements for the African American community. You could be doing alot more for that Community and you are NOT NOW DOING IT.

  7. I hate the changes!

    I lost matches that are 10 cM and longer, and I know they are IBD matches because they match me and a full Acadian woman on a certain location.

    A lot of people with Acadian ancestry are related to each other,sharing 17th Century ancestors. Of course, there are going to be a lot of people matching each other in the same location.

    I lost matches that I know that are IBD because they are 10 cM or more AND they match other people in the same location.

    The people have Acadian ancestry

    Triangulation.

    They are not IBS.

    I have found that very small segments can be IBD. It doesn’t have to be 10 cM or more.

    There are actually ancestor chromosome clusters. I found some Acadian ones with my matches.

    They are not necessarily meaningless pile-ups.

    I am upset with losing paternal matches. They are people that are white and can give clues to my African American father’s paternal white ancestors. Many of these have Acadian family lines that fit with my father’s Louisianian roots.

    Because of these new algorithms, I am afraid that I am going to be missing out on other IBD matches that are mistakenly identified as false positives because of some pile-up.

    Another thing

    How am I going to figure out my African American father’s white DNA matches without a chromosome browser? I haven’t uncovered a single paternal white ancestor yet. I really would like to find who my Acadian ancestors were.

    I am convinced that AncestryDNA was created for only White New World Colonials with well-developed family trees.

    I haven’t seen anything to show me otherwise.

    • The cost of the test itself is $99. For circles, you also need to subscribe at some level to Ancestry. They have different levels. You’ll need to check their website for that info.

  8. Faith based genetic genealogy. I don’t like it. I also don’t like the fact that my DNA data is being used for this new circular logic tool, but without a subscription, I can’t see it. If I unlink my tree from my DNA results or link my DNA results to a 3 person tree, will that take me out of these darned circles I’m not allowed to see?

      • I’m not a big fan of the shaky leaf hints. Not sure I’d miss it. Lack of matching segment info leaves me making too many guesses. How are small town folks and endogamous groups supposed to use this? What is AncestryDNA really doing now, simply grouping shaky leaf hints by shared ancestor? That’s not a bad idea, but it’s a really modest change.

        Although it looks slick and sexy, the new system seems to be compounding guesses upon speculation. No real triangulation, right? Too many people are going to look at these circles and assume too much. What a shame.

        Not many people feel like they should have to transfer their results to FTDNA or GEDmatch. I’m tired of playing salesperson for FTDNA and GEDmatch. I’m about ready to give up on AncestryDNA.

        I don’t mean to sound dramatic, but AncestryDNA has left me gravely disappointed. What disappoints me most is that most customers don’t seem to understand what they’re missing and how weak the hints and circles really are.

        I’d be AncestryDNA’s biggest fan if they’d allow us to share our own matching segment info with each other. But it looks like they’re fully committed to the big guessing game.

  9. Funny, regarding “when” ancestry.com would implement a chromosome browser… I was going to suggest that you and a few other genetic genealogists do a “Finding your roots” with management of ancestry to show them how DNA actually works. Show examples of using x matches to elliminate sections of your tree, etc. If they saw it first hand, and how it affected their own trees, “maybe” they would think again. I know you are all busy, but you could share the efforts. I truly think that they probably don’t understand how it works….

  10. One of the problems with Ancestry is that it really only works if you are a US citizen. For whatever reason they have built a wall around the borders and created, for example Ancestry.ca for Canada. As an Ancestry.ca customer I cannot see the trees of Ancestry.com unless the owner invites me to share. Unlike FT-DNA and 23andme it is not possible for Canadians to take the ancestry DNA test. Because Acadians are mostly Canadians, and were all from “Nouvelle-France” initially, there will be problems in implementing the circles across this barrier. So if you are an American with Canadian roots, good luck…

    I wanted to post my 23 and me results to FT-DNA but could not, because they are V4. It’s too bad because I had MtDNA done at FT-DNA.

    I too think that GEDmatch is the best product available and it’s free…

  11. Excellent two posts of before and after Roberta. I know you hinted (very strongly) to them that a chromosome browser is at the top of the list and they denied the request to implement. Maybe a public shitstorm hitting them on all social media and other feedback channels like email would be helpful.

    If you think so too, you might want to add a part to your post for people to easily click on and tell Ancestry.com what they think should be implemented next 😉

  12. Excellent article, Roberta. I have no sharkey leaves, and only 12 fourth cousin matches. I may not even have any DNA matches , just tree matches.!, My autosomal results show my ethnicity is 91 % Ireland, and I have never had one shakes leaf. !, Brick walls all over the place.!,, I am mostly patient and don’t give up to easily after 30 years. ,
    Love you blogs, keep up the good work.

  13. I saw the “new and improved” thing today. Same old. I tried to click on some of the pictures and was directed to a page offering to sell me the information for about $50. No thanks.

    • I wonder if you have an outdated browser. On my computer at work, if I clicked on a match, I would get a message that Ancestry wanted me to subscribe (even though I’m already a member). At home it worked just fine on my newer computer. Since Ancestry began the upgrades, the DNA page won’t come up at all on my work computer.

  14. Well this change didn’t work out very well for me. I went from 24 shakey leaves to 6. One of the leaves they removed was the one that solved my toughest brick wall. I was able follow the lead that her match provided and verify it by having other members of that branch take the Ancestry test. If the original AncestryDNA results would have been this strictly curtailed, I would have never solved this mystery. No circles either.

    • They made a big mess and now they’re claiming to have cleaned it up. Hopefully they didn’t throw out a lot of babies with the bathwater. How could we ever know?

      And now they’ve added a new gimmicky circle tool. Yay!

      To borrow a phrase from Mike Stivic, I don’t know if I’d go so far as to say that this new product is “new and improved” but the previous iteration was clearly “old and lousy.”

  15. Just wish the “Circles” were based on a common DNA segment, and not based on commonality in trees. ie all who match a segment on Ch1. Like a triangulation group with ease of communication.

    What is the point on giving us info we already have? I guess it is just a shiny new toy (window dressing/) for the LOW-information genetic genealogist.

  16. 152 pages of matches > now 17
    7561 matches > now 832
    7 4th cousins 95% confidence > now 9 4th cousins very high confidence
    226 5th-8th cousins moderate confidence > now 27 high confidence
    971 5th-8th cousins low confidence > now 367 good confidence
    6357 5th-8th cousins very low confidence > now 429 moderate confidence
    0 leaves > now 0 leaves
    zero circles
    still zero confirmed matches

  17. I am glad that ancestry has “improved” their matching technology. I spent time scratching my head as it seemed as if the found out my maternal line originated in south carolina and they proceeded to match me with every person in the database that had ties to south carolina. While my family was big it was not that big. I was also pleased to see family that I knew were family move up on priority on my matching list. However I still have matches that I am scratching my head wondering where they came from. For now as an african american I am accepting the fact because of our history dna is going to bring up ties that were lost. I wish someone would start doing work in this area to help us along but the hints as to how these leads could pan out wet the appetite for more.

  18. Great article, Roberta. I was curious what happened to M.C., listed on your previous post with an incredible 43,000+ people on his or her tree. Being new to genealogy, but not DNA testing, I’ve found that Ancestry is so filled with well, I have say, garbage, I was hoping the new methodology would help in eliminating the tendency of some members to just throw every name they find on their trees without sufficient proof. I’ve even seen the same “owner” with several variations of the same tree. One thing I’d love to see is a group of pedigrees that have been vetted and certified by accredited genealogists, trees that I can be more comfortable in accepting as reasonably reliable. If the new “circles” are full of junk and everyone else is, as I see so frequently, just copying names from one tree to another, how can we rely on the information? Through DNA analysis, of course, but that’s the point of your superb articles. Thanks.

  19. I found a rather interesting problem in a DNA Circle. One member of the circle does not have this particular ancestor as part of her pedigree tree. The kit is administered by another user, would Ancestry use another tree to put her in the circle? This same match dropped off my match list, but still matches my mother. However, when I compared her kit to mine in Gedmatch, this looks like a legitimate match. I zeroed in on this match b/c I noticed the change in “circles”. I have to say that losing a match of this size diminishes my confidence in the new and improved matching system.

    Chr Start Location End Location Centimorgans (cM) SNPs
    3 95726403 119605506 17.7 4820
    3 120502771 128345011 7.3 1932
    6 148878 3518221 10.6 1188
    Largest segment = 17.7 cM
    Total of segments > 7 cM = 35.5 cM
    Estimated number of generations to MRCA = 4.3

  20. I’ve viewed my new results pages at Ancestry. After reading your article (and thank you very much for it) and reading Ancestry’s explanation of ‘circles’ I still have a question. I have a private tree on Ancestry. On my home DNA page their are a group of pictures which are circles (I guess) but then there is a statement which indicates I have no known circle matches at this time. Then I am advised to make my tree public. Does this mean I do have circle matches but to see them I need to make my tree ‘public’ ?

    • I don’t know. You might try it and see. You can always make your tree private again after you find out. Keep in mind that that change may take 3-4 hours to take effect as that is how often they update their database.

      • Yes, now I have ‘circles’ connections after making my tree ‘public’. Alas, they are cousins and trees which have been very well documented and also DNA proven…..certainly not much help. Unfortunately, their trees don’t go back any farther than mine or my cousins’ do. I will now go back and make my tree ‘private ‘.

        Here’s a little bit of information which some may not know. If you have a match on your Ancestry DNA and the tree is private before you write the person do surname search of all your DNA matches. Use all the surnames you may be interested in and voila even ‘private’ tree matches will cough up the name 🙂 . Of course, this may not really be the way you match them but at least it’s a start especially if they don’t answer your emails :-(.

        I still have 4500 matches (don’t know how many I had before) and only lost two shaking leaves.

  21. I’m pleased with the results of the change. I went from 18,863 matches down to 5,061 and from from 186 shaky leaves to 119. I had several new matches with shaky leaves that I didn’t have before.

    I have 19 DNA circles. Most are husband/wife pairs, but not all.

    Roberta, I don’t have a circle for Fairwick Claxton/Clarkson or Agnes Muncy or James Lee Claxton/Clarkson. In looking at your Fairwick circle posted in the article, I see that we DID have one match in common, D. J.. He was one of my DNA matches, but he is no longer in my match list. I do however still have a DNA match with Bduvall100 for James Lee Claxton/Clarkson. In my tree I have Fairwick’s name spelled as Clarkson so I’m trying a little experiment and changed it to Claxton. I’ll have to wait and see if anything changes.

  22. This is a fabulous report, as always, Roberta. It sounds great, but when are ancestry going to stop being US centric and let some of us in other countries participate! I’m in Australia and have raised this with them several times. We used to have a DNA tab, now even that’s gone from the home screen!

  23. Two people do not make a circle. I have three DNA & tree matches but each one is a single person. My matches went from ~7,000 to 600 and several shakey leaves disappeared (and they should have never been there in the first place). My problem is like yours – no trees out on the site and no response to my note.

  24. I still feel ancestrydna is playing cat and mouse. They are providing us with total shared cMs. What we need is longest segment/block.

  25. I am very interested in the “pileups” and ways to identify them. I did the Family Finder at FTDNA so can’t participate in Ancestry’s efforts to weed them out. In your matches, did you notice one at Chromosome 19? Among the matches my sister and I have – that are not shared by either parent’s kits – are 102 matches on Chromosome 19 between 19,167,147 and 35,366,988. At 20,190,111-34,687,643 alone we have 55. I am betting this is a pileup.

      • Through Gedmatch I had already discovered a “pile-up” area on Chromosome 15, between about 25-29 covering about 10cm, which at the time I called for myself a “hotspot” and realized that there was something wrong with these matches being true ‘recent’ ancestors. If this is indeed individual, then the question becomes what are the causes of hotspots or pile-ups, at the time I thought that it might mean that ‘by chance’ my DNA through that area happened to ‘look’ very generic or there was a lot of missing no-calls. Because of Ancestry’s paternalistic arrogance so that everything is kept in a black box it seems to me that the genetic community will have to find out for themselves where the pile-ups occur and whether their individual or shared by many. After all it is a very good discovery they made and it is helpful to know that these areas exist, but then the next step is to find out where they are.

  26. Very disappointed with the upgrade. I’m trying to get us over the “brick wall” and the majority of my non AA cousins are gone. They gave us a lead to possible connections before 1870. Wonder if this is by design?

  27. Good Morning,

    I am on my way to the 50th wedding anniversary of a 1st cousin. I was looking at his father’s line and found that he is descended from Aladephia Speaks, born 1796 North Carolina. This line eventually lived in Allen County and Wells County, Indiana.

    Cindy Beverforden

  28. This is a good rundown of the new features. I was hopeful that this would make it easier to work with my matches in spite of the fact that AncestryDNA does not provide a chromosome browser. My mother was adopted and I have a few close matches that seem to be really good leads to her parental family. But on Saturday a predicted fourth cousin contacted me. I looked at her family tree and found no apparent connection. Then learned she was on gedmatch, so logged in and she was not included in my matches, so did a one-to-one analysis and the largest segment of cMs was 2.3 with only a total of 8.4 cM combined in small segments, there was a 3249 SNP count on one. I thought that there should be at least 7.0 cMs in one block? Then another predicted fourth cousin who I know shares one block of 14 cMs (total of 48 cMs) on FTDNA was lost in the Ancestry update. This makes me question the quality of the matches again. Is there a logical explanation for this? I’ll continue to do my testing with FTNDA until AncestryDNA supplies a chromosome browser. I don’t like being in the dark.

      • Yes, if AncestryDNA would provide tools similar to FTDNA in addition to their present features, they would be supplying the best of all worlds. I’m glad that I tested with them since a few close matches have provided me with my mother’s parental family, which is great, but the majority of my testing will be done through FTDNA so that I’ll have better information to work with.

  29. It looks like Ancestry is still the Prime Player in the arena of dna testing as far as numbers go. In the last months, these are the numbers of my new matches at Gedmatch:

    Ancestrydna – 47
    23 & Me -30
    FTDNA -15

    Just random sampling, but gives an idea of who is doing most of the dna testing.

    Does anyone know if 23 and Me’s results can still be sent to Promethease to get a report???

    • Hi,

      your assumptions are wrong. What you are predicting in your statistically too small sample (that’s another problem but you wrote “random sample”) is the number of people from those three big DNA services companies that have moved their data to GedMatch. This number is correlated with:

      1) The ability and difficulty of downloading the raw data

      2) The need and motivation to download the data and upload it to GedMatch. Here is where I think your assumption goes wrong. 23andme and FTDNA customer have chromosome browser and other excellent tools to identify the chromosome & segment where people match them. They also both have a large pool of atDNA matches. Now Ancestry on the other hand has NO TOOL to identify chromosome & segments and is therefore 100% dependent on a third party site like GedMatch to identify where these people match. The DNA circles also seems to be prone to a lot of errors as posted by others (both false positive and false negative) hence GedMatch gives you all the details and you can easily sort the wrong ones out and even let the tool do the triangulation for you (although for that small monthly charge).

      3) The number of customer who did atDNA test

      I think that the main driver behind this statistic that you see is reason no 2, not reason no 3. There are other websites available that clearly see FTDNA and 23andme fight head-to-head on the number of customers tested but given that 23andme offers only atDNA tests I see them as no 1 in that regard.

      Let’s all be careful when we make assumptions on the business and size of these companies as a blog like Roberta’s has a lot of influence. Not sure if our comments show up in Google results but you can see then how big wrong perceptions are becoming “truth” just because people find them on one of the best DNA genealogy blogs and through Google!

      Just my two cents 😉

  30. it’s weird that my mom dna’s wasn’t put into any dna circle even though there’s a few matches where she shares the same common ancestor with.

  31. Pingback: Sarah Hickerson (c1752-?), Lost Ancestor Found, 52 Ancestors #48 | DNAeXplained – Genetic Genealogy

  32. Pingback: Chromosome Browser War | DNAeXplained – Genetic Genealogy

  33. Now, before I share this with you, I want to be very, VERY clear – just because we share DNA and a common genealogy line does NOT MEAN that we are genetically connected via this genealogy path. However, having said that, it’s a very good hint and a wonderful place to start…………………………………………………How can we not be connected if we share DNA and a common genealogy? Someone out there…………Please help me understand this.

    • Let’s say you share two genealogy lines with this person, or three – separate lines. You could be related to them genetically from any or all of them – or from an unknown line. So just because the pedigree matches doesn’t mean that is where the DNA is from.

  34. Pingback: The Long Finn Logjam: A Rebellion against Recombination | IowaDNAProject

  35. Roberta, one of my researcher cousins found a really bad trend with DNA Circles. A person had circles leading to tree information that is wrong! Now people will begin saying “My line is proven by DNA.” when Ancestry is not “proving” anything more than expanding the amount of bad information out there!

  36. Pingback: 2014 Top Genetic Genealogy Happenings – A Baker’s Dozen +1 | DNAeXplained – Genetic Genealogy

  37. Pingback: How Phasing Works and Determining IBD Versus IBS Matches | DNAeXplained – Genetic Genealogy

  38. Pingback: Ancestry Gave Me A New DNA Ancestor – And It’s Wrong | DNAeXplained – Genetic Genealogy

  39. Pingback: Testing Ancestry’s Amazing “New Ancestor” DNA Claim | DNAeXplained – Genetic Genealogy

  40. Pingback: Dissecting AncestryDNA Circles and New Ancestors | DNAeXplained – Genetic Genealogy

  41. Pingback: What is a Population Bottleneck? | DNAeXplained – Genetic Genealogy

  42. Pingback: The 1 Million Mark and Effective Matches | DNAeXplained – Genetic Genealogy

  43. Pingback: Autosomal Matchmaking Vendor Comparison | DNAeXplained – Genetic Genealogy

Leave a Reply