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. Pingback: Autosomal DNA Matching Confidence Spectrum | DNAeXplained – Genetic Genealogy

  2. Pingback: DNAeXplain Archives – Introductory DNA | DNAeXplained – Genetic Genealogy

  3. Pingback: Ancestry’s New “Amount of Shared DNA” – What Does It Really Mean? | DNAeXplained – Genetic Genealogy

  4. Pingback: The Ancestry 200 | DNAeXplained – Genetic Genealogy

  5. Pingback: Concepts – How Your Autosomal DNA Identifies Your Ancestors | DNAeXplained – Genetic Genealogy

  6. Pingback: Concepts – Identical by…Descent, State, Population and Chance | DNAeXplained – Genetic Genealogy

  7. Pingback: Concepts – CentiMorgans, SNPs and Pickin’ Crab | DNAeXplained – Genetic Genealogy

  8. Pingback: 2016 Genetic Genealogy Retrospective | DNAeXplained – Genetic Genealogy

  9. Pingback: Concepts – Segment Size, Legitimate and False Matches | DNAeXplained – Genetic Genealogy

  10. Was reading your article to see a better explanation of how Circles are calculated, only to find that you are also a Wilkes County Vannoy. Small world!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s