The 1 Million Mark and Effective Matches

one million

Last week, Ancestry.com announced the millionth customer in their autosomal data base.  On January 18th, 23andMe did the same.  I don’t have exact numbers from Family Tree DNA, but they can’t be terribly far behind.  So, let’s look at the effectiveness of these matches at the roughly 1 million mark between the various vendors.

comparison chart

Black bold highlights the vendor’s positive aspects and red bold notes the drawbacks and places where each vendor could stand improvement.  I’ve underlined the two red issues I feel are the most serious.

*1 – Both 23andMe and Ancestry provide communications with others whom you match through internal message systems.  However, you have to request permission at 23andMe with anyone you match to communicate with them, and then additionally to share their DNA.  The 23andMe the 1404 number is how many people I match and the 162 number is the number of people that have accepted communications from me.  Not all of those 162 are sharing DNA.

*2 – At 23andMe, this would be the number of people sharing DNA results with me.  Ancestry has no tools that allow comparison of DNA segments.  At Family Tree DNA this would be all of my matches.

*3 – 23andMe cuts your matches off at 1000 unless you are communicating with your matches or you have an outstanding “introduction sent” request.  Of the 1404 people I match, 138 are sharing genomes, 24 have accepted communications but have not shared genomes, and 12 have declined.  The balance of my 1404 are either those to whom I’ve requested an introduction and they haven’t replied at all or some that I haven’t gotten around to inviting yet.  Ironically, my last of 1404 matches (in percentage of shared DNA order) is my known cousin who would have been purged had we not been sharing genomes.  You don’t have to send introductory invitations to those you match at either Family Tree DNA nor Ancestry and neither of those companies have an arbitrary cutoff, although Ancestry.com did a massive match purge when they implemented phasing.

*4 – At 23andMe, I can request to communicate with all 1404 people I match.  Of those, 162 have agreed to communicate or share genomes.  I can only communicate with those 162 people.  That doesn’t compare very well to either 1040 nor 5481 – and it shows how much genealogical benefit I’ve derived from 23andMe as compared to both Ancestry and Family Tree DNA.

*5 – At Ancestry, a minimum level subscription is required at $49 per year to see matching trees.  Not all participants have trees uploaded, and many trees aren’t public, so are not available for tree matching.  Otherwise, all trees connected to DNA results are included in matching function.

*6 – At Family Tree DNA, testers are encouraged to upload GEDCOM files or create trees in their account, and matching surname hints are given, but no actual ancestor matching in trees is performed.  Each participant must look at the tree of their matches, if provided.

*7 – 23andMe no longer hosts family trees on their site.  They have entered into collaboration with subscription service, MyHeritage.  Family Tree DNA is the only one of the vendors who hosts their own trees and does not require an additional subscription for that service, or for tree matching.

*8 – I have fewer matches at Family Tree DNA now than I did in November of 2014 when I had 1875 matches.  I have submitted a query to Family Tree DNA and they assure me this match number is accurate.

Commentary

The disparity between the 23andMe and Ancestry match numbers, since both vendors have 1 million autosomal results in their data bases, is suggestive of how many matches may have been pared from my match list at 23andMe.

The number of effective matches that can be usefully utilized, and how they can be utilized, are quite a bit different than the total number of matches implies without further analysis.

Both Family Tree DNA and Ancestry have unique strong points that make them stand out as vendors.

23andMe, since I can only work with or communicate with about 10% of my matches, is the least useful, for me, for genealogy.  I found their health services, which 23andMe is no longer allowed to offer following a dust-up with the FDA, very beneficial.

The tree matches and DNA Circles at Ancestry are very useful, but the fact that Ancestry provides absolutely no tools such as a chromosome browser or the other comparison tools that both 23andMe and Family Tree DNA provide makes Ancestry’s tree matches terribly frustrating eye candy in the candy shop behind a hermetically sealed window we can’t get through.  Tree matches and Circles are suggestive of an ancestral connection, but without comparison and triangulation tools, your match to an individual could be through a different, potentially unknown, line, and you have no tools at Ancestry to confirm or deny.  People are left to assume that the tree matches and Circles are proof, and unfortunately, they do in droves.

Thankfully, Family Tree DNA accepts transfers from Ancestry, V3 chip transfers from 23andMe (not the V4 chip since Dec. 2013) and GedMatch accepts files from all 3 vendors.  Those are the only avenues to actually compare the DNA of those who tested at Ancestry to triangulate and prove ancestral matches.

The great news in all of this is that more than 1 million people have tested, and probably more than two million in total – although there is clearly some overlap between vendors.  With every person that tests and that we match in one place or another, it increases our odds as genealogists to confirm our genealogy or break through those pesky brick walls.

Footnote:  The prices for the tests are the same, at $99, unless a sale is taking place at one of the vendors.  Both 23andMe and Ancestry also sell the aggregated anonymized DNA data for other purposes.  Both 23andMe and Ancestry will request that you sign (digitally authorize by clicking a box) an informed consent agreement for your non-anonymized (or less anonymized) data to be utilized or sold as well.  Family Tree DNA is the only one of these three firms that does not sell your DNA data in any form.

35 thoughts on “The 1 Million Mark and Effective Matches

  1. Roberta,

    Because of the differences in pluses and minuses for the various companies, I have settled on a methodology that maximizes their benefits for me. I start at AncestryDNA where there are more trees and more matches. I use their search box (while viewing all DNA matches) and look for a surname and location I’m actively researching. Once I find a promising suspect, I contact them through Ancestry (or I try to find their email address by searching posts on the community bulletin boards) and talk them into copying their raw data to GEDMATCH. If they respond at all, I’ve been successful in doing that most of the time. Having a family in common (potentially) usually fires them up. Once their data is on GEDMATCH, I do a one-to-one to get the segments we share. Then do triangulations in ADSA (GEDMATCH version) to find triangulated groups for those segments. Then I go to FTDNA ADSA and look in the same regions for triangulated groups and look for names that appear in both groups to link them. Then I start contacting the members of those groups – tell them who I think the MRCA might be (or at least the family and location) and start comparing notes. In a sense I am turning the triangulation process on it’s head by trying to find a MRCA before actually doing the triangulation. I find that this (mostly) skirts the deficiencies of Ancestry. Although I would SURE love to be able to triangulate all those matches that I have on Ancestry!

    Other things I can do on Ancestry: I can send a match a list of people I match who also have the same surname and location (my search results) and ask them to see if they also match any of them. (A poor man’s in-common-with) Then I know other matches there I should contact and try to talk into moving to GEDMATCH. I also do searches for female line surnames of putative ancestries. For example, if I am following a male to male surname and want to “prove” that a couple is in that path, I look for matches to the wife’s family. If I can find 2 or 3 people in Ancestry who have the wife’s ancestors (but not the husband’s family) among my matches, it is a strong indication that the couple is in my direct ancestry. I did this for several successive couples up my path to a common ancestor who was proven by Y-DNA.

    Of course I have to be very careful about these approaches because, for more distant matches, poor confidence (IBS) matches and endogamy (being related in more than one way) can really screw this up. But I’ve found that if I can find 2 or 3 people who can find the same target family, it tends to bolster my suppositions.

    Anyway, I would die and go to heaven if I got segment information from Ancestry, but it’s amazing how useful Ancestry’s DNA results are even without it – simply because there are so many matches and so many public trees.

    • This is an excellent strategy. I just wish we didn’t have to do so much subdivided work. My biggest issue with Ancestry matches is getting them to upload to GedMatch. Their most typical response is to go silent. The one I most desperately needed wouldn’t.

      • I’ve been pretty lucky, but it helped that I made a little PDF file with each of the steps to get the data downloaded from Ancestry and uploaded to GEDMATCH described in lurid detail. Also, I tell them a bit about the advantages (to them) of doing it – like painting their admixture to find Amer Indian (LOL) and getting me to help them interpret their DNA results (double-LOL). And, of course, finding other people who match in the same way this person does to me so that we can all exchange genealogies and photos of that branch of the family!

      • Roberta – Pretty much the things I mentioned. That I will help them interpret their results, that they will get more matches including people who tested at the other companies, that having more people who match on the same segment will make it easier to find the MRCA we share, that we can identify particular pieces of DNA that came from a certain ancestor of theirs so that if more people match them at those locations later we’ll know how they are related, that they can paint their ethnicity and see whether they are Amer Indian or even Pigmy (smile – my brother agreed to test because he wanted to see if he had any Pigmy DNA), that we can determine if their parents share any DNA, etc. And also I usually tell them that I’ve been working on this line for 40 years and it is a brick wall for me and if they would just do this thing (copy their raw data to GEDMATCH) they can help me try to knock down the wall and I would be eternally grateful! People like to feel like they are helping you out. And they are! I’ve broken through two 40 year old brick walls with DNA.

  2. Roberta, Could you add Gedmatch to the table to show how many matches you have their and how many you can contact? I find a few at Gedmatch without an e-mail address.
    Jay

  3. Roberta,
    Nice post! However one thing you have not mentioned is that it depends where your ancestors are from as to how useful the various companies are. Since my Dad’s families came from Norway in the 1880s we have almost no useful matches at ancestry and hundreds at 23andme, most in Norway. Maybe Norwegians are just more inclined to accept shares.
    Also a major irritation at family tree DNA is not being able to compare your matches to each other for triangulation.
    Kitty

  4. Its great to have all those matches but personally don’t have much use for the 400 + I have at Ancestry.Com. For me if its below a 3rd cousin its so hard to chase and I am semi retired and still don’t want to spend the time on below 3rd cousin? Ancestry has had in the past certainly the best Family tree capability though they now seem to be screwing with it every week and not sure that’s really improving it its gotten slower and more UP and DOWN ?? FTDNA which I belong to also has a JOKE of a Family Tree its really pretty pitiful looks like something they hired elementary school kids to write it is slow beyond belief and disjointed like the pieces really don’t fit together ! So even with Ancestry.Com. issues and they certainly have issues their FAMILY-TREE far exceeds
    FTDNA (these guys have tons of work to do). But FTDNA have the tools for at least doing some analysis without me having to build a bunch of spreadsheets. Just my view.

  5. First question : I have two distant male relatives. “A” tested with FT-DNA Y-DNA test while “B ” chose Ancestry.com. I asked “B” what was his haplogroup, and his answer was “what is that?”. If I get “B” to upload on GEDMATCH, will his haplogroup appear? Can the Y-DNA of “A” data be uploaded on GEDMATCH?

    Second topic: at 23andMe, there is a workaround the 1000 mark. It is called “Countries of Ancestry”. In addition to listing the origin of the grandparents of your match, the exact location of the matching segments on each matching chromosome is listed. It can be uploaded to your computer as a .csv file and imported into a spreadsheet for hours of enjoyment. You have access to all the files of the people who share their genomes with you so you can see their matches as well. The cutoff is 5cM so beware of IBS, but it provides many more possibilities. Of course, nothing can be done about those who decided to remain anonymous.

  6. Another great article, Roberta.
    Slightly OT…
    At FTDNA I like to check my Family Finder matches by clicking on “Match Date” so I can see who is a new match since the last time I checked. But I’ve been having problems with FTDNA showing folks who were already past matches suddenly reappearing as new matches. When I asked them about this problem, one of their programmers said it was probably folks who changed their privacy permissions and that would cause them to look like new matches. But since then I’ve had a match who reappeared 5-6 times in one month. (Surely she’s not changing her permissions that often!) And just this week I had about a dozen old matches appear on the same day as new matches. As a retired programmer, this certainly does not give me a “warm and fuzzy feeling” about FTDNA’s software algorithms and practices. And the effectiveness of the “Match Date” sort is greatly diminished. Anyone else having this same new (old) matches problem?

  7. Thanks for this comparison. I am absolutely new to DNA genealogy, but on your previous recommendation, I and a possible second cousin have forwarded DNA samples off to FTDNA this week for autosomal testing, to see if we have a match, and of course, to see what else we can find.

    Do you have any posts that outline a newcomer’s guide to how to analyse FTDNA results please? I’ll need to be prepared for when my results come back in a month or so. We both live in Australia if that makes any difference.

  8. FTDNA does allow tree searching, in theory and only up to a point. In the box at the top left marked “Enter a name or place to begin searching” you enter a name. Matches will only be visible if you match the person, or if they have marked their trees as “public”. I have some very common surnames, so to avoid wasting both my time and irrelevant surname matches, I have set my very detailed tree to “private”. Generally other people tend to do the same, so in practice this is next to useless – but it is there.

  9. Thanks for another great post. I decided the “Circles” were a joke when I have “cousins” who have also uploaded to GedMatch. When we share 22 cMs and 20 of those are triangulated to a different ancestor, I am doubting that the 2 cMs have anything to do with the ancestor in the Circle. I honestly think with so many of their customers still on base pairs, the “scientists” at Ancestry have no clue about how to triangulate.

    I pretty much only use Ancestry to compare trees when I find a real match on FTDNA or GedMatch.

    I admire Don for doing so much work. I admit I have gotten burned out and am barely keeping up with my spreadsheet.

  10. Thanks for the great article and analysis. With Ancestry, have you used Chrome’s AncestryDNA Helper add-on extension which is free? It is a good analytical tool for matches.

    I think uploading into GEDMatch for free is the future for all 3 companies.

    I suggest people put their email address inside all their profiles & messages to improve the chances for communication.

    • I use Jeff’s Chrome Add-On mostly to find people who tested or uploaded to the other databases. I think using the Ancestor download to determine possible ancestors can be good if you have 2nd or 3rd cousin matches (I don’t) but can lead to jumping to some very wrong conclusions. Just because Daniel Boone’s father shows up a few dozen times, you are NOT therefor descended form Daniel Boone. It is just as likely (probably more likely) that most of those folks grafted their unrelated kin on to the tree.

      I find it morally reprehensible that a for-profit company, such as Ancestry, forces their customers to rely on a free service provided by some of the real heroes of genetic genealogy (GedMatch), rather than spend some of their profits providing customers with basic tools..

      • Marci Bowman, I agree with you 100% about Ancestry.com, but still use the Chrome Add-On since it is free to work the DNA matches they provide me which are clues. I am fortunate to have DNA from a mother, uncle, three 2nd cousins, and a 3rd cousin on Ancestry.com.

  11. Another great article, Roberta,

    As with you and others, I have autosomal DNA testing with FTDNA, 23andMe, and Ancestry.

    For me, FTDNA is by far and away the best company for autosomal testing because of the ease to contact matches and the tools to help evaluate my matches.

    Ancestry may have made the coveted million mark, but without the tools to evaluate the matches, it is almost meaningless when a match is made. It makes a difference, especially when a match is important to you. It is like telling you – YOU HAVE A MATCH! It is probably to one of these surnames that you have in common, but we don’t know which one it is. Take a look at their tree and pick the one that looks the best to you.

    It would be like a woman going on the Maury Povich Show to see if someone is her son’s father, then being told – We have a match – someone is the father. then opens the envelope to pull out the photos of 5 men. “Take a look at these guys, which one do think is the father of your son?” You look them over carefully, “He looks good, that must be the right one!”

    Looking at the potential possibilities assumes you have a tree to look at. It seems more people have locks on their trees or no tree attached to their results. I presume they get to see my tree, but I don’t get to see theirs.

    I get tired having to go through all these matches that are 1) private 2) have no tree attached, 3) have fewer than 5 people listed on their tree. I have contacted Ancestry to suggest a filters separate these out. I guess they didn’t like my idea.

    This problem of privacy seems to be carrying over to FTDNA. People that used to have trees have changed their privacy settings.

    I don’t even bother with 23andMe any more. It has just become too labor intensive and time-consuming working through their system. I will say the closest match that I have gotten from all three companies was on 23andMe. That person worked to track me down to make that match.

    To your point, Roberta, a company can boast they have made the one million mark, but if they don’t provide the tools to effectively allow the consumer to use those results, then the number is meaningless.

    To me, FTDNA provides the best overall service for autosomal testing, but there is still room for improvement across the board for all companies.

  12. Roberta,

    Concerning Ancestry and 23andMe sharing it’s customers info, I just came across this article from Bloomberg.

    http://www.bloomberg.com/news/articles/2015-07-21/google-s-calico-to-scour-ancestry-com-data-for-longevity-genes

    I am an AncestryDNA customer and had I realized when I started that they would be giving/selling my DNA & tree info I would not have used them in my pursuit. Thanks again for sharing your knowledge that helps us novices make sense of this great, but bewildering world of DNA and finding our roots.

    Anita

  13. Write a review of AncestryDNA

    Love it or loathe it, you know AncestryDNA could be improved. Let future customers know what to expect and give AncestryDNA your opinion by writing a review at Amazon.com. If you didn’t buy your test through Amazon, don’t worry. You can still write a review there:

    http://www.amazon.com/Ancestry-DNA-Genetic-Testing-Test/product-reviews/B00TRLVKW0/ref=cm_cr_dp_qt_hist_two?ie=UTF8&filterBy=addTwoStar&showViewpoints=0

  14. Hi Roberta –

    I love the blog.

    On the topic of effective matches, there may be a major curveball involving DNA Circles, but I haven’t seen much commentary about it.

    Ancestry says the reason they came up with DNA Circles is because of the way people inherit atDNA. They say:

    “Only 4 percent of DNA Circles (that have between 3 and 30 members) have three or more people that share the same matching segment. In the remaining 96 percent of DNA Circles, no more than two people in the Circle share any one particular segment. In other words, even in DNA Circles with 30 descendants, usually only two or three descendants will all inherit the same segment.”

    They go on to say that the chances that five 3rd cousins share the same segment is zero. ZERO!

    So the algorithm looks to see cousins who match – just not all on the same segment – and makes a circle out of them. Fine.

    But this begs the question about those six people that match me at 20-25cM, all on the same segment. Is Ancestry saying we don’t share a common ancestor? Could it be that some “pile-up” areas could be part IBS and/or a sign of shared geography instead of a common ancestor? I have to admit, it could help explain why my success rate with matches is 100% above 4th cousin and about 2% at 4th cousin and below. Credit to genetic genealogist Diahan Southard for originally asking this question.

    Thoughts? Thanks.

  15. Thanks Roberta – I sincerely hope you’re right.

    Re 23andMe – I’m at a 30% response rate (330/1100), but it’s work. I invited everyone – even the 11cM matches in the back. I invited the top 200 matches twice. Most importantly, I go in every few weeks, sort by date, and invite all my new matches. People are just more attentive when they first sign up. I also use my own custom greeting when inviting people, listing my surnames by geography.

    I’ve been hoping that one DNA company or the other would drop the price to the magical $49 per kit. I even convinced myself that 23andMe might get their first given their other sources of revenue. But they don’t even have sales….

    • I agree that people are much more likely to answer when their results are fresh. I also send a custom message. Some people have reported a better response rate since they are only selling tests for genealogy and not for health.

  16. Pingback: Autosomal DNA Testing 101 – Tips and Tricks for Contact Success | DNAeXplained – Genetic Genealogy

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