Archives

All My GEDmatch Articles

Apr7

GEDmatch is one of my favorite sets of genetic genealogy tools but it can be hard to get started with. I have done many presentations over the years attempting to help people master the basics there, most recently for Family Tree University (click here). Plus I wrote a GEDmatch Basics article for their magazine (click here) which explains all the column headings on the one-to-many report. [UPDATE 15 Apr 2024: Click here for the interview I did for the FTU podcast.] Impact-Site-Verification: d124adea-1323-4f60-a2fb-f1fdf846aa56

Over the past 12 years I have written many posts about GEDmatch, most of which are still useful. So I decided to make a new tag – GEDmatch tools – which brings up just my posts that explain a specific tool. Now I can see which tools I still need to write about!

Here is the list of my articles in reverse chronological order (click any title to go to that post):

GEDmatch: What’s New and my Basics Talk

Mar6

The people behind the current GEDmatch have been working hard to improve its usability and its appearance. Tom Osypian, the QIAGEN GEDmatch Product Manager, gave a talk at the recent Rootstech about what’s new there. I hope some of you went to that! I was sad not to make it this year, other than virtually.

Meanwhile on this Thursday March 7, I will be giving an updated version of my talk on GEDmatch Basics for Family Tree University.  This webinar is designed to help the new user understand how to use the terrific tools on that site.

Now to list the main new features.

  • Most every function now has a drop down list of your kit numbers
    (cut and paste still works)
  • You can import your family tree from the FamilySearch website if you have an account there
  • You can click over to the People who match both of two kits function from the one to many.
  • Emails are now partially concealed.
  • Use of computerized collecting will shut down your account.

Personally my favorite new feature is the click point called Match on each match in the one to many tool. That takes you to a pre-filled form for the person you did the one-to-many for to compare them to the person in the resulting list. This gives you a quick way to see which of your relatives on GEDmatch match the new match, and thus what family line they are likely from.

I also tried the import of our family tree from FamilySearch for my brother’s kit and it worked beautifully. One small problem was that my Norwegian grandmother has a foreign character in her middle name so she came over as ? Lee (just her surname). I have reported that problem.

Showing the first three characters of the email plus the provider seems a good compromise to increase privacy while allowing you to see when kits belong to the same person. Clicking on that partial email address takes you to the user lookup page for that user which gives you the pseudonym or name of the kit owner, their email address, and their GEDCOM numbers if there are any

As to computerized collecting, if you are a programmer, it may seem efficient to write yourself a script to scrape the information you want but computers are too fast and that bogs down the GEDmatch servers and thus affects everyone else’s response time.

Using AI for Genealogy by Steve Little

Feb26

One of the most unusual talks at the recent i4GG conference (videos coming soon) was the one about the use of AI for genealogy by Steve Little, the AI program director for the National Genealogical Society (NGS).  I learned that it was how you phrased your question that could lead to more accurate answers, e.g. “you are a professional genealogist … ” I found out that AI, particularly the paid versions, could extract text from documents, even handwritten ones and translate in context. Here is my favorite slide from that talk. Personally my first impression of ChatGPT had been that it was great at sounding good while making stuff up.

Slide from Steve Little’s talk, used by permission

Steve will be speaking at RootsTech at 8 am Thursday this week and will also be available at the NGS booth as per his post on FaceBook.

Amusingly, in my own talk about using bioancestry to solve unknown parentage cases, I had experimented with using AI generated images to illustrate a few of my points. For example, when I asked the deepAI image generator for a Hungarian violinist I got this image whose hands are imperfect, but it still adds pizzazz to the slide.

No sooner has my favorite DNA conference (i4GG) ended, than it is time to get ready for Rootstech! No I won’t be there in person this year, too much to do to prepare for our move to Connecticut. Hope everyone has a great time. I will attend virtually, so if you are logged in there, you can click here to see if you are related to me! As all my ancestors are fairly recent immigrants (earliest 1860s), I have only 434 relatives at Rootstech, the closest being a fifth cousin. Oh well.

What’s new at GEDmatch and i4GG

Jan25

Every year genetic genealogists gather in San Diego in February for the i4GG conference founded by CeCe Moore and Dr. Tim Janzen. This year is the tenth anniversary and I am honored to be one of the presenters again. Click here for the i4GG web site.

In past years I have talked about the new features at GEDmatch. In 2022, this was a particularly dense lecture. (click here for the slides), as there were so many new and enhanced features. Clustering was taking the community by storm and GEDmatch has two versions of that, one of which even includes tree building.

{UPDATE 9 Feb 2024] My talk for i4GG this year is about how to use Ethnicity to solve DNA cases. This year, I will talk about GEDmatch yet again, but as there are fewer new features, I will include some of the ways that site has helped me solve DNA puzzles.

One nice new feature is that the site now sends you an email when you have a new match. You can select both the frequency of those emails and the match size which triggers an email.

Whenever I get a new match on the One-to-Many, the first thing I want to see is who else they match among my relatives, so it has always puzzled me why they did not include a button to do that. Needless to say this was a feature I requested. Now it is finally here! There is a column called ICW tool which has the word Match which when clicked takes you to the function People who match both, or 1 of 2 kits with the kit numbers filled in, ready to use.

One-to-Many image showing new ICW tool

middle section of the free version of One-to-Many limited version

 

Come to my lecture to hear more!

Much Ado About DNA Hacking

Nov8

The recent panic about hacking at 23andme in the press seems overblown to me. What exactly would someone do with my DNA? There is nothing in there of any monetary value nor do I have health risks that need to be private. Perhaps knowing which celebrities are Jewish or Chinese might be of use to some bad actors. The fact is that those lists are for sale on the dark web. Click here for an interesting article about that.
We have all been advised to guard our online privacy but our DNA is not our social security number nor our credit card so I am not worried about this yet. The hackers were able to use login credentials that were leaked from other sites to access those people’s accounts at 23andme. Then they could see information about other 23andme users whose DNA matched the compromised account. The type of information exposed was ethnicity, other relatives, and family tree information, plus whatever you said about yourself. This does not seem worrisome to me. My actual DNA was not exposed and even if it were, it would take a very DNA savvy hacker to use it to create a fake relative of mine.

Most of the DNA sites are now forcing two factor authentication (2FA) on their users when they log in. That is where a text or an email is sent to you when you log in to be sure it is you. This should prevent “credential stuffing” hacks in the future. If you try to log into 23andme, you will also discover that you must now change your password there. If your relative is deceased and their email of record is no longer available you may be out of luck. Perhaps customer service can help you.

Here is the text of the recent email all my Jewish accounts received:

“Specifically, there was unauthorized access to one or more 23andMe accounts that were connected to you through DNA Relatives. As a result, the DNA Relatives profile information you provided in this feature was exposed to the threat actor.”

The moral of this story is not to use the same passwords on more than one site. Several of my favorite passwords were leaked in various hospital system breaches. Google is kind enough to tell me when I try to log in to a site with a compromised password. My recommendation is to use several passwords which you vary by including a 2 or 3 character indicator of the site name.  So for example add “23m” somewhere in your 23 and me password. Most of us have browsers which remember our passwords for us and if they forget, we can use the forgot password link or have the site text us a code. I keep a text file of my passwords with written descriptions of which password used rather than spelling them out. Naturally I use unique, different, and difficult passwords with 2 factor security on sites that access money.

Personally, I am not leaving 23andme although I did change my password there. I am sad that many of the features that I love, like the DNA comparison tools, are temporarily closed down. I look forward to their return once the breach has been understood and dealt with.