Archives

My talk about GEDmatch and more about GEDCOMs

Jun6

Finally! The video of the GEDmatch basics talk I did on May 26 for Verogen is now available. Best to view it full screen at youtube in order to see the images well.

Click here for the slides for that talk

GEDmatch, a DNA tools site, was originally created to compare GEDCOMs, a function you can still use it for. A GEDCOM is a plain text file of your family tree formatted so that any genealogy program can understand it. Click here for the wikipedia entry explaining this in more detail.

In my talk I emphasized that it is best to upload a privatized GEDCOM with no more than 10 generations of ancestors then connect it to the DNA test for that person. This will help you use the relative matching tools. I promised in the presentation to explain how to create a GEDCOM. So here are a few of the many ways.

Continue reading →

My Recent Talk on BioAncestry

Feb2

The interpretation of the origins of your ancestors from your DNA is called different things by different companies: ancestry composition, admixture, or, incorrectly, ethnicity. The latter term is borrowed from anthropology and refers to a shared cultural heritage and does not necessarily include shared DNA although it often will. Click here for the definition of ethnicity online. Ann Turner and Debbie Kennett, two genetic genealogists I admire, both like the word bioancestry instead of ethnicity, so that is the word I will attempt to use from now on.

I did a recent talk for the Southern California Genealogy Society’s (SCGS) webinar series on why the predictions vary so much from company to company and whether they are at all accurate. The slides are here and the webinar, free to SCGS members, is archived here (log in to view it). Click here for the schedule of SCGS webinars.

I explained that each company has different reference populations and they were originally focused on Europeans. I suggested reading the article at the ISOGG wiki (click here). I discussed that our ancestors moved around more than you might realize so that bioancestry predictions are not accurate at a country level. Your admix can mainly be determined on a macro scale: the North, South, East, and West of each continent. Some populations were isolated and inbred and thus are easier to predict from the DNA.

Continue reading →

My new puppy’s DNA says she is mainly Heeler

Jan23

Did you know that your dog has 39 pairs of chromosomes? Many more than us humans. The pet DNA testing companies now have features similar to the people testing sites, relative matching, messaging, raw data download, health results, and even a chromosome browser and painting.

Rockie at about 6 weeks (photo by my friend Rochelle)

Last fall I was given a puppy, found by the side of the road near Shiprock, NM by friends of mine. Of course I had to test her DNA. This time I decided to use Embark (this link should give you a discount*). Last time I used Wisdom (click here for that post) which now claims similar features to Embark. The Embark test shows only 38 chromosomes, thus not the sex chromosome, the XY or XX pair. The test is an easy cheek swab and the results came back within a few weeks.

It is important to understand that we humans deliberately created dog breeds, originally to help us hunt or herd, but also for protection or lap pets. The Victorians developed most of the more recent breeds we know today. Wikipedia has a good article on dog breeds (click here) which says “In 2004, a study looked at the microsatellites of 414 purebred dogs representing 85 breeds. The study found that dog breeds were so genetically distinct that 99% of individual dogs could be correctly assigned to their breed based on their genotype…” That sounds like these tests are likely to be accurate as to breeds.

Embark Home Page for Rockie showiin one message (red arrow added by me)

When I logged into Embark after her results came in, I saw the image above.  I was surprised that she had no border collie. because that is how her black and white markings looked to me, but apparently the white is from her Great White Pyrenees ancestor. On that home page there are many areas to explore. First I checked her health results, of course, and there was nothing bad there, phew.

Next I looked at the Breeds page which had the same image as above plus a chromosome painting by breed of her DNA. I can see that she has heeler (australian cattle dog) on both sides but Pyrenees only on one side. Both are herding breeds that might have been useful in the South West for cattle ranchers and sheepherders.

Continue reading →

The Westbye Mystery: Finding a Y Tester

Dec7

When I first started playing with clustering of my Ancestry DNA matches long ago, I found a group of presumed relatives descended from one Thorkhild Westbye of South Imjelt farm in Skougar, Vestfold. These people were not known to be my relatives, so something in my documented tree or their tree was likely incorrect. It so happens that my great great grandad Jørgen Wold, father of my great grandmother Maren Wold Lee, was the foreman on that farm.

My Ancestry DNA clusters from long ago – Added naming is mine

I blogged about the mystery of our DNA matches to the Westby(e) family a while back (click here) and my theory that Jorgen’s father was a Westbye, not the father of record. One of my action items was to find a male line Westbye who would test his Y, since that chromosome passes almost unchanged from father to son (click here for more about the Y) and can be used for deeper ancestry and paternal lines. Autosomal DNA is accurate for close matches but cannot distinguish in this case: between a half third and a half fourth cousin.

If the Y DNA were to match my Wold male line cousin Mike, who has tested, this would help confirm that Jorgen’s father is the Westbye and he and Thorkhild are half brothers. Looking at their eyes in the picture below, they could easily be closely related. If the Y does not match, then likely Torkhild Westbye was my great grandmother Maren’s father, not Jørgen. This would mean that our small matches to Jorgen’s mother’s side could be from some other further back ancestor. Although a failure to have a Y match could also indicate some other break in the line, so I would need more testers …

After much searching of family trees on Ancestry I was unable to find a male Westbye in America to test. However the tree at GENI provided me with a number of Norwegian relatives, one of whom has a great tree at MyHeritage. Thus I was able to contact him.
Continue reading →

New DNA Match List Features at MyHeritage

Nov23

MyHeritage has added labels (colored dots) and favorites (stars) to the DNA matches lists. These are extremely similar to the ones at AncestryDNA. One advantage at MyHeritage is that when you select multiple colored dots to display, you are shown all the matches marked with either one, whereas Ancestry only shows the matches who have both. Another advantage is that MyHeritage gives you 30 colors as opposed to the 24 at Ancestry.

On a DNA match, the left are icons for the new features, labels and favorites, above the new location of the notes icon (red arrow my addition). Clicking the square for a label slides in a panel on the right as above.

The downside of the MyHeritage implementation is that you can only see and edit these colored labels on the DNA match lists, not on the actual match page or its in common with list. According to the blog post that My Heritage wrote on how to use this feature (click here), those pages will have the labels in the future. Also when you export your match list from MyHeritage there is no indication of those labels in the resulting CSV.

So how might you use this new feature? First of all, for myself, I use the favorites star for matches I want to come back to later. However when working on an unknown parentage case, I use the star for just the paternal side which is helpful for various automated tools.

If you have already assigned colored dots on Ancestry, my advice is to use the same colors on your MyHeritage labels for the same groups. Personally I have assigned a color to each great-grandparent line, except my Bavarian line which has very few testers and no matches that I can confirm other than the one 2nd cousin that I convinced to test.

Then I have a few fifth grandparents who seem to have many tested descendants that match us, so they get colors too. Next I assigned colors for as yet undetermined matches from the same localities that my ancestors are from, those include labels for Norway unknown, Germany unknown, and Ashkenazi unknown. Click here for my discussion of how I use the very similar the Ancestry labels.

Continue reading →