FTDNA has rolled out the long-awaited update to their MyOrigins ethnicity estimator–and it includes Sephardic in the reference panel! I was shocked to see such a high percentage of Sephardic in my results, which had merely been suggested by GEDmatch before, but in a smaller amount. Disappointingly, I found some changes in other catagories that raised an eyebrow. I have to question whether this Sephardic component can be considered accurate.
I would love to hear your thoughts on whether the accuracy of FTDNA’s MyOrigins has improved with their new version. Weigh in on the comments section below!
I got excited when I saw the 11% Sephardic component. GEDmatch had suggested there was significant Sephardic ancestry in my Ashkenazi grandfather. This family line has y-DNA tested as being Mizrahi in origin, surprisingly. Haplogroup R-M124 very well could have entered Europe through Sephardic lands like Roman Italy and Spain, based on the historical evidence.
Then I saw they had me as 0% Western European–before I was 42%, and Ancestry has me at 48%. The entire category has disappeared in the updated results. This is an issue because I have lots of Western European ancestors in my tree, including some relatively recent immigration from Pomerania, Germany.
Furthermore, they clocked me in at almost 1/4 Southeastern European (Greece, Italy, Balkan). I have an extensive tree with no know ancestry from any of these regions. This is also clearly problematic.
However, some results appear to have genuinely improved with the new update.
My Scandinavian used to be grossly underestimated at 2%, and is now 17%. This is more in line with my tree, having recent immigration from Norway.
Eastern European went from 21% to 7%. This is probably also more accurate.
Additionally, the increased British Isles checks out compared to what I know about my tree. I have recent immigration from Lincolnshire, England on one branch. Prior to the update, FTDNA had me at 0% British Isles.
Because of the total disappearance of Western European when it should be close to 50%, and the surprise large amount of Southeastern European, I’m concerned about how reliable these results are. This makes me question the validity of the Sephardic result along with it. As they say, percentages must always be taken with a grain of salt. In this case, a large grain indeed.
Have you noticed any major changes to your results?
Elizabeth Onheiber 
As I said in my comment on your Facebook post, mine has improved! Finally picking up close to 100% British Isles like the paper trail on all my kits. Weird that the trace regions are so different though. The biggest shocker with the trace regions is that they’ve added 6% Jewish Diaspora to my mum’s test, which I suppose could be possible since I haven’t had any cousin matches on her paternal line but still a massive shock since it hasn’t come up with any other testing companies and doesn’t reflect the known trail.
LikeLiked by 1 person
Glad to hear your results were consistent with your research! Thanks for sharing
LikeLiked by 1 person
It’s a shame yours weren’t 😓 Hopefully with more refinements more people will be happy. Thank you for sharing your story! 😊
LikeLiked by 1 person
Have you updated this? I exported my data there (autosomal from my heritage) and it found nothing, on my heritage it found distant cousins that I know are related from 3-7 generations ago. Names I’d never heard of, that as soon as my brother saw it, he’s saying.. Oh yeah that’s paternal, grandfather’s mom’s maiden name, oh yeah that’s a third cousin on maternal line, they moved to Canada at the turn of the century.. Etc. I got a Sephardic shock too 9 %, 89 % Ashkenazi, however. Although I don’t know enough to interpret it on gedmatch. I uploaded to my family so I’m waiting on the origins. (& ancestry results not for a month at least). Be well!
LikeLiked by 1 person
Hi Teri, I’m sorry but I’m not sure what the question is. This post has not been updated, if that’s all you’re asking. Happy hunting!
LikeLiked by 1 person
Thanks for the quick response. I meant, you can retest (not physically test, but have the data reevaluated) the raw data, with the upgraded algorithms. And add them to new sites. Many people have reported after transmitting some of the changes were shocking, either by the population sampled, rejigging errors (like those Koreans who got Japanese because they tested Japanese who were Korean immigrants within generations, or for Jewish matches – tweaked the algorithm to account for endogamy. If it stays the same you might be able to narrow down which ancestor gave it to you because there’s so many new people who have tested since then.
LikeLike
My old results had 33% British Isles, and 38% Scandinavia. This made sense as I have grandparents from both Norway and Scotland. The new algorithm gave only 18% British Isles, 0% Scandinavia and added 78% West and Central European. It also added ‘noise’ from Central America and Oceania. My wife had all of her 15% Scandinavian removed (she is Swedish by surname), and was replaced with West/Central European. Her Iberian (Sephardic?) was replaced with European Jewish at 2%.
LikeLiked by 1 person
My previous results gave my mother Ashkenazi DNA for 10% (I had 8%) and now it disappears in mother’s results, and I have 2% now only 🙂 Funny. Unless, there was very small Ashkenazi in my father’s ancestry, and together with mother’s combined into my 2%.
PS: Thanks for this post, I referred to my as additional resource.
LikeLike
Glad you found it useful! I experienced the same thing with Eastern European decreasing and Southeastern European increasing.
LikeLike