Monday, May 5, 2014

DNA News of the Week: MyOrigins vs AncestryDNA


Finally we got a look at the new myOrigins ethnicity charts, which replaces Population Finder at Family Tree DNA. Sorting through all this brings to mind talk show guests who purport to communicate with dead relatives. They ask audience members, for instance, do you know someone with an R in their name. Is it a male? Is his name Robert or Roger? Not too impressive. The DNA companies do tell us instead of asking about our ancestors origins. At the same time giving themselves plenty of wiggle room by not completely committing to anything specifically. You might be told some your ancestors were from Italy. The fine print says give or take a few thousand miles. Basically your ancestors were from Europe somewhere, but who knows exactly where? That's difficult to exactly pinpoint with DNA. Broad regions can be predicted reliably however.
Actually the old Population Finder results came close to committing to a specific area. My mother was said to be 18% Mozabite with an error rate of less than 1%. Her new results have completely backed away from that. She now has 10% Mediterranean and  8% Middle Eastern. Only 2% is now in the Morocco/ Algeria area. My Middle Eastern is only 4% now, and it was 13% with a high level of confidence. So the old Population Finder was quite inaccurate. We should have been given margins of error close to 50%.
Turning to our head to head competition between AncestryDNA and myOrigins; we find we need to translate the regional descriptions in order to compare. Western Europe seems to be covered by European Coastal Plain.

Round 1. European Coastal Plain at myOrigins and Western Europe at AncestryDNA. This is where we see the most stark contrast. Ancestry only gave me less than 1% for this area. My Origins gives me 50% for this region. Both could make some sense. I do have a substantial number of French Canadian and German matches, which could point to my inheriting more DNA than would be expected considering how far removed I am from these ancestors. Ancestry does say I could have plus 3% since they do present a margin of error. If you calculate shares of inheritance based on generations I could have around 3% of German and combined French? I could have inherited more than that, but I would not think it would be 50%. I would have to give that round to Ancestry.

Round 2. Our next round considers British Isles ancestry. Ancestry breaks that down between Britain and Ireland. Combine these two result I get 60% British Isles. MyOrigins gives me 29%. I would say 29% would be closer if you break my heritage down by generation. That round goes to MyOrigins. Ancestry gave themselves lots of wiggle room on this one with high margins of error however.

In Round 3 we compare Eastern European results. Here we have a tie with both results around the same percentage. AncestryDNA gives me 6% with a high margin of error, and myOrigins gives me 8%.

Round 4 Scandinavian. I am not aware of any Scandinavian in my family. I've heard that Ancestry gives unusually frequent results, and high percentages for Scandinavian. I will give them the benefit of the doubt and assume this relates to the Vikings in Scotland. This round goes to AncestryDNA.

Round 5 Finland/ Northwest Russia. May relate to Hungarian? This round goes to AncestryDNA. No comparable myOrigins result.

Round 6 European Jewish. I would give to AncestryDNA because I believe there should be a trace of European Jewish which myOrigins does not show. I would like to see myOrigins show more trace regions.

Round 7 Pacific Islander Melanesia. Don't know what to make of that? It's only 1% but is completely puzzling to me? myOrigins didn't show an equivalent. I'll just scratch this round.

Round 8 Middle East is another tie round with both results very close. AncestryDNA gives me 5% and myOrigins gives me 4%. The edge does go slightly to myOrigins because they gave me a more specific area. Half point extra myOrigins.

Round 9 Native American is also very close. AncestryDNA gives me 6% and myOrigins gives me 3%. I'll give this round to AncestryDNA because FTDNA pulled their old estimate of 5.2% Native American and their specific result of Central American Native American. The old estimate had a low margin of error and I don't know why they are backing down on that?

Round 10 Africa. This result leads me to believe that the myOrigins populations are becoming more accurate. I had no African with their previous Population Finder. My mtDNA turned out to be African, so it's confirmed we do have some African. Both AncestryDNA and myOrigins give the exact same result of 2%. Half a point to AncestryDNA for a specific location Senagal. It's a tie round otherwise.

Round 11 Mediterranean. Ancestry gives me 7% and myOrigins 3%. I tend to go with the higher results because my cousin's Y DNA of J2b pointed us in that direction. I would give this round to AncestryDNA.

Round 12 Iberian. This isn't an area well covered by the new myOrigins. Ancestry gives me 5%, which I believe should be higher since one of my Grandmother's, Graciela Del Castillo, would have had a high percentage of this regional DNA. This round goes to AncestryDNA.

Final results Ancestry gets 7 1/2 points and myOrigins gets 1 1/2 
Plus 3 tied Rounds. 

myOrigins is brand new and I'm sure changes will be made at some point which may reconcile some of my results. 

Below are some charts I made. The first compares my percentages between AncestryDNA and myOrigins. The second would be a chart I made showing my known ancestral breakdowns.

Comparison Between myOrigins and AncestryDNA

My knowledge of my Ancestry


MHD said...

I am so enjoying your DNA series, Annette! Please keep it up! I hope to see you soon.

CR Campbell said...

Look for Vikings in Ireland as well. Just saying as you didn't mention it, but that may be because you already know they were from western Ireland...shrug ...:) Wesley johnston has a pretty good map showing their encampments. Google "viking ireland map" to locate it... Reading Anatole Klyosov has helped me when looking into my deep ancestry. You can find some of his research scattered around the Net in PDF... very interesting. Not much better location wise, but he puts dates to locations and when combined with some history, you can make an educated guess...:) Good post!