Can there be profiling errors in the DNA results?

[Nuno Rodrigues]

Not an error in the DNA itself of course but how the profiling is done. I am Portuguese and only know my mother thus I was interested in finding out who the other half was, especially because she is very light skinned and I am your typical Iberian.

The results came back as 51% Italian, 29% Iberian and 20% Scottish, this was a huge surprise and my mother must have Scottish ancestors, can´t convince her to take a DNA test though. The Italian part is where things get weird, I did a search for relations and got about 800 results, all of them were people with Portuguese names and not a single Italian, neither in name nor ethnicity.

I was thinking of sending another sample to FamilyTreeDNA, my first was to Myheritage but to me it makes no sense whatsoever that I should be part Italian, they do have plenty of Italians in their database, just as many as Portuguese. I could understand a disparity in the results, but 800 vs 0 is to me an extremely weird result.

Since most of the names that came up in the results are Portuguese I can only assume my father is actually Portuguese.

I asked Myheritage and their reply made no sense, could be that many of the Portuguese in the list had Italian heritage too, then why did they all have typical Portuguese names without exceptions and not a single Italian on the list? Not even an Italian name.

Could their be errors in the methods of ethnicity profiling?

[John McCoy]

"Ethnic origins" from DNA is not an exact science, at least not at the level of distinguishing among different "ethnic" groups in Europe. Some would argue that it is more "art" than "science". At best, the results can be regarded as suggestions about your possible ancestry. One of the big problems, as you will probably understand, is that European history, in virtually every part of Europe, is full of migrations, wars, social upheavals, etc. that have mixed the "ethnic" groups.

Different vendors have developed their own methodology in hopes of making better predictions of "ethnic" ancestry. The statistical methods will be similar, but each vendor has attempted to develop a library of "reference samples" that are believed to represent in some way the hypothetical ethnic groups, with varying degrees of success. However, it is very difficult to demonstrate objectively that one method or one set of "reference groups" is better than another, for lack of any external standard.