Pardon my French!

Flavour of the month seems to be revised Ancestry DNA test results, which seem to be more blunt, and to offer less insight into ethnicity before records began.

I have been thinking about a blog on this topic, but it seems to me that, unless we have data to show the stratification of the samples, we cannot know the skewness. In other words, when revised estimates show stronger or weaker association with various ethnicities is that simply because the overall sample has been “swamped” by the predominant ethnicities among those with the inclination and resources to submit a test, and to relate the results to hard facts about actual ancestors?  The great majority of matches thrown up by Ancestry have yet to attach a tree.

For example:-

• if my English, red-haired, blue-eyed, fair-skinned ancestors were from Scandinavia via Normandy (as opposed to direct from Denmark);
• and no one from Normandy is tested (I gather it is illegal in France);
• and other white English and American folk like me have tests in their droves;

is it not likely that the Scandinavian component would seem to diminish in proportion? Does this actually mean that Norman ancestry is in truth more likely than before such a shift in data? (Absence of evidence not being evidence of absence).  The light colouring genes must be from somewhere, but if Saxon / north Germanic, why the name Dennis or Dennies?

What happened to my other regions?

… Ancestry asks, and answers:

We’re all intrigued by unexpected regions that show up in our ethnicity estimate—even if they don’t fit what we know about our family’s past. Here are some reasons why some of those regions may not appear in your new estimate.

First, we have more data. We estimate your ethnicity by comparing your DNA to the DNA of people who are native to a region. We call these people a reference panel. We now have 13,000 more samples in our reference panel, which means our ability to estimate your ethnicity is even better.
Second, with more data, we have been able to narrow down and better define our regions.
Third, DNA analysis is complex, cutting-edge science. We have developed even more powerful mathematical algorithms that help improve the accuracy of your DNA results. It’s like having a more powerful antenna that lets us pick up a clearer signal from a radio station.

In my case Scandinavian is far from unexpected and does fit with what I know of my family’s past.  What is unexpected is that the revised analysis excludes this background entirely.

Ancestry now has 13,000 samples on top of the previous 3,000.  But that is still a tiny sample compared to populations, and, indeed, the exponential numbers of previous generations.

Having read, and seen on TV, The Vikings by Neil Oliver, I will discount Swedish Vikings as it appears they were eastward-looking.  The populations of Norway, Denmark and Normandy, the likely pathways of my light-coloured ancestors is about 15 million (Wikipedia 2017), though the Normandy DNA sample may not be significant.

The population of the more heavily Viking-affected countries of the British Isles (excluding Wales) is about 67 million.  Many emigrated from these places to America, especially in the earlier British colonies, to Canada, and to Australia and New Zealand (population about 29.4 million combined, though that includes first nation peoples and other immigrants).

Ancestry also asks:

Does this new Ancestry 2018 update mean that your previous results were wrong?

And answers:

Most likely not.

But wait!  The problem here is that the original range given for Scandinavian is 0-30%, so a complete absence of Scandinavian in the revision cannot be claimed to be wrong, well, not based on the raw results, anyway.  But another problem arises.  It is known that the Normans were Norsemen, from Scandinavia.  When the Normans invaded England, they did not wipe out the already settled, Anglicised, Christianised, Danes; they would eventually mingle and have children with them.  How, then, can genetic analysis of living people separate them, except by the very few who can trace ancestry back to the Norman invasions of England and Ireland?  And, of course, the fewer still who can trace their Ancestry back to the Viking invasions of England and Ireland  a few centuries before.

It has been suggested that ethnicity is not necessarily genetic, but a social construct, essentially how one identifies oneself.  I am English, so was my father, and his and so on.  However, at some point no one identified themselves as English.  Similarly, at one time there were no people identifying as American, even though there were many peoples living there with their own identities and proud histories and cultures.  Ethnicity for Roma Gypsies is entirely genetic.  Is it so unreasonable to wonder how my ancestors identified themselves?  It my be impossible to prove with any scientific precision, but, though I accept the vested interest of Ancestry and other companies offering DNA analysis, one must go on the information available from DNA profiling and with documentary evidence.

There are, and have been for a long time, people named Dennis all over England. My Dennis or Dennies ancestors had light colouring and lived at the edge of what had been the Danelaw. I find it hard to believe that this is not significant.