Today (6th June 2018) I received an update to the Ethnicity Estimates created from my DNA at Ancestry.com. I thought it was worth reviewing the updates to see what value, if any, they add to my genealogical research.
As a background I should mention that previous Ancestry estimates for British folk tend to include both Scandinavian and Western Europe ethnicities. It’s not hard to understand this problem, given our shared history with our mainland European neighbours. However this does sometimes create a perception problem “Do I have Viking ancestors ?” . The answer, for anyone with British ancestry is a resounding yes. Ancestry themselves addressed in this blog post from 2015 “The Viking in the Room“. I was hoping that this might get addressed in this update.
I’ve written about my own known regional British heritage in previous posts, however it’s probably easier to give a simple summary of my ancestors. Based on my traditional paper-trail research I believe I’m 50% North Yorkshire, 25% East Anglia, 12.5% Scots-Irish (but probably lowland Scots) and 12.5% Durham/Scottish Borders. For more visual folk (or aren’t too familiar with the British geography, I have a nice map. The black- pins are my great grandparents.
Ancestry Ethnicity Updates
This would appear to be quite a major update that Ancestry have made to their Ethnicity Estimate. The headline figure is that the reference population database has been increased from 3,000 individuals to over 16,000. More interestingly, Ancestry have enhanced the algorithm used to match people to their reference populations. Rather than matching the reference samples to single SNPs of your DNA (i.e.the four bases: cytosine [C], guanine [G], adenine [A] or thymine [T] that make up you DNA) they are now matching longer stretches of your DNA. Whilst I’m no DNA expert I can imagine that this would produce better results. The Ancestry website gives a short, and more eloquent, run-down of their improvements, as below:
With advancements in DNA science and more data, we’re able to divide the world into more regions. With more regions to work with, we can typically make a more nuanced estimate.
Larger Sample Sizes
We determine your ethnicity estimate by comparing your DNA to samples of DNA from people who have a long history in a region. As we get more samples, our picture of what DNA from a region or group “looks like” gets better. We’ve added more than 13,000 new samples to the original 3,000 in our reference database to give us our clearest picture yet for each region.
Improved Ways to Analyze Your Data
DNA is made up of strings of four different letters: A, C, G, and T. Our old algorithm looked at one letter at a time, and based on where that letter appeared in your DNA, it decided where that bit of DNA came from. Without getting too technical, our new algorithm reads longer stretches of your DNA at once, making it easier to identify regions of the world where you ancestor once roamed.
My Updated Results
Ancestry do a very good job of presenting the new results and showing you what’s changed. In my case this means:
I think this is an improvement. If it is typical for all testers with British heritage then it’s good to see the disappearance of the trace regions. The flip side is that the more precise Scandinavian ancestry doesn’t seem to add much. The majority of my ancestors come from parts of England that was once under the Danelaw, but that’s over a thousand years ago.
The other improvement I see is that Scottish ethnicity has been grouped with Irish, which seems to make sense given my rather basic knowledge of the history of both peoples. It also neatly fudges put my Ulster Plantation ancestors, who were probably Scots.
Speaking of fudges Ancestry has (if I remember correctly) re-drawn the maps of the regions, as you can see this allows quite a lot of overlap in the regions, although I do feel a little sorry for those folks with Orcadian and northern Scottish Highlands ancestry, who seem to not be put in any region. By the Way – I should add at this point that the map is “zoomable” so don’t worry if you can’t see your country’s name up in lights. If you zoom in then even the smallest country gets a name-check.
It’s always nice to see the improvement in Ethnicity Estimates, even though the science is still very much a work-in-progress. The flip-side is that I’m not really the target audience for such products. Those targets are based in North America where the combination of high disposable incomes, incomplete record-keeping and diverse ancestries provides a flourishing market for ethnicity products that may reveal unknown “heritage”.
For me, DNA is an important tool in my genealogical research, but that is through the cousin-matching algorithms that help validate my paper-based research. I’ll leave me Viking drinking horn in the corner for later !