DNA Discovery 2020
Blaine and Angie had been scheduled to be in Wellington on the 27th and 28th March as part of a wider Tour within New Zealand. The Tour was organised by NZ-based Genies Michelle Patient[3], Fiona Brooker and Paul Alpe. Michelle and Fiona were quick to reorganise the Tour as a series of virtual visits when it was clear that travel from the USA to NZ was going to be a problem. It was really great that Blaine and Angie agreed to hold the lectures live online in lieu of the planned ‘in person’ visit.
And so it was that, last Friday and Saturday, Blaine and Angie came into my study along with Michelle, Fiona and Paul and lots of other people I had not met before.
In a broad sense, a lot of what was talked about I was already familiar with. I have been working on my husband and my family tree for about 25 years and working with DNA-related information for about 7 years. However, what these lectures did do for me was to open up a new way to think about some things and introduce me to some intricacies of working with DNA that I hadn’t really thought too much about before.
There is always more to learn no matter how long you have been researching. It was a very worthwhile and enjoyable two days.
There was a lot said and discussed over the course of two days. This blog talks about just one of the highlights for me. There were many more. Perhaps I will write about them later? This one stands out in particular because it has prompted me into a new line of inquiry.
Rethinking Ethnicity Estimates
I have, on occasion, posted in various groups or forums that I wish ‘ethnicity estimates’ would be called ‘bio-geographical estimates’ since that it what they are. There is an obvious problem with renaming them so … it’s very clunky and just doesn’t roll off the tongue! Consistent with that, Blaine, while identifying them as ‘bio-geographical estimates’, said he would be referring to them in his talk as ‘ethnicity estimates’. Until someone comes up with a better option, it’s just simpler that way. I shall do the same here.
The so-called 'ethnicity estimates' are not something I have taken much interest in the past. In fact I would say that, for the most part, I have gone as far as ignoring them. From time to time, I have had a quick look e.g., when they get updated, but I have never taken them particularly seriously; my preference being to focus on the DNA match information, which I have worked extensively with, in conjunction with the paper trail. My mistake it seems!!
A key message from Blaine was that there is something to learn from the 'ethnicity estimates' and we should explore them. However, there is a proviso. Blaine urged that, in working with our estimates we need to bear in mind that the:
- 'Ethnicity estimates' are fundamentally limited by the reference panels and algorithms used. These vary across companies. Consequently, the estimates you get are only as good as the reference panels and the process of analysis used, will vary from company to company and will improve over time as reference panels get larger and analysis improves.
- 'Ethnicity estimates' can only detect the DNA you have – that is, we need to remember that our genetic tree is a subset of our family tree and that, because segments from specific ancestors are passed down randomly, they can be lost quickly. Somewhere about 6-8 generations back, or so, ancestors fall off our genetic tree.
To date, I have had a quick preliminary look only. ‘Preliminary’ is the operative word here – I spent about an hour just to get a sense of what specifically I want to investigate - what would my research question be? What information do I have/could I get to help me with this?
Firstly, a look at my 'ethnicity estimates' across the testing companies - 23andMe, Ancestry, FamilyTreeDNA, MyHeritage and LivingDNA. There is a bit of juggling to do given that they all use different labels for their estimates, and FamilyTreeDNA was quite hard to line up. My ethnicity estimates look something like this:
 |
| As at: 3 April 2020 |
With Blaine’s words ringing in my ears – “look for consistency, averages, patterns”, I noted that Great Britain/Western Europe with a bit of Ireland is reasonably consistent both across companies and with my paper trail so far. Of the rest, it is the Scandanavian I find most intriguing. On the face of it, there isn’t much of it and only 2 of the above companies have detected it. The average amount across the 5 companies [ 1.4% plus 4% = 5.4% divided by 5] leaves me with 1.08%.
To date, I have no paper trail evidence of any Scandanavian connection whatsoever but the absence of evidence does not mean it is not there. I do know that I have some Scandanavian matches but that may mean that member(s) of one of my British Ancestor(s)’s family married a Scandanavian and had descendents who are my Scandanavian matches. It does not necessarily mean that I have a Scandanavian ancestor.
MyHeritage shows no Scandanavian at all in its ethnicity estimates but a quick look at my match list on MyHeritage shows that I have 73 matches from Sweden (closest shares 23cM across 2 segments), 53 from Norway (closest 44.1cM/2), 45 from Denmark (closest 22.7/3), 35 from Finland (closest 26.9 cM/4) and 1 from Iceland (12cM/2). These are worth investigating further, especially the 44.1cM match.
Looking very quickly across the shared matches of the top matches in each of these countries and looking at the occasional European names that came up amongst the many Scandanavian ones, I do have a hunch where this may be in my tree. It is definitely paternal and most likely my Father’s Mother’s side.
Next, I looked at the ethnicity chromosome map on 23andMe to see where 23andMe thinks the Scandanavian DNA is. Looking at the speculative 50% view, Scandanavian (1.4% as included in the chart above) looks like this:
When I looked at the less speculative 70% view, the amount reduces to 1.0% and looks like this:
To date, I have no paper trail evidence of any Scandanavian connection whatsoever but the absence of evidence does not mean it is not there. I do know that I have some Scandanavian matches but that may mean that member(s) of one of my British Ancestor(s)’s family married a Scandanavian and had descendents who are my Scandanavian matches. It does not necessarily mean that I have a Scandanavian ancestor.
MyHeritage shows no Scandanavian at all in its ethnicity estimates but a quick look at my match list on MyHeritage shows that I have 73 matches from Sweden (closest shares 23cM across 2 segments), 53 from Norway (closest 44.1cM/2), 45 from Denmark (closest 22.7/3), 35 from Finland (closest 26.9 cM/4) and 1 from Iceland (12cM/2). These are worth investigating further, especially the 44.1cM match.
Looking very quickly across the shared matches of the top matches in each of these countries and looking at the occasional European names that came up amongst the many Scandanavian ones, I do have a hunch where this may be in my tree. It is definitely paternal and most likely my Father’s Mother’s side.
Next, I looked at the ethnicity chromosome map on 23andMe to see where 23andMe thinks the Scandanavian DNA is. Looking at the speculative 50% view, Scandanavian (1.4% as included in the chart above) looks like this:
 |
Scandanavian
– 23andMe – Speculative (50%) view of 1.4%
|
When I looked at the less speculative 70% view, the amount reduces to 1.0% and looks like this:
 |
Scandanavian - 23andMe - Less Speculative (70%) view of 1.0% |
Next, I looked at my chromosome mapping profiles on the DNA Painter website [4] to see if I have anything painted on to segment 18. On my main profile (which contains all segments with an identified common ancestor), I found that chromosome 18 is pretty bare apart from a small segment painted on the maternal side which I share with a known maternal relative.
So, what about the autocluster I recently downloaded from MyHeritage and uploaded to DNA Painter and added to my experimental profile? Do any of those clusters map onto chromosome 18? No, they don’t but a quick look down the DNA segment information I downloaded from MyHeritage tells me that I do have Scandanavian matches who share DNA with me on Chromosome 18. None of them show up in the autoclusters.
So, what do I know and what do I need to find out?
I have a couple of brick walls that this Scandanian connection may sit behind. I have a hunch as to roughly where in my tree it may fit. If I can gather a bit more information, I may be able to convert my hunch into a testable hypothesis. I have DNA matches in Scandanavia who may know more than me, some with reasonable amounts of DNA shared. I need to look across my Match lists with the other companies for other matches from Scandanavia and see what I can learn about them. I also need to look more closely at the Scandanavians in my MyHeritage match list, particular those I match on chromosome 18, and make contact with them. What might they know that I do not? Fortunately for me, most Scandanavians speak good English. As part of my investigation into the Scandanavian matches, I also need to make use of shared matching to consider how these matches may group together and fit together and consider what that may mean in terms of common ancestry. Lots of work to do …
So … do I think it is worth paying some attention to ethnicity estimates? Yes, it is, as long as you bear in mind their limitations as per Blaine’s advice outlined above and consider them as just one piece of evidence that needs to be weighed up against all other available evidence, including DNA-related information and paper trail. I may have answered differently before last Friday. Thank you, Blaine!
So, what about the autocluster I recently downloaded from MyHeritage and uploaded to DNA Painter and added to my experimental profile? Do any of those clusters map onto chromosome 18? No, they don’t but a quick look down the DNA segment information I downloaded from MyHeritage tells me that I do have Scandanavian matches who share DNA with me on Chromosome 18. None of them show up in the autoclusters.
What Do I Know and What Do I Need to Find Out?
This was a quick review – as I said above, about 60 minutes worth just looking at information I already have ‘at my fingertips'. Clearly, I have more work to do if I want to find out whether I have a Scandanavian ancestor or not and/or where my Scandanavian matches fit in my tree.So, what do I know and what do I need to find out?
I have a couple of brick walls that this Scandanian connection may sit behind. I have a hunch as to roughly where in my tree it may fit. If I can gather a bit more information, I may be able to convert my hunch into a testable hypothesis. I have DNA matches in Scandanavia who may know more than me, some with reasonable amounts of DNA shared. I need to look across my Match lists with the other companies for other matches from Scandanavia and see what I can learn about them. I also need to look more closely at the Scandanavians in my MyHeritage match list, particular those I match on chromosome 18, and make contact with them. What might they know that I do not? Fortunately for me, most Scandanavians speak good English. As part of my investigation into the Scandanavian matches, I also need to make use of shared matching to consider how these matches may group together and fit together and consider what that may mean in terms of common ancestry. Lots of work to do …
So … do I think it is worth paying some attention to ethnicity estimates? Yes, it is, as long as you bear in mind their limitations as per Blaine’s advice outlined above and consider them as just one piece of evidence that needs to be weighed up against all other available evidence, including DNA-related information and paper trail. I may have answered differently before last Friday. Thank you, Blaine!
Notes
[3] Yes
I do know that Michelle is Australian and I did carefully say 'NZ-based Genies' rather than 'NZ Genies'-
[4] DNA Painter [Website]
[4] DNA Painter [Website]
Comments
Post a Comment