PDA

View Full Version : Review on Your Genetic Genealogist



AntG
02-16-2017, 11:12 AM
I was very interested to read the review by CeCe Moore on her blog, Your Genetic Genealogist.
You can read the review here: http://www.yourgeneticgenealogist.com/2017/02/the-new-living-dna-test-review-of-my.html

In it, CeCe highlights the "significant overestimate" of DNA from the British Isles and that "It is immediately obvious that something is off with the Living DNA estimate, since my grandmother was of full Finnish ancestry and all three of the other companies accurately detect that (21% - 23%). Conversely, Living DNA only estimates 12.6% in their Europe East category for me, which includes Finnish DNA. (On a side note, I consider this a misnomer. Finland is generally not considered to be part of Eastern Europe.)".

I personally am still excited about Living DNA but I suspect more refinement is needed. I know they have built upon the POBI work but I wonder what their international sample base is. I am still puzzled why my own results show all British Isles plus 1.2% from Balochistan (I've not received anything outside Europe on the other big 3 testing companies). LivingDNA did correctly identify my largest region as South West Scotland and Northern Ireland.

I know we've only got the 'Standard' view at the moment but I'm struggling with the definition of the three views:
Cautious: (Coming Soon) We group similar populations together in order to provide the highest certainty of results possible.
Standard: We highlight the sources of your ancestry estimate that are most certain. Ancestry that cannot be attributed to one of these sources is shown as being 'unassigned'.
Complete: (Coming Soon) Our best estimate of your overall genetic makeup.

"highest certainty" vs "most certain" vs "best estimate". Huh? Is it just me?

MacUalraig
02-16-2017, 12:53 PM
I was very interested to read the review by CeCe Moore on her blog, Your Genetic Genealogist.
You can read the review here: http://www.yourgeneticgenealogist.com/2017/02/the-new-living-dna-test-review-of-my.html

In it, CeCe highlights the "significant overestimate" of DNA from the British Isles and that "It is immediately obvious that something is off with the Living DNA estimate, since my grandmother was of full Finnish ancestry and all three of the other companies accurately detect that (21% - 23%). Conversely, Living DNA only estimates 12.6% in their Europe East category for me, which includes Finnish DNA. (On a side note, I consider this a misnomer. Finland is generally not considered to be part of Eastern Europe.)".

I personally am still excited about Living DNA but I suspect more refinement is needed. I know they have built upon the POBI work but I wonder what their international sample base is. I am still puzzled why my own results show all British Isles plus 1.2% from Balochistan (I've not received anything outside Europe on the other big 3 testing companies). LivingDNA did correctly identify my largest region as South West Scotland and Northern Ireland.

I know we've only got the 'Standard' view at the moment but I'm struggling with the definition of the three views:
Cautious: (Coming Soon) We group similar populations together in order to provide the highest certainty of results possible.
Standard: We highlight the sources of your ancestry estimate that are most certain. Ancestry that cannot be attributed to one of these sources is shown as being 'unassigned'.
Complete: (Coming Soon) Our best estimate of your overall genetic makeup.

"highest certainty" vs "most certain" vs "best estimate". Huh? Is it just me?

Yes very interesting, quite critical but she does appear to have signed up to their Affiliate program.

Jessie
02-16-2017, 01:50 PM
It looks like it is inflating British ancestry in people. The problem is most likely a lack of enough other populations in their database. Anyway it is early days and this company is only in its infancy and is offering a more in-depth AC than any other dna company. It might take a while to get more accuracy.

MacUalraig
02-16-2017, 02:04 PM
I wonder if they concentrated too much on UK people during the testing phase. They must have known that lots of Americans with colonial ancestry would buy it and that they would have broader mixes - I think German is the largest ethnic group in the USA ahead of Irish? So a test that reports such as EA/SE Eng may not go down well, even if it pleases its UK audience.

ADW_1981
02-16-2017, 02:49 PM
I was very interested to read the review by CeCe Moore on her blog, Your Genetic Genealogist.
You can read the review here: http://www.yourgeneticgenealogist.com/2017/02/the-new-living-dna-test-review-of-my.html

In it, CeCe highlights the "significant overestimate" of DNA from the British Isles and that "It is immediately obvious that something is off with the Living DNA estimate, since my grandmother was of full Finnish ancestry and all three of the other companies accurately detect that (21% - 23%). Conversely, Living DNA only estimates 12.6% in their Europe East category for me, which includes Finnish DNA. (On a side note, I consider this a misnomer. Finland is generally not considered to be part of Eastern Europe.)".

I personally am still excited about Living DNA but I suspect more refinement is needed. I know they have built upon the POBI work but I wonder what their international sample base is. I am still puzzled why my own results show all British Isles plus 1.2% from Balochistan (I've not received anything outside Europe on the other big 3 testing companies). LivingDNA did correctly identify my largest region as South West Scotland and Northern Ireland.

I know we've only got the 'Standard' view at the moment but I'm struggling with the definition of the three views:
Cautious: (Coming Soon) We group similar populations together in order to provide the highest certainty of results possible.
Standard: We highlight the sources of your ancestry estimate that are most certain. Ancestry that cannot be attributed to one of these sources is shown as being 'unassigned'.
Complete: (Coming Soon) Our best estimate of your overall genetic makeup.

"highest certainty" vs "most certain" vs "best estimate". Huh? Is it just me?

I believe the lack of non-British samples is causing an over estimate of UK type ancestry. For someone like myself, I get nearly perfect analysis with LivingDNA because I am the 'model' tester, but someone like the blog poster with Finnish ancestry, it is not nearly as accurate. I noticed a Belgian was getting overestimates for UK regions because no substantial French or Belgian samples are used in the model.

It should be noted that Anglo-Canadians are not Americans by any stretch ;) We're about 150 years removed (at most, usually 100 or less) from Britain or Ireland, where as Americans are about 300-400 years. There is a very big difference here, and I can see how tracing ancestors to specific regions in UK will be far more difficult for them.

02-16-2017, 03:23 PM
I wonder if they concentrated too much on UK people during the testing phase. They must have known that lots of Americans with colonial ancestry would buy it and that they would have broader mixes - I think German is the largest ethnic group in the USA ahead of Irish? So a test that reports such as EA/SE Eng may not go down well, even if it pleases its UK audience.

Hi MacUalraig, Regarding Germans being the largest Ethnic grouping in the US, that is correct it is the largest "reported", followed by "Irish", but if you look at the statistics, the next largest, I believe is "Americans", i.e people reporting themselves as Americans, probably denoted deeper colonial ancestry, where no affinity is related to the motherlands, people in the Southern states for example, There people reporting as "Americans", allot of them are White, with mostly British & Irish surnames, so make of that what you will.

ADW_1981
02-16-2017, 03:34 PM
Hi MacUalraig, Regarding Germans being the largest Ethnic grouping in the US, that is correct it is the largest "reported", followed by "Irish", but if you look at the statistics, the next largest, I believe is "Americans", i.e people reporting themselves as Americans, probably denoted deeper colonial ancestry, where no affinity is related to the motherlands, people in the Southern states for example, There people reporting as "Americans", allot of them are White, with mostly British & Irish surnames, so make of that what you will.

Absolutely, I have been attacked for suggesting the same, but often Anglo descent people are attacked for identifying themselves as anything besides Canadian. We do not hyphenate our origin, but often everyone else does. I have suggested the same for Americans, I tend to think the majority of the settlement in the former "Thirteen Colonies" was English or British, and I would be curious how many of them self-reported census data as "American". Of course they would.

AJL
02-16-2017, 03:36 PM
I think German is the largest ethnic group in the USA ahead of Irish?

This is contentious.

Self-reporting of Americans' ethnicity likely underestimates English ancestry significantly, to a large degree because the founding of the country was itself based on a war against Britain. English Americans, who by most accounts formed the majority at any early time in the nation's history, simply became "Americans." There was a systematic pattern to repudiate things English, including Webster's deliberate choosing of what were then often rare spelling variants, to create more distinctions between the US and Britain.

I don't disagree that German and Irish are both extremely prominent contributors to the US gene pool but the settlement from these places was mostly from the 19th century on, and I am not certain this effect could have outstripped the massive founder effect of mainly English-descended people from the original 13 Colonies in the 17th century.

Finally, the bulk of the non-outlying CEU population sample (Caucasian Europeans from Utah), which might be regarded as a fairly good stand-in for colonial white Americans in general, tends to place somewhere around southern England, like Kent, which is where you'd expect someone who is mainly English, somewhat German, and a bit Irish, with some other bits and bobs, to place.

{EDIT}
sgdavies beat me to the punch as I was typing.

02-16-2017, 03:41 PM
This is contentious.

Self-reporting of Americans' ethnicity likely underestimates English ancestry significantly, to a large degree because the founding of the country was itself based on a war against Britain. English Americans, who by most accounts formed the majority at any early time in the nation's history, simply became "Americans." There was a systematic pattern to repudiate things English, including Webster's deliberate choosing of what were then often rare spelling variants, to create more distinctions between the US and Britain.

I don't disagree that German and Irish are both extremely prominent contributors to the US gene pool but the settlement from these places was mostly from the 19th century on, and I am not certain this effect could have outstripped the massive founder effect of mainly English-descended people from the original 13 Colonies in the 17th century.

Finally, the bulk of the non-outlying CEU population sample (Caucasian Europeans from Utah), which might be regarded as a fairly good stand-in for colonial white Americans in general, tends to place somewhere around southern England, like Kent, which is where you'd expect someone who is mainly English, somewhat German, and a bit Irish, with some other bits and bobs, to place.

{EDIT}
sgdavies beat me to the punch as I was typing.

LOL, got to be fast on that keyboard :P but your correct.

MacUalraig
02-16-2017, 04:40 PM
The point is taken although I was really just trying to emphasise the importance of non-UK/Ire data for the testing companies. It does give me a chance to mention that I did a university dissertation on the peopling of America and the work drew upon a massive report based on analysis of surnames in the 1790 census - obviously from 1850 onwards there is specific census data to draw upon.

The report ref was
‘Report of the committee on linguistic and national stocks in the population of the US’,
American Historical Association Proceedings (1931), pp107-452

and it is available online if you've not read it before. There is lengthy discussion about the Germans :-)

and my humble dissertation which was about Irish v Scots Irish Kennedys is

http://www.academia.edu/3630733/Peopling_of_the_United_States_with_special_referen ce_to_the_Kennedys

MitchellSince1893
02-16-2017, 07:54 PM
Here is the estimated ethnic makeup of European Americans in 1790

Country, #of immigrants up to 1790, Population in 1790.
England* 230,000 1,900,000 (59.7%)
Ulster Scot-Irish* 135,000 320,000 (10.0%)
German 103,000 280,000 (8.8%)
Scotland* 48,500 160,000 (5.0%)
Ireland 8,000 200,000 (6.3%)
Netherlands 6,000 100,000 (3.1%)
Wales* 4,000 120,000 (3.8%)
France 3,000 80,000 (2.5%)
Sweden and Other 500 20,000 (.6%)

So the British and Irish made up 84.9% of the White population in 1790.

This source says in 1699, 90% of white population was of English origin. http://www.rootsweb.ancestry.com/~nycoloni/ethmigclem.html

wombatofthenorth
02-17-2017, 10:57 PM
"LivingDNA estimates that 81.7% of my DNA comes from Great Britain and Ireland. That is a significant overestimate."

Hmm I wonder if it is not as I feared. Those happy they finally got a much higher direct British & Irish score and thinking it meant because the test was better and could actually pick out modern English much better from Continental might be incorrect and it might just be that they tuned it to make tricky calls bias more towards B & I since they are UK based and have tons of UK testers. So it may not really be any better at all, just biased differently (and perhaps so far over to arguably make it a worse overall test)?

Of course we still need to see more results from people with lots of European but not a whole tons and tons of UK ancestry.

sktibo
02-18-2017, 01:34 AM
Hmmm.. I just read over the review again. While I agree with her that it overestimates British and Irish, there are a few things that make me think she may be exaggerating the amount of the overestimate.
1. "At 23andMe, I am 24.8% British/Irish and 22% Finnish. At AncestryDNA, I am 0% Great Britain, 10% Irish and 21% Finland/Northwest Russia.At Family Tree DNA, I am 27% British Isles and 23% Finland and Northern Siberian." Doesn't include the Western Europe under Ancestry. Pretty much every British tester I've seen the results for got a chunk of Western Europe, and it often is larger than Britain. IIRC, Ancestry is planning on merging their Great Britain and Western European clusters because they have so much trouble telling them apart. Looked up her full results and found: Scandinavia, 42%, Finish/N Russia, 21%, Europe West, 20%, Ireland 10%.
2. In the family tree picture she posted, there's a bunch of surnames which clearly look to be of English/British origin even though they aren't highlighted in red. I'm looking at Purdy, Cole, and Moore. Not saying these are British for sure, I don't have access to her tree, but it indicates that there's probably some admixture in these people if they carry these names.
A quick look at a few of her posts and it's often and bout Finnish ancestry, so it looks like her bias is towards this region. I get the impression she's a bit disappointed that it estimated her Finnish as low, and her British so high.
For the record,
She's not wrong, it does overestimate British. I'm just trying to be critical of her review.

wombatofthenorth
02-18-2017, 03:22 AM
Hmmm.. I just read over the review again. While I agree with her that it overestimates British and Irish, there are a few things that make me think she may be exaggerating the amount of the overestimate.
1. "At 23andMe, I am 24.8% British/Irish and 22% Finnish. At AncestryDNA, I am 0% Great Britain, 10% Irish and 21% Finland/Northwest Russia.At Family Tree DNA, I am 27% British Isles and 23% Finland and Northern Siberian." Doesn't include the Western Europe under Ancestry. Pretty much every British tester I've seen the results for got a chunk of Western Europe, and it often is larger than Britain. IIRC, Ancestry is planning on merging their Great Britain and Western European clusters because they have so much trouble telling them apart. Looked up her full results and found: Scandinavia, 42%, Finish/N Russia, 21%, Europe West, 20%, Ireland 10%.
2. In the family tree picture she posted, there's a bunch of surnames which clearly look to be of English/British origin even though they aren't highlighted in red. I'm looking at Purdy, Cole, and Moore. Not saying these are British for sure, I don't have access to her tree, but it indicates that there's probably some admixture in these people if they carry these names.
A quick look at a few of her posts and it's often and bout Finnish ancestry, so it looks like her bias is towards this region. I get the impression she's a bit disappointed that it estimated her Finnish as low, and her British so high.
For the record,
She's not wrong, it does overestimate British. I'm just trying to be critical of her review.

1. OTOH, Broadly NW Europe still is Broadly and it can just as easily be assumed to be say German. And if this skews people even farther to the UK (as it is tests already give Germans a decent chunk of B & I) it might actually make it even a bit trickier to sort out and maybe slightly even more initially misleading at first.

2. Yeah it did seem like maybe she underestimated her UK lines a bit on her tree, but I was imaging she must know something we don't, but perhaps not.

Anyway, hopefully a bunch of considerably German, French, Dutch, Swiss, Austrian, etc. test. And it will be interesting to see how the whole Baltic, Finland, EE works out for people. On the face of it having a group with Finland, ALL of Western Russian and the Ukraine seems a bit puzzling.

deadly77
02-18-2017, 06:57 AM
She also posted a link to her review on the ISOGG Facebook page and there was a bit more discussion with her in the comments on there. She confirms most of the colonial names as having English origin, but says the Moores are likely Palatine Germans or might be Irish. Her main point appears to be the underestimation of Finnish ancestry from her grandmother. She then said she had an extensive call with the CEO of Living DNA last night, and the company did not seem to have any issues with her post and acknowledged the weaknesses in their algorithms regarding Finnish, Scandinavian and German predictions. She then said she was very encouraged and hopeful regarding their future. She then said David Nicholson also wanted to communicate to everyone that this is a "Living DNA" test in that it will be improved and changing over time. They are very committed to fine-tuning their results and she was confident that they will do so. She found him to be very open to feedback and, even constructive criticism.

I think we're at very early days with the data from the first wave of testers. I'm interested to see how this gets refined in the future.

sktibo
02-18-2017, 09:34 AM
She also posted a link to her review on the ISOGG Facebook page and there was a bit more discussion with her in the comments on there. She confirms most of the colonial names as having English origin, but says the Moores are likely Palatine Germans or might be Irish. Her main point appears to be the underestimation of Finnish ancestry from her grandmother. She then said she had an extensive call with the CEO of Living DNA last night, and the company did not seem to have any issues with her post and acknowledged the weaknesses in their algorithms regarding Finnish, Scandinavian and German predictions. She then said she was very encouraged and hopeful regarding their future. She then said David Nicholson also wanted to communicate to everyone that this is a "Living DNA" test in that it will be improved and changing over time. They are very committed to fine-tuning their results and she was confident that they will do so. She found him to be very open to feedback and, even constructive criticism.

I think we're at very early days with the data from the first wave of testers. I'm interested to see how this gets refined in the future.

That is fantastic news. It seems like a rarity these days when people are open to and accepting of criticism. I remain hopeful for the future living DNA updates