99.9% European, 0.1% East African? (facial, family tree, grandmother, races)

[alcatel]

bump
for what it's worth, i subsequently tested with AncestryDNA, which showed 100%European and 0%African. Not sure wht conclusions to draw. Seems like an inexact science.

[Mimidae]

Someone mentioned that east Africa can include the Middle East. There were Moorish pirate invasions of Ireland. I also just read a theory that some of the earliest settlers of Ireland may have come from Africa (before mostly being wiped out by subsequent settlers).

If the earlier test was correct, it's such a miniscule amount that I'd think that the ancestor predates slavery in America. .01% - that would have to be so, so many generations back.

Anyway, thanks for the update!

[AppalachianGumbo]

Originally Posted by alcatel

bump
for what it's worth, i subsequently tested with AncestryDNA, which showed 100%European and 0%African. Not sure wht conclusions to draw. Seems like an inexact science.

99% European on one test and 100% European on another is pretty exact you are entirely European in ancestry. 0.1 East African, it's statistical noise. You need to understand these test are on an algorithm and is purely a statistical exercise. If you were 3% East African, that's something to bring up at the family dinner table.

[BSL1534]

Even 3% could be noise. These are estimates based upon comparing the tested DNA with the DNA of people in other regions. While I think most of the big companies do well in a very general sense---that is, people with 100% European ancestry are not coming back as 25% East Asian, or whatever, the ethnicity breakdowns are still estimates, and hardly useful when the estimates attempt too much precision. The differences between what we would say are easily distinguished ethnic groups are not so clearly different at the level of biology.

Small percentages---say 1-5% are likely meaningless, but might indicate a distant ancestor from the region. Those less than 1% estimates...probably should be ignored in the overwhelmingly majority of cases.

[AppalachianGumbo]

Originally Posted by BSL1534

Even 3% could be noise. These are estimates based upon comparing the tested DNA with the DNA of people in other regions. While I think most of the big companies do well in a very general sense---that is, people with 100% European ancestry are not coming back as 25% East Asian, or whatever, the ethnicity breakdowns are still estimates, and hardly useful when the estimates attempt too much precision. The differences between what we would say are easily distinguished ethnic groups are not so clearly different at the level of biology.

Small percentages---say 1-5% are likely meaningless, but might indicate a distant ancestor from the region. Those less than 1% estimates...probably should be ignored in the overwhelmingly majority of cases.

Continental (would be Europe) levels are fairly accurate.Sub-regional (to Europe; Ireland, British, French etc..) can be a tough call.

Larger companies such as 23andMe (only because I'm familiar with their methodology) 1% is considered DNA wise, above the noise range. 23andMe has confidence levels. Conservative, Standard and Speculative. Now, if you have 1% East Asian in Conservative, 2% in Standard and 3% in Speculative chances are those assignments are correct.. If you see 1% East Asian only in Speculative, one needs to be careful. However even at 50% confidence (in 23andMe's Speculative mode), the only next closest match for that genomic sequence was East Asian. Anything else would go into an "unassigned" category if no assignment to a Continental region could be made at all. So it still may be correct. Not that you wanted to know all of that, but I threw it in there.

2% is pretty indicative of probable ancestral roots. 25% East Asian in Europeans? Most have none with 23andMe. If they do, most don't even reach 1%. If they do, maybe .01 EA and usually Doug McDonald's test would identify this as noise.
If someone has 3% Sub Saharan African with 23andMe, it's pretty accurate especially if someone is from the American south or knows they have a African ancestor. Sub-Saharan African alleles is one of the most unique and identifiable it's rarely a wrong call. However, you cannot judge "ancestral distance" based on percentages, it doesn't work that way. I'll throw in for good measure, you cannot judge physical phenotype by autosmal testing either. These test use 90% of the DNA regions that are non-genetic regions. If you have a grandparent who is SSA or East Asian your racial fraction may be 25% SSA/East Asian does not mean you are going to be 25% on an autosomal admixture test. Percentages can also very widely between siblings. I've seen siblings from the same parents range from 10% Native American to 22%. The Native American from the maternal side.

To say someone's results would be "meaningless" is a bit presumptuous. I think in someone 99% Euro with no known ties to East Africans is statistical noise, I can be wrong myself. Maybe that do have an East African ancestor. Who knows. However, that small with no ancestry known in your genealogy, I wouldn't think about it too hard.
5 years ago and beyond that, 15% was in the noise range this was due to the algorithms used at that time, samples and the type of testing SNP's are pretty standard these days vs. STR's. Accuracy has made leaps and bounds with the top companies. Though less than 1% can be noise, it is not always. It does not mean the result is meaningless. I have seen results with people .05% Native American verify their documented linage. Tim Janzen who tested many people in his family has smaller percentages of Native American. Would you tell him his results were "meaningless?" I would not. He comes from a documented line of Wyandot Indians and this further confirms his biological connection to them. This is why people test. On the other hand, how do we tell people who have 0.5 Native American is meaningless if it can be accurate for someone else? I believe the entire picture needs to be looked at on an individual basis. Regional levels are pretty accurate, sub-regional can be static. For those who don't understand that, regional is East Asian with sub-regional being Chinese, Japanese, Korean.

This article (Tim Janzen) was before the 23andMe's update so he is using the older and now defunct Ancestry Painting. I believe with Ancestry Composition he was 0.5 Native American, going from memory.
Exploring Potential Native American Ancestry Using 23andMe

With 23andMe their East Asian and Native has a very high accuracy and recall rate. When it's called, even at a lower confidence
it's the next closet call that fits the model. For example, I'll post the African and Native American/East Asian. Pretty high precision. As I say, people should do their genealogy to verify if at so possible.

East Asian & Native American 0.99 0.99
Native American 0.99 0.86
East Asian 0.97 0.97
Japanese 0.98 0.92
Korean 0.86 0.62
Yakut 0.96 0.78
Mongolian 0.89 0.53
Chinese 0.93 0.91
Southeast Asian 0.95 0.70

Sub-Saharan African 0.99 0.99
West African 0.97 0.96
East African 0.95 0.89
Central & South African 1.00 0.89

"As you can see, our precision numbers are very high across the board, mostly above 90%, in a few instances dipping down to and below 80%. That means that when they system assigns an ancestry to a piece of DNA, it is very likely to be accurate. You can also see that as you move up from the sub-regional level (e.g. Britain and Ireland) to the regional level (e.g. Northern Europe) to the continental level (e.g. Europe), the precision approaches 100%".
https://www.23andme.com/ancestry_composition_guide/

[England Dan]

I have also heard years back ( from memory,) that Ireland and the West of Britain was settled by North Africans, in small parts. There are a lot of Southern Irish, Southern Welsh, some West English that are very dark with black curly hair and brown wide eyes. ( not just Tom Jones!). Strange how recent DNA found wales in the North different from the South. Brits and Irish tend to be dark brown hair ( ok, blue eyed majority but a lot of Welsh are brown eyed and tan. Most Celts tan very well, not all are freckly), so these moors might have fitted in better than Scandanavia, Germany ect:Was it pre Christ Tin, sheep trading that brought people from the Med over.

[Snowball7]

The delineation of anthropological DNA is really questionable.
The results seem to be tailored to making money by selling test kits. I'd like to see
what their classification of "British/Irish" DNA really is. Same can be said for all.
Except for the most major classifications, I think they are just out to make money.
Pick up a copy of Carleton Coon's "The Races of Europe" and compare real racial
subsets with these modern nationalist terms.
There are non-physical feature distributions of DNA. For example you may have someone with 10%
of a DNA marker and another person with the same 10% but one will appear to somewhat represent
that group in his or her features, while the other will not. This doesn't mean the DNA
isn't there, it just means that the genetic application of it is manifest somewhere other
than in observable features due to dominant and regressive tendencies which exist in all
the DNA types.

[AppalachianGumbo]

Originally Posted by Snowball7

The delineation of anthropological DNA is really questionable.
The results seem to be tailored to making money by selling test kits. I'd like to see
what their classification of "British/Irish" DNA really is. Same can be said for all.
Except for the most major classifications, I think they are just out to make money.
Pick up a copy of Carleton Coon's "The Races of Europe" and compare real racial
subsets with these modern nationalist terms.
There are non-physical feature distributions of DNA. For example you may have someone with 10%
of a DNA marker and another person with the same 10% but one will appear to somewhat represent
that group in his or her features, while the other will not. This doesn't mean the DNA
isn't there, it just means that the genetic application of it is manifest somewhere other
than in observable features due to dominant and regressive tendencies which exist in all
the DNA types.

SNP's are used on non-genetic areas of the DNA. These are DNA sequences, alleles. To make a database of Irish for example, they can sample 300 people. 200 have a sequence of ABC while 50 have DEF and the remaining 50 have GHI. Since 200 people in the Irish database carry ABC as a more common marker, this will be dubbed an "Irish Marker" therefore concluding most people of Irish ancestry will cluster with that sample. The test are not ingenious and are based on statistical analysis.
There is no such thing as "British" DNA "East Asian" DNA, only alleles carried in populations found more common. They can be found in other groups but more common in British populations.

True, they are a fore-profit companies catering to people's interest as most products geared to people, especially the American public. However most do the best they can in ancestral prediction. Most of these test are geared to European ancestries. Native Americans, Middle East and the Pacific are very under sampled, whereas Europeans with 23andMe are 5k+ in a database the other populations are less than 200. They get people in the ballpark at continental levels.

[Snowball7]

Well, someday I may take one of these tests just to see if it's right or not,
because I already know exactly what my ethnic background is, to the percentile.
If you take a look at Coon's Plates, in detail, you'll see what I meant.
Sampling the current populations of these places isn't as reliable or valuable
an indicator as is assumed. Here is a free online copy of the 875 page book.
https://archive.org/details/racesofeurope031695mbp

[suzy_q2010]

Coon's theories are pretty much incompatible with what modern genetics tells us.