PDA

View Full Version : 23andMe Ancestry Composition Results



Pages : [1] 2 3 4

Scarlet Ibis
12-07-2012, 05:11 AM
My mom's results:

[img]

DMXX
12-07-2012, 10:50 AM
I am entirely Middle-Eastern on mine (99.1%). Looks like this is only going to be informative for Europeans or those of continental mixes.

[Edit]: Note that the "Middle-East" here, for whatever reason, is Syria, Turkey and Iran. Not sure what this grouping was based on.

Little bit
12-07-2012, 12:03 PM
The most verifiable in my group is my Polish mother-in-law, all 4 GP traced as being Kashubian:
http://i953.photobucket.com/albums/ae20/LB234/th_Screenshot2012-12-07at52723AM.png (http://s953.photobucket.com/albums/ae20/LB234/?action=view&current=Screenshot2012-12-07at52723AM.png)

The Eastern European get's inherited by her son and our kids in a nice, intuitive way:
Her son: 51.7% E Euro
Our son: 30.7% E Euro
Our daughter: 26.2% E Euro

The rest of us with Irish, English, and German get more Nonspecific Northern European than anything. Here's my mom with an Irish grandparent, English grandparent, German grandparent, and American colonial (Euro) grandparent:
http://i953.photobucket.com/albums/ae20/LB234/th_Screenshot2012-12-07at54111AM.png (http://s953.photobucket.com/albums/ae20/LB234/?action=view&current=Screenshot2012-12-07at54111AM.png)

I saw someone note on a 23andme thread that French and German percentages are low and may get fixed, so we'll see. The most exciting part for me is that after almost 4 years of looking for the genetic fingerprint of my father through my DNA, I may have it:
http://i953.photobucket.com/albums/ae20/LB234/th_Screenshot2012-12-07at60044AM.png (http://s953.photobucket.com/albums/ae20/LB234/?action=view&current=Screenshot2012-12-07at60044AM.png)

I verified the Scandinavian from him because all my Scandinavian RF cousins come from him. The Native American was unknown before now: never showed up before and frankly at that level is probably untraceable, and maybe just noise?

DMXX
12-07-2012, 02:09 PM
After checking a few of the results it seems like the "Middle-Eastern" category here is basically equivalent to "West Asian" on ADMIXTURE runs.

The Arabian peninsular sharing contacts I have score as a mix of North African and Middle-Eastern, whereas Iranians tend to predominantly be Middle-Eastern.

Clinton P
12-07-2012, 03:10 PM
Here is my ‘Ancestry Composition’. Accurate, but boring. :\

Speculative
99.8% British and Irish
0.1% Finnish
0.1% Nonspecific Northern European

Standard
99.6% British and Irish
0.4% Nonspecific Northern European

Conservative
95.0% British and Irish
3.7% Nonspecific Northern European
1.2% Nonspecific European

Clinton P

Silesian
12-07-2012, 04:26 PM
Three generations; 23andMe Ancestry Composition Speculative estimate Results:

Silesia Poland-1st father
Eastern European-77.9%
French-German-4%
British-Irish-1.3%
Non specific Northern European-7.5%
Sardinian-0.1%
Nonspecific Southern European-2.5%
Nonspecific European-6.5%
Middle Eastern and North African-0.2%

Polish-German mix/2nd generation son
Eastern European-64.7%
French-German-9.7%
British-Irish-2.4%
Nonspecific Northern European-8.9%
Balkan-3.9%
Nonspecific Southern European-3.6%
Nonspecific European-6.8%


Polish-German-Irish-U.K./3rd generation grandson
French-German-17.6%%
British-Irish-16.7%
Scandinavian-4.5%
Nonspecific Northern European-23.7%
Eastern European-27.9%
Italian-0.7%
Balkan-0.2%
Nonspecific European-6.8%
Nonspecific Southern European-2.9%
Non specific European-5.4%
East Asian and Native American0.2%
Middle Eastern and North African-0.1%

Alanson
12-07-2012, 05:57 PM
Here is very interesting result of my ancestry composition but one thing is bothering me. Well I am mixture of Caucasian and Arabian. It seems that the Caucasian-Anatolian component went to the ME and North Africa. The North African score here is also interesting because it's 8%. Well my 4% European has been almost the same in all of the result and most often it turns out to be of NorthEastern Euro in the other tests. Though what is really bothering me is the missing 19.4% that seems to be unassigned. I am guessing it's Central Asian most likely

http://s7.postimage.org/sj8vf4hcb/Untitled56.png (http://postimage.org/)
upload pic (http://postimage.org/)

Jean M
12-08-2012, 02:09 PM
It looks like the new tool can identify British and Irish easily enough. My result is similar to Clinton's - just a fraction higher B & I (99.9% on speculative, 99.8% standard.)

AJL
12-08-2012, 07:52 PM
It looks like the new tool can identify British and Irish easily enough.

Yes and no.

My mother is mainly Isles ancestry by a fair margin, and ends up with 43.9% British + Irish, 3.0% French + German (which is slightly low but close), and a whopping 50-odd percent Nonspecific Northern European/Nonspecific European, in addition to 0.1% South Asian (some remote Romani ancestry is plausible but not documented). Her sister -- my aunt -- has 31.7% British + Irish, 0.1% Scandinavian (again plausible considering coastal English/Irish/Scottish ancestry), 0.5% Iberian (probably feeding into known French ancestry), and the rest Nonspecific Northern European/Nonspecific European.

Almost all that "nonspecified," by paper trail, is English, Scottish, Irish, or Welsh, with a small amount of Dutch.

My mother's contributions to my DNA show up as 25.4% British + Irish, 4.9% French + German (more accurate than my mother's assesment), and the rest Nonspecific Northern European. My paternal contributions end up primarily Ashkenazi with some Eastern European and fairly large Nonspecific European and Unassigned blocks, much of which probably represents Sephardic ancestry.

So I think it is safe to say that while a very high proportion of British + Irish is excellent evidence of such ancestry, a somewhat lower proportion of 30-50% with a large amount of Nonspecific Northern European does not preclude majority or perhaps even full Isles ancestry.

basque
12-08-2012, 09:13 PM
Standard:
99.6% British/Irish
0.2% Finnish
0.2% Nonspecific Northern European.

Speculative:
99.7% British/Irish
0.2% Finnish
0.1% Nonspecific Northern European.

Conservative:
96.7% British/Irish
3.0% Nonspecific Northern European
0.3% Nonspecific European.

basque (euro mutt) :rolleyes:

MJost
12-08-2012, 09:50 PM
Ancestry Composition tells you what percent of your DNA comes from each of 22 populations worldwide. The analysis includes DNA you received from all of your ancestors, on both sides of your family. The results reflect where your ancestors lived 500 years ago, before ocean-crossing ships and airplanes came on the scene.

99.5% European (95.5%)
Northern European (86.5%)
34.5% British and Irish
16.6% French and German
6.3% Scandinavian
29.2% Nonspecific Northern European

Southern European (5.2%)
0.3% Sardinian
0.2% Italian
4.8% Nonspecific Southern European

4.6% Eastern European (4.6%)
3.1% Nonspecific European (3.1%)

0.3% Sub-Saharan African

0.2% East Asian & Native American
0.2% Native American

< 0.1% Unassigned
100.0%
MJost

Tregowath
12-09-2012, 02:42 AM
So I think it is safe to say that while a very high proportion of British + Irish is excellent evidence of such ancestry, a somewhat lower proportion of 30-50% with a large amount of Nonspecific Northern European does not preclude majority or perhaps even full Isles ancestry.

My mother and her two sisters are 1/4 English (from Cornwall and Northamptonshire; 1850s immigration) and 3/4 colonial, most of which can be traced to early colonial New England and appears to be British based on documented genealogy, surnames, and presence in locations where the population was overwhelmingly of British origin. They only scored 31%, 34%, and 40% British Isles/Irish, respectively. They also scored between 13 and 23 percent French/German, of which there is only about 1-3% French (and no German) documented in the family tree (the French is at least six generations back from a New Netherlands Huguenot family). They each got between 37% and 42% Non-specific Northern European.

The French/German just seems to be high no matter how I look at it. My Aunt Harriet shows both branches of her chromosome 2 as entirely French/German. There just aren't enough places for French and Germans to hide in our family tree to account for that.

By comparison, I got 47% British/Irish and only 6% French/German. I have about 1/8 German ancestry on my father's side, so the 6% for me is fairly accurate.

...and they're still fiddling with the numbers, so the small NA segment my mother had has disappeared, but I still have a small NA segment that I received from my mother :)

thetick
12-09-2012, 02:55 AM
It looks like the new tool can identify British and Irish easily enough. My result is similar to Clinton's - just a fraction higher B & I (99.9% on speculative, 99.8% standard.)


I think it's more likely your sample was used in the population sample. If you answered your AF questionnaire then most like it was and then of course your DNA is a high match to your own DNA. With that said 23andme targeted their largest customer base ( European and Jewish ancestry) and this is exactly what we are seeing in this results.

I'm thrilled with my results and match my known ancestry very closely. My only concern is the German/French percentage appears smaller than expected for me and others..also the above poster thought it was quite high. I suspect there is an issue there since many at 23andme have mentioned the German / French percentages seemed off.

Boudicca
12-09-2012, 04:08 PM
Does anyone know if there is a noise threshold? My mother gets 0.1% SSA from one parent on chr 11 and 0.1% East Asian from the other on the same chr! They're not in the same place, but the East Asian is on the non specific North Euro half and the SSA is on the British/Irish half. Her mother was 100% British so if it is from her mother then I don't know how it got there! All her ancestors were from the East Midlands of England apart from a GGGGGrandmother who was Irish. My mother's father was unknown American with some ancestry from Hungary so I thought that might explain the non specific North Euro half with the East Asian.

Jean M
12-09-2012, 04:42 PM
I think it's more likely your sample was used in the population sample.

Yes I came to that conclusion after the reply on this thread from AJL. If it is any consolation to anyone, the lineages of my four grandparents do trace back to Britain as far as I can go. Admittedly that is not all that far on most lines - 18th-19th century. The surnames so far uncovered are all British except for one Norman, which is likely to be pretty standard. I have no known Welsh lines, which could be a problem for some, if others in the population sample are also short on Welsh.

thetick
12-09-2012, 05:37 PM
Does anyone know if there is a noise threshold? My mother gets 0.1% SSA from one parent on chr 11 and 0.1% East Asian from the other on the same chr!

.1% is the definition of noise. Keep in mind almost no one is anywhere close to 100% a component even the sample populations, so that .1% SSA and East Asian is much more likely not in your recent ancestry. Roughly speaking I would only take anything in the 1-2% seriously and even then there's a good chance of still being noise.

Alanson
12-09-2012, 06:35 PM
Does anyone know if there is a noise threshold? My mother gets 0.1% SSA from one parent on chr 11 and 0.1% East Asian from the other on the same chr! They're not in the same place, but the East Asian is on the non specific North Euro half and the SSA is on the British/Irish half. Her mother was 100% British so if it is from her mother then I don't know how it got there! All her ancestors were from the East Midlands of England apart from a GGGGGrandmother who was Irish. My mother's father was unknown American with some ancestry from Hungary so I thought that might explain the non specific North Euro half with the East Asian.

African slaves were present in England, it can also have come via Berber ancestry. Since the Berbers did raid as far as Ireland at one time to.

Jean M
12-09-2012, 09:50 PM
@ Alanson

It's amazing how colourful the history of my country becomes, viewed from Canada! :\

A few African slaves were brought to England in the 18th century by slave-owners in the West Indies, but it was established fairly soon that slavery was unacceptable on these shores. A very few West Indian slave-owners brought back to England their own children by part-Africans. In some cases they could have easily passed as white. I researched one case (to do with a house I was working on), where the son of an English slave owner and West Indian octoroon married in England and had descendants. So that's a teeny, teeny number in a population of many millions even then.

Barbary pirates raided these shores to take slaves. If you have evidence that they went in for rape as well as pillage, it's news to me.

In short, the chances of any individual Briton having even a tiny bit of SSA ancestry from the last few centuries is roughly equivalent to the chance of being struck by lightening. It is not impossible, but I wouldn't bother investigating the possibility if all you get from Ancestry Composition is noise level SSA.

Manakel
12-11-2012, 02:17 AM
278279

So I have .2% Sub-Saharan African on a few different websites. My 23andme results as well as Douglas McDonald show the same segment on my ancestry composition as being Sub-Saharan African. I'm American with full European ancestry. I don't think this is genetic noise however and i'm curious as to whether this sub-saharan dna is from the last few hundred years or from the original immigration out of Africa during the Neolithic Era. I am spanish, italian and Irish and both Spain and Italy are known to have a tiny fraction of african influence in their ancestry and i'm wondering if that's a possibility. Or possibly an African slave here in the U.S. since my Dad's family as been in the U.S. since the early 1800's. I understand that .2% has no real bearing at all genetically however it would be interesting to know if I have an African Ancestor within the last few hundred years. Has anyone with an ancestral background from a country that has no history of African slavery like that of Japan or Scandanavia had any tiny amounts of sub-saharan african markers? If that was the case than I would definitely agree that this is just genetic noise or ancient African influence from the Neolithic migration out of Africa.

Jean M
12-11-2012, 08:50 PM
whether this sub-saharan dna is from the last few hundred years or from the original immigration out of Africa during the Neolithic Era.

I'm not entirely sure what you have in mind. Our species spread out of Africa maybe 100,000 or 50,000 years ago (depending on which theory you favour. That was long before the Neolithic (which was when farming began - about 11,000 years ago in the Near East.) But there was another migration then I think.

Out-of Africa

The group of Homo sapiens who left Africa originally would of course be pretty much like their relatives who stayed in Africa genetically. The group who left would not encompass all the genetic variety that remained in Africa. They would be a sub-set of it. But more importantly all sorts of genetic changes have taken place since then in people all across the world. Sub-Saharan Africans today are not exactly like their remote ancestors of 100,000 years ago. Asians are not exactly like their remote ancestors who first arrived in Asia. Etc. Picture a bunch of mutations arising in one continent, which then get spread around in that continent over millennia in which there was very little movement between continents, so people mainly interbred within that continent. Now picture that repeated in each continent. So today there are on average continental differences between people. It is those continental differences that get labelled "race". It is those continental differences that 23andMe and other companies can pick up. So if your .2% Sub-Saharan African is not just noise, it won't have anything to do with the Out-of-Africa scenario.

Neolithic

As far as I can work out on available evidence, there was another migration out of Africa around the time of the Neolithic, when the Sahara was green. An African mtDNA L2 has been found at an early Neolithic site in the Near East. Then some of the people who had filtered north into the Near East and adopted farming seem to have spread it into North Africa and also contributed to its spread across the northern shores of the Mediterranean. This may have contributed some small part of the SSA that shows up at low level in Iberia.

Moorish dominion

One study calculated that all Southern Europeans have inherited 1%–3% African ancestry with an average mixture date of around 55 generations ago. That would fit the period of the conquest of much of Iberia by the Umayyad Caliphate. The Arabs began the slave trade which took so many sub-Saharan Africans into servitude far from home.

Jean M
12-14-2012, 12:00 AM
I think it's more likely your sample was used in the population sample.



Yes I came to that conclusion after the reply on this thread from AJL. If it is any consolation to anyone, the lineages of my four grandparents do trace back to Britain as far as I can go. Admittedly that is not all that far on most lines - 18th-19th century. The surnames so far uncovered are all British except for one Norman, which is likely to be pretty standard. I have no known Welsh lines, which could be a problem for some, if others in the population sample are also short on Welsh.

If you hover over the line for British and Irish, and click on the i for information symbol, you find that the sample included 843 23andMe customers and 89 public. So I think I can stop worrying that my particular ancestry (specifically lack of Welsh) might have biased the sample. There should be some Welsh in that lot. :)

History-of-Things
12-14-2012, 12:36 AM
Sorry for the re-post. Coming back from a hiatus I always get it wrong on which thread to post, somehow.

Aha. Well. My results:

Conservative Estimate:
98.9% European
.2% British and Irish
55.4% Nonspecific Northern European
.1% Iberian
<.1% Southern European
43.2% Nonspecific European
1.1% Unassigned

Standard Estimate:
100% European
9.6% British and Irish
72.8% Nonspecific Northern European
.2%Iberian
<.1% Nonspecific Southern European
17.4% Nonspecific European

Speculative Estimate:
100% European
37.4% British and Irish
18.8% French and German
1.6% Scandinavian
36.8% Nonspecific Northern European
.3% Iberian
.8% Nonspecific Southern European
4.3% Nonspecific European

It would appear that my genes have "left the Old-World building" and therefore are largely nonspecific. Still, it beats the heck out of the Genographic Project's attempt at such things. It is interesting that the continental elements are almost equally split by my mother's and my father's sides, when I would have expected my mom's German roots to be more pronounced than my father's Huguenots. The Scandinavian is on my mother's side (no surprise--New Sweden bits and bobs), as is the Iberian (a tiny, remote mystery).

AJL
12-14-2012, 01:00 AM
If you hover over the line for British and Irish, and click on the i for information symbol, you find that the sample included 843 23andMe customers and 89 public. So I think I can stop worrying that my particular ancestry (specifically lack of Welsh) might have biased the sample. There should be some Welsh in that lot. :)

Yes, I think what's more likely is (1) what History-of-Things alludes to, that some of my mother's Nonspecific Northern European matches are only to colonials, so they represent a part of the gene pool that was once British or Irish but is now more common in the US and elsewhere abroad than in the mother countries themselves and therefore not easily traced; and (2) other of my mother's Nonspecific Northern European regions represent rather inbred regional sub-populations that are simply not covered sufficiently in the present sampling for her to find matches.

Yorkie
12-15-2012, 07:55 AM
It looks like the new tool can identify British and Irish easily enough. My result is similar to Clinton's - just a fraction higher B & I (99.9% on speculative, 99.8% standard.)

I'm actually a little surprised at your 99.9% Bit/Irish score on speculative. I would have thought that you might register at least some Scandinavian and French/German with your ancestral connection to the Danelaw county of Lincolnshire. The Danelaw area dominates my ancestry and I get on speculative [rounding off figures] 75% Brit/Irish, 2% Scandinavian, 5% French/ German and 18% non-assigned Northern Euro. Maybe my score reflects a lot of Norman surnames, some Huguenot and a dash of more recent Norwegian in the 18% non-assigned NE?

Jean M
12-15-2012, 11:12 AM
@ Yorkie

Since the Danelaw ran over a big chunk of England, the genetic input from Denmark (both from Danish Vikings and Angles before them) should be a common part of the British gene pool. So would we expect it to show up as "foreign" after all these centuries of blending into the English mix?

I have 23andMe cousin matches to a few Scandinavians, but no Scandinavian input showing in the Ancestry Composition Results. That may mean that the cousin matches are the result of a relative of mine going to Norway or Sweden within the last few generations.

Yorkie
12-15-2012, 05:26 PM
@ Yorkie

Since the Danelaw ran over a big chunk of England, the genetic input from Denmark (both from Danish Vikings and Angles before them) should be a common part of the British gene pool. So would we expect it to show up as "foreign" after all these centuries of blending into the English mix?

I have 23andMe cousin matches to a few Scandinavians, but no Scandinavian input showing in the Ancestry Composition Results. That may mean that the cousin matches are the result of a relative of mine going to Norway or Sweden within the last few generations.

No, I certainly don't see 'Viking' dna as 'foreign' but rather, as you say, a 'common part of the British gene-pool'. I should have been more clear. What I mean is that, since the Danelaw is one of the areas of Britain where Scandinavians made a relatively high genetic impact, I would expect you to have 'a bit extra', so to speak, and thus get a percentage of the 'Scandinavian' as I do.

DMXX
12-15-2012, 06:36 PM
So would we expect it to show up as "foreign" after all these centuries of blending into the English mix?


From my perspective, 23andMe's unspecific sampling within countries has blurred much of the variation that would exist.

I was very surprised to see I came up as completely Middle-Eastern despite several of the Kurds I'm currently sharing with had European and East Asian segments. I use those Kurds as a comparison because, in those whose Dodecad, Harappa or Eurogenes results I'm aware of, they tend to score more Middle-Eastern-related ADMIXTURE components (e.g. West and/or Southwest Asian) and less European or East Asian than myself.

Alanson
12-15-2012, 09:31 PM
I don't like this program at all, because 19.4% i get nothing. The program needs improvement in my opinion. As well I don't have much Berber or North African ancestry, and it made it into 10%. It's either the Arabian component is similar to the North African, because it can't jump it seems like this. The Caucasian-Anatolian all went to the Middle East. They need to improve it.

Jean M
12-16-2012, 11:50 AM
No, I certainly don't see 'Viking' dna as 'foreign' but rather, as you say, a 'common part of the British gene-pool'. I should have been more clear. What I mean is that, since the Danelaw is one of the areas of Britain where Scandinavians made a relatively high genetic impact, I would expect you to have 'a bit extra', so to speak, and thus get a percentage of the 'Scandinavian' as I do.

Not sure about that Yorkie. The people behind the People of the British Isles (http://www.peopleofthebritishisles.org/) project keep saying that it is easy to identify Norwegian Viking ancestry (as in Orkney), but impossible to distinguish between Danish Vikings and Angles. The majority of Vikings within the Danelaw were Danes. As ever with the Vikings, one can't say more than "majority", because Viking war-bands recruited widely. Burials found in Castle Mall, Norwich, included four who stood out as Viking in the DNA analysis, with links to Orkney, Norway and the Western Isles.

warwick
12-18-2012, 01:48 AM
A couple results: 3% Scandinavian:

282

Jewish blocks:

283

Splitview:

284

J Man
12-27-2012, 02:53 AM
Hmmm my results make sense to a degree I suppose but my mother's and great uncle's do not make much sense at all really. For example my great uncle is from Northern Ireland. He was born and raised there and the vast majority of his ancestors before they reached Ireland came from Britain I am sure yet he only scores 64.6% British and Irish on ''Speculative'', 39.0% in the British and Irish category on ''Standard'' and 3.8% British and Irish on ''Conservative''. That really makes no sense unless as some seem to think here his matches are with individuals from the USA who's ancestors left Ireland and Britain a long time ago.

AJL
12-29-2012, 01:07 AM
Hmmm my results make sense to a degree I suppose but my mother's and great uncle's do not make much sense at all really. For example my great uncle is from Northern Ireland. He was born and raised there and the vast majority of his ancestors before they reached Ireland came from Britain I am sure yet he only scores 64.6% British and Irish on ''Speculative'', 39.0% in the British and Irish category on ''Standard'' and 3.8% British and Irish on ''Conservative''. That really makes no sense unless as some seem to think here his matches are with individuals from the USA who's ancestors left Ireland and Britain a long time ago.

Yes, I think that's likely what's going on. For example, your great uncle is probably one of both my and my mother's more recent matches (when I phase my maternal chromosome, he is in my top-5 at Gedmatch) but in Ancestry Finder that only shows as "Canada-Canada-Canada-Canada" for him unless he had other matches that overlap with my mother's and mine with recent ancestry all from the same Old World country.

Kyte
12-30-2012, 09:04 AM
My standard results:

292

I seem to be more of a mix than most individuals from the Balkans, even when taking into account that one grandparent is Anatolian. However it is still the single biggest entity on there. Also would be great if they put more detail into the MENA portion. Anyone know if this will happen?

azuly24
01-02-2013, 06:16 AM
I am Mexican which right away tells you I am a mix, but I didnt realize that the mestizo population was as mixed as when I saw my chromosome view from 23andme. However not too surprised with the Native American and European mixes but a little surprised with the middle east and african segments. 47% European, 42% Native American, 3% African and 1% Middle Eastern.


http://i22.photobucket.com/albums/b313/azuly24/chromo-view_zps10f68d6f.jpg

DMXX
01-02-2013, 08:28 AM
My standard results:

292

I seem to be more of a mix than most individuals from the Balkans, even when taking into account that one grandparent is Anatolian. However it is still the single biggest entity on there. Also would be great if they put more detail into the MENA portion. Anyone know if this will happen?

It's a good (and necessary) question given how 23andMe has defined their "Middle-East" category (Turkey, Iran, Lebanese Druze only).

When you have Saudi Arabians and Kuwaitis scoring predominantly North African rather than Middle-Eastern, you know something is wrong in the methodology. In this case, the "Middle-East" is pretty much taken as Anatolia and the Iranian plateau. No surprise that populations immediately south of this region score better with North Africans given the (likely) greater interactions between the Arabian Peninsula and North Africa.

However, since this is itself a tremendous step-up from their old Ancestry Painting, I don't think we're due for an update on this in a good few years.

GTC
01-02-2013, 08:33 AM
I am Mexican which right away tells you I am a mix, but I didnt realize that the mestizo population was as mixed as when I saw my chromosome view from 23andme. However not too surprised with the Native American and European mixes but a little surprised with the middle east and african segments. 47% European, 42% Native American, 3% African and 1% Middle Eastern.


That's the most colorful chromosome chart I've seen to date.

Being mostly of Irish and British ancestry, my chart shows just a few shades of blue and gray.

Kyte
01-02-2013, 01:53 PM
It's a good (and necessary) question given how 23andMe has defined their "Middle-East" category (Turkey, Iran, Lebanese Druze only).

When you have Saudi Arabians and Kuwaitis scoring predominantly North African rather than Middle-Eastern, you know something is wrong in the methodology. In this case, the "Middle-East" is pretty much taken as Anatolia and the Iranian plateau. No surprise that populations immediately south of this region score better with North Africans given the (likely) greater interactions between the Arabian Peninsula and North Africa.

However, since this is itself a tremendous step-up from their old Ancestry Painting, I don't think we're due for an update on this in a good few years.

Fair, but I'd imagine it won't be as difficult to tweak results every so often as it was to change the entire system from the Ancestry Painting to the Ancestry Composition. I remember talk on the 23andMe forum about updates on the system as participants increase, so hopefully we'll get increased detail every so often. A distinction between Caucasian, Mesopotamian, Arabian and perhaps even Iranian, would be pretty useful.

J Man
01-03-2013, 12:07 AM
Yes, I think that's likely what's going on. For example, your great uncle is probably one of both my and my mother's more recent matches (when I phase my maternal chromosome, he is in my top-5 at Gedmatch) but in Ancestry Finder that only shows as "Canada-Canada-Canada-Canada" for him unless he had other matches that overlap with my mother's and mine with recent ancestry all from the same Old World country.

Thanks for the reply AJL. It makes more sense to me now. :)

SC11
01-06-2013, 02:31 AM
I am Mexican which right away tells you I am a mix, but I didnt realize that the mestizo population was as mixed as when I saw my chromosome view from 23andme. However not too surprised with the Native American and European mixes but a little surprised with the middle east and african segments. 47% European, 42% Native American, 3% African and 1% Middle Eastern.


http://i22.photobucket.com/albums/b313/azuly24/chromo-view_zps10f68d6f.jpg

It's not too surprising that you have Middle Eastern being from Latin America. I am part Spanish-New Mexican and it's typical for us to have Middle Eastern DNA

Onetalllnerd
01-07-2013, 11:38 PM
Here's my results! I'm Mexican. :D http://imgur.com/R7Y3d.jpg

AJL
01-08-2013, 12:40 AM
Here's my results! I'm Mexican. :D http://imgur.com/R7Y3d.jpg

Interesting. The 1.9% Ashkenazi as well as azuly24's small amount is probably not Ashkenazi, but Judeo-Spanish (Sephardi) genes that were later incorporated into the Ashkenazi community. The "Middle Eastern & North African" may be partly that but more likely Berber, though some may also arise from the Phoenician colonies of Iberia.

You both have slightly more Italian than I'd have thought. Perhaps this represents some link from when much of southern Italy was under Spanish control, and which runs from Spain to Italy rather than the other way around.

The small amount of East Asian might be from the Philippines, or might represent a vestige of Spanish genes that is now more common there than in Spain.

Onetalllnerd
01-08-2013, 12:54 AM
Interesting. The 1.9% Ashkenazi as well as azuly24's small amount is probably not Ashkenazi, but Judeo-Spanish (Sephardi) genes that were later incorporated into the Ashkenazi community. The "Middle Eastern & North African" may be partly that but more likely Berber, though some may also arise from the Phoenician colonies of Iberia.

You both have slightly more Italian than I'd have thought. Perhaps this represents some link from when much of southern Italy was under Spanish control, and which runs from Spain to Italy rather than the other way around.

The small amount of East Asian might be from the Philippines, or might represent a vestige of Spanish genes that is now more common there than in Spain.

That's very interesting hmmm. I sent in samples from three of my grandparents. One of my grandpas died long ago so I didn't have the chance to test him. I'll post up my grandparents results when I get them. One of my cousins also has way more than Italian, around 20%. I might as well post his too. Here: http://i.imgur.com/VES9M.png

AJL
01-08-2013, 01:22 AM
Thanks, that'll be interesting. With 20% Italian, I would think some actual Italian ancestry is likely, but I didn't think there was much Italian migration to Mexico.

Veridicus
01-09-2013, 04:50 AM
I am kinda of confused with my result of 7.1% Ashkenazi with no known Jewish ancestors and I am pretty familar with my great great grandparents. This seems quit high for my most recent immigrant from mainland Europe to be German around 1860's. And of all the records I have of them they seemingly are are not Jewish.

How reliable is this designation - it seems that the reference population (1015 which is the most of all Euro populations) is soley 23andMe customers who identify as Ashkenazi and that your Ancestry Compositon results are based on shared segments with them in the Relative Finder. There does not seem to be any SNP's for this group?

AJL
01-09-2013, 05:42 AM
How reliable is this designation - it seems that the reference population (1015 which is the most of all Euro populations) is soley 23andMe customers who identify as Ashkenazi and that your Ancestry Compositon results are based on shared segments with them in the Relative Finder.

The Ashkenazi gene pool is unique due to prolonged inbreeding, and since about 10% of the 23andme user base has Ashkenazi ancestry, this is a very reliable forecast. If you had between 0.1-3.0% Ashkenazi, you might theorize that the shared ancestor was non-Ashkenazi, but with over 5%? No.

Your figure of 7% fits with one of your great-grandparents being completely Ashkenazi (in other words, you are about 1/16th Ashkenazi). Depending on your age, this may fit with someone born around 1800-1900 but most typically around 1850. The most likely thing is (1) your German ancestor who immigrated in the 1860s did not declare themselves as Ashkenazi but was; or less likely (2), there is another fairly recent Ashkenazi ancestor in your tree around the same time whom you have not identified, possibly owing to an error in genealogy, or to some other irregularity, such as adoption, marital infidelity, name change, etc.

Bear in mind that there is no magic technique to tell a Jewish Müller or Schmidt from a German Müller or Schmidt from name, census records, or family myths.
Given that your yDNA and mtDNA seem typical of western Europe, I would be looking at a non-direct line, i.e., not your Y or mtDNA ancestors.

Veridicus
01-09-2013, 05:54 AM
The Ashkenazi gene pool is unique due to prolonged inbreeding, and since about 10% of the 23andme user base has Ashkenazi ancestry, this is a very reliable forecast. If you had between 0.1-3.0% Ashkenazi, you might theorize that the shared ancestor was non-Ashkenazi, but with over 5%? No.

Your figure of 7% fits with one of your great-grandparents being completely Ashkenazi (in other words, you are about 1/16th Ashkenazi). Depending on your age, this may fit with someone born around 1800-1900 but most typically around 1850. The most likely thing is (1) your German ancestor who immigrated in the 1860s did not declare themselves as Ashkenazi but was; or less likely (2), there is another fairly recent Ashkenazi ancestor in your tree around the same time whom you have not identified, possibly owing to an error in genealogy, or to some other irregularity, such as adoption, marital infidelity, name change, etc.

Bear in mind that there is no magic technique to tell a Jewish Müller or Schmidt from a German Müller or Schmidt from name, census records, or family myths.
Given that your yDNA and mtDNA seem typical of western Europe, I would be looking at a non-direct line, i.e., not your Y or mtDNA ancestors.

Thanks for the response! Man, now I am gonna have to go back and do a second look at my records. I am having my mother tested next with 23andMe since 2 of her Great Grandparents were both from Germany around that time. Of course it will probably be difficult to confirm whether they were Jewish or not since having acess to records in Germany is hard to come by for a person in the US and even if found probably would not say much about whether they were Jewish.

I am so curious about this - it's annoying!

I was just wondering since European Jews are similar to Europeans how do they know that these segments are not non-Jewish? Even though they were pretty homogenous they still have European looking signatures and are still related to other Middle Easterners. So would it be possible that these segments are non-Jewish? It seems to me that 23andMe should have used SNPs and not customers that self identified as Jewish - that's a cultural/Religous designation not a genetic one. Maybe I am not understanding how they arrived at this result??

Also, in looking at my chromosome view on Ancestry Composition the X chromosome is completly absent of Ashkenazi - it is about 2/3 Northern European and 1/3 non specific European. I am also assuming that they do not know what chromosomes are from which parent. As such there is never any Ashkenazi on both pairs just on one pair over 12 chromosomes. Does this mean that it is coming from one parent and if so can it still be my mom since her X is absent any yet it is her father that these Ashkenazi signitures are most likely coming from since his mother's parents were from Germany?

AJL
01-09-2013, 04:42 PM
I was just wondering since European Jews are similar to Europeans how do they know that these segments are not non-Jewish? Even though they were pretty homogenous they still have European looking signatures and are still related to other Middle Easterners. So would it be possible that these segments are non-Jewish? It seems to me that 23andMe should have used SNPs and not customers that self identified as Jewish - that's a cultural/Religous designation not a genetic one. Maybe I am not understanding how they arrived at this result??

Also, in looking at my chromosome view on Ancestry Composition the X chromosome is completly absent of Ashkenazi - it is about 2/3 Northern European and 1/3 non specific European. I am also assuming that they do not know what chromosomes are from which parent. As such there is never any Ashkenazi on both pairs just on one pair over 12 chromosomes. Does this mean that it is coming from one parent and if so can it still be my mom since her X is absent any yet it is her father that these Ashkenazi signitures are most likely coming from since his mother's parents were from Germany?

Ancestry Composition calls Ashkenazi genes "European," but this is not entirely accurate. My grandfather is entirely Ashkenazi and his Global Similarity Plot is right in the Near East, besides Druze. By "signatures" do you mean yDNA, mtDNA, or autosomes? I can assure you that Ashkenazi autosomes have been the subject of much study and are not that close to most European populations: certainly not Germans. Select yDNA and mtDNA haplogroups now found among Ashkenazim may certainly have arisen around the Rhine, but uniparental markers are not as good as predictors of overall ancestry as autosomal data. To go back to my grandfather, based on his yDNA/mtDNA profile, he should appear to be completely Eastern European, but his autosomes belong around Lebanon.

The lack of any Ashkenazi on your X doesn't exactly prove which line your ancestor is along, because the X often recombines in females. However, because the X is not transmitted from male to male, it does make it somewhat more likely that you are looking for a more male transmission path: perhaps on your father's side, or perhaps your mother's father's father's, but in fact any line is possible.

Your mother's results will be interesting, of course. She has two Xs, and if she does have substantial Ashkenazi on her X, you can narrow your search to her, and more particularly to her father's mother, or her mother, rather than her father's father. If she doesn't have much Ashkenazi, you can look to your father's side, from either of his parents. Also, if you test a parent, Ancestry Composition data can be split to see from which side you have inherited a particular ancestry. If your father is available to test, his X would be interesting too.

Though I can see why you're confused that a religious group should have genetic significance, genetic-religious confluences can arise from inbreeding of others with the same belief system and can be found in other religious minority groups -- e.g. Parsis (South Asian Zoroastrians), Anabaptists (Mennonites, Hutterites, Amish), as well as in any population that has been isolated from others for long periods. There weren't massive amounts of people becoming Ashkenazi in Europe over the last thousand years, apart from an influx of Sephardim about 1500. However, quite a number of people ceased practising Judaism for various reasons. So if you find significant blocks of matching (several percent of your genome with some blocks over 8 cM) to people with Ashkenazi ancestry, the direction is most likely coming from Judaism to Christianity, rather than the reverse.

Veridicus
01-09-2013, 08:50 PM
Ancestry Composition calls Ashkenazi genes "European," but this is not entirely accurate. My grandfather is entirely Ashkenazi and his Global Similarity Plot is right in the Near East, besides Druze. By "signatures" do you mean yDNA, mtDNA, or autosomes? I can assure you that Ashkenazi autosomes have been the subject of much study and are not that close to most European populations: certainly not Germans. Select yDNA and mtDNA haplogroups now found among Ashkenazim may certainly have arisen around the Rhine, but uniparental markers are not as good as predictors of overall ancestry as autosomal data. To go back to my grandfather, based on his yDNA/mtDNA profile, he should appear to be completely Eastern European, but his autosomes belong around Lebanon.

The lack of any Ashkenazi on your X doesn't exactly prove which line your ancestor is along, because the X often recombines in females. However, because the X is not transmitted from male to male, it does make it somewhat more likely that you are looking for a more male transmission path: perhaps on your father's side, or perhaps your mother's father's father's, but in fact any line is possible.

Your mother's results will be interesting, of course. She has two Xs, and if she does have substantial Ashkenazi on her X, you can narrow your search to her, and more particularly to her father's mother, or her mother, rather than her father's father. If she doesn't have much Ashkenazi, you can look to your father's side, from either of his parents. Also, if you test a parent, Ancestry Composition data can be split to see from which side you have inherited a particular ancestry. If your father is available to test, his X would be interesting too.

Though I can see why you're confused that a religious group should have genetic significance, genetic-religious confluences can arise from inbreeding of others with the same belief system and can be found in other religious minority groups -- e.g. Parsis (South Asian Zoroastrians), Anabaptists (Mennonites, Hutterites, Amish), as well as in any population that has been isolated from others for long periods. There weren't massive amounts of people becoming Ashkenazi in Europe over the last thousand years, apart from an influx of Sephardim about 1500. However, quite a number of people ceased practising Judaism for various reasons. So if you find significant blocks of matching (several percent of your genome with some blocks over 8 cM) to people with Ashkenazi ancestry, the direction is most likely coming from Judaism to Christianity, rather than the reverse.

Thanks AJL you have benn most helpful! I do share larger segments over 8cM - I believe the longest ones were around 15-18cm with designation of 3rd cousin. Well I do look forward to figuring this out with the help of my moms results. I am now going to GEDmatch - should I expect anything with Ashkenazi and what tool would be best?

Thanks again!

AJL
01-09-2013, 09:10 PM
Thanks AJL you have benn most helpful! I do share larger segments over 8cM - I believe the longest ones were around 15-18cm with designation of 3rd cousin. Well I do look forward to figuring this out with the help of my moms results. I am now going to GEDmatch - should I expect anything with Ashkenazi and what tool would be best?

Thanks again!

You're most welcome. Given the length of the longest of those segments, I'd say again that you're probably looking at a fully Ashkenazi great-grandparent. Gedmatch could be very useful: I'd suggest going to the Admix page after your results are processed and putting your genome through Jtest, then Oracle 4, and reporting back. Whatever admix tool you choose, pay attention to components like West Asian, Southwest Asian, Southern, etc., in the raw reports (these have different names with different tools, but essentially you are looking to see what proportion of your genome fits in the Eastern Mediterranean/Near East).

Veridicus
01-09-2013, 10:02 PM
You're most welcome. Given the length of the longest of those segments, I'd say again that you're probably looking at a fully Ashkenazi great-grandparent. Gedmatch could be very useful: I'd suggest going to the Admix page after your results are processed and putting your genome through Jtest, then Oracle 4, and reporting back. Whatever admix tool you choose, pay attention to components like West Asian, Southwest Asian, Southern, etc., in the raw reports (these have different names with different tools, but essentially you are looking to see what proportion of your genome fits in the Eastern Mediterranean/Near East).

Ok! Here are the results:

1 NORTH-CENTRAL_EURO 25.90
2 ATLANTIC 21.99
3 EAST_EURO 13.93
4 SOUTH_BALTIC 11.17
5 WEST_MED 8.36
6 ASHKENAZI 7.69
7 EAST_MED 6.14
8 WEST_ASIAN 3.83
9 SOUTH_ASIAN 0.99

PureEvil
01-09-2013, 11:24 PM
Speculative Estimate:
321

Standard Estimate:
322

Nothing too exciting...

AJL
01-10-2013, 01:07 AM
Ok! Here are the results:

1 NORTH-CENTRAL_EURO 25.90
2 ATLANTIC 21.99
3 EAST_EURO 13.93
4 SOUTH_BALTIC 11.17
5 WEST_MED 8.36
6 ASHKENAZI 7.69
7 EAST_MED 6.14
8 WEST_ASIAN 3.83
9 SOUTH_ASIAN 0.99

Thanks -- the Ashkenazi and Eastern Mediterranean are both non-trivial (exceeding noise levels). Did you try the Oracle?

Veridicus
01-10-2013, 01:49 AM
Thanks -- the Ashkenazi and Eastern Mediterranean are both non-trivial (exceeding noise levels). Did you try the Oracle?

Yeah - do difference, here is the complete results: I don't know how to read the latter two - what are some of those populations like 'NO'? What are the @ numbers?

# Population Percent
1 NORTH-CENTRAL_EURO 25.89
2 ATLANTIC 21.98
3 EAST_EURO 13.93
4 SOUTH_BALTIC 11.17
5 WEST_MED 8.36
6 ASHKENAZI 7.68
7 EAST_MED 6.14
8 WEST_ASIAN 3.83
9 SOUTH_ASIAN 0.99
10 MIDDLE_EASTERN 0.03

Single Population Sharing:

# Population (source) Distance
1 West_&_Central_German 6.29
2 NL 6.87
3 DK 7.69
4 English 7.92
5 Cornish 8.92
6 AT 9.21
7 NO 9.42
8 Orcadian 9.45
9 South_&_Central_Swedish 10
10 IE 10.35
11 Scottish 11
12 North_Swedish 11.86
13 FR 12.34
14 HU 14.52
15 Serbian 14.96
16 PT 18.09
17 RO 18.33
18 South_Finnish 18.7
19 ES 19.21
20 North_Italian 20.21

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 82.4% DK + 17.6% AJ @ 2.56
2 79.4% NO + 20.6% AJ @ 3.49
3 78.9% South_&_Central_Swedish + 21.1% AJ @ 4.38
4 86.4% NL + 13.6% AJ @ 4.38
5 88.7% DK + 11.3% Mandean @ 4.45
6 88.4% DK + 11.6% Assyrian @ 4.46
7 83.9% DK + 16.1% GR @ 4.47
8 88.4% DK + 11.6% Armenian @ 4.51
9 87.9% DK + 12.1% Kurdish @ 4.53
10 88.6% West_&_Central_German + 11.4% AJ @ 4.56
11 86.7% DK + 13.3% TR @ 4.58
12 83.3% DK + 16.7% South_Italian_&_Sicilian @ 4.6
13 87.8% DK + 12.2% IR @ 4.66
14 89.4% DK + 10.6% Druze @ 4.66
15 74.9% DK + 25.1% RO @ 4.9
16 84.4% English + 15.6% AJ @ 4.92
17 88.9% DK + 11.1% IQ @ 4.95
18 89.8% DK + 10.2% Samaritan @ 4.98
19 70.7% DK + 29.3% Serbian @ 5.01
20 80.1% DK + 19.9% Tuscan @ 5.23

Veridicus
01-10-2013, 01:57 AM
I also did the EUtest which is, from what I read, the same but without the Ashkenazi alleles. It seems that the Ashkenazi is mostly 'hiding' behind East Mediterranean 1.87%, Middle East 1.69%, and the West Asian 1.21%. Does this sound reasonable? Here are the two results: By the way what is South Baltic?

JTest:

SOUTH_BALTIC 11.17%
EAST_EURO 13.93%
NORTH-CENTRAL_EURO 25.89%
ATLANTIC 21.98%
WEST_MED 8.36%
ASHKENAZI 7.68%
EAST_MED 6.14%
WEST_ASIAN 3.83%
MIDDLE_EASTERN -
SOUTH_ASIAN 0.99%

EUtest:

SOUTH_BALTIC 11.40%
EAST_EURO 14.50%
NORTH-CENTRAL_EURO 26.36%
ATLANTIC 22.89%
WEST_MED 9.16%
EAST_MED 8.01%
WEST_ASIAN 5.04%
MIDDLE_EASTERN 1.69%
SOUTH_ASIAN 0.95%

AJL
01-10-2013, 04:28 AM
I also did the EUtest which is, from what I read, the same but without the Ashkenazi alleles.

Did you try the Oracle 4 with Jtest? EUtest is like JTest without an Ashkenazi reference, so is not well suited to exploring the possibliity of such ancestry. Given that your results in the mixed mode are heavy with northern Near Eastern populations as well as Ashkenazi, I'd say the available evidence supports Ashkenazi ancestry.

Some of the populations are countries using a two-letter code, e.g. NO = Norway, FR = France, etc.

Veridicus
01-10-2013, 04:51 AM
Did you try the Oracle 4 with Jtest? EUtest is like JTest without an Ashkenazi reference, so is not well suited to exploring the possibliity of such ancestry. Given that your results in the mixed mode are heavy with northern Near Eastern populations as well as Ashkenazi, I'd say the available evidence supports Ashkenazi ancestry.

Some of the populations are countries using a two-letter code, e.g. NO = Norway, FR = France, etc.

Yeah there was very little difference:

# Population Percent
1 NORTH-CENTRAL_EURO 25.90
2 ATLANTIC 21.99
3 EAST_EURO 13.93
4 SOUTH_BALTIC 11.17
5 WEST_MED 8.36
6 ASHKENAZI 7.69
7 EAST_MED 6.14
8 WEST_ASIAN 3.83
9 SOUTH_ASIAN 0.99

Thanks!

DaddyIcy
01-10-2013, 06:37 AM
Hello, I am really curious about my results. I did DNA tribes in the past and got a 15,000 to one Jewish score and it said I scored higher than 97% of Jews. I just recently tested with 23 and Me and get 0% Ashkenazi. As far as I know I am all Italian. I have one great grandfather who didn't know his father. However, when I put my results into the admix estimates on Gedmatch.com I get results that are mostly all Ashkenazi or other jewish such as georgian jews or Moroccan jews, iraqi etc.... Some of them are as high as 98%. I am just trying to figure out why I am getting these results. Could anybody tell me which would be most accurate? Thanks for your time.

GTC
01-10-2013, 10:32 AM
Hello, I am really curious about my results. I did DNA tribes in the past and got a 15,000 to one Jewish score and it said I scored higher than 97% of Jews. I just recently tested with 23 and Me and get 0% Ashkenazi. As far as I know I am all Italian. I have one great grandfather who didn't know his father. However, when I put my results into the admix estimates on Gedmatch.com I get results that are mostly all Ashkenazi or other jewish such as georgian jews or Moroccan jews, iraqi etc.... Some of them are as high as 98%. I am just trying to figure out why I am getting these results. Could anybody tell me which would be most accurate? Thanks for your time.

I don't know about accuracy when it comes to these tools (I've read some adverse criticism of DNA Tribes) but it's quite feasible to have Italian-Jewish ancestry:

http://en.wikipedia.org/wiki/Italian_Jews

J Man
01-10-2013, 03:41 PM
I am curious as to which setting people on here find to be the most accurate when it comes to reading your Ancestry Composition results? ''Speculative'', ''Standard'' or ''Conservative''?

NK19191
01-10-2013, 05:32 PM
I am Iranian. Here is my ancestry composition result.



http://img703.imageshack.us/img703/446/ancestrycomposition.png

Scarlet Ibis
01-10-2013, 10:46 PM
I am curious as to which setting people on here find to be the most accurate when it comes to reading your Ancestry Composition results? ''Speculative'', ''Standard'' or ''Conservative''?

Good question. I have so much "nonspecific" stuff, that it doesn't really make too much of a difference for me either way.

Conservative:

http://i50.tinypic.com/mjpf6s.jpg


Standard:

http://i45.tinypic.com/34ta5vr.jpg


Speculative:

http://i46.tinypic.com/6jq23t.jpg

AJL
01-11-2013, 01:07 AM
Hello, I am really curious about my results. I did DNA tribes in the past and got a 15,000 to one Jewish score and it said I scored higher than 97% of Jews. I just recently tested with 23 and Me and get 0% Ashkenazi. As far as I know I am all Italian. I have one great grandfather who didn't know his father. However, when I put my results into the admix estimates on Gedmatch.com I get results that are mostly all Ashkenazi or other jewish such as georgian jews or Moroccan jews, iraqi etc.... Some of them are as high as 98%. I am just trying to figure out why I am getting these results. Could anybody tell me which would be most accurate? Thanks for your time.

Ancestry analysis tools can't hold a candle to real people to compare to. Sicilians and Calabrians are often a little difficult to distinguish from Jews on genetic analysis calculators, and if you don't match any Ashkenazi people at 23andme, it's extremely unlikely that you have Ashkenazi ancestry. You still might have other kinds of Jewish ancestry as GTC mentioned, but if none of your matches have Jewish ancestry I would lean toward the view that DNA Tribes (which is often hilariously wrong) and these other tools are simply confusing two different eastern Mediterranean populations.

AJL
01-11-2013, 01:16 AM
I am Iranian. Here is my ancestry composition result.


Thanks -- looks fairly typical for Iranians. "Middle Eastern" as a single term is somewhat ambiguous, because most BGA calculators nowadays divide "West Asian," peaking around the Caucasus, from a more southerly component (corresponding fairly well with Semitic peoples).

J Man
01-11-2013, 01:19 AM
Speculative seems to match best with my own known ancestry.

Scarlet Ibis
01-11-2013, 04:34 AM
As far as the Northern European part of my ancestry composition is concerned, the speculative part is probably the most accurate, too. The only reason why, though, is because the proportion of British & Irish on the other settings is too low, in proportion to the French & German. It's skewed too much by the majority of it being pushed into "nonspecific northern european."

I did find out very recently that I have post-1800 ancestry from Germany on my dad's side, but the ratio is still off.

Ancestry Composition seems to get a little....confused sometimes when it comes to Americans of mixed backgrounds. And it's not just because we weren't used as reference samples. I'm sharing with a few Europeans & East Asians who didn't fill out their ancestry survey, but who get clearer results. Which is fine. It's an ambitious feature of 23andme, but not perfect.

Also, just as a side note, it would be cool if they would give a few ethnicity breakdowns for non-europeans, too. It would have been cool to see if they were able to pinpoint my Korean ancestry as Korean.

geebee
01-11-2013, 06:11 AM
http://i1205.photobucket.com/albums/bb429/gjsb56/AC-conservative-Gary_zps232e9138.jpg
http://i1205.photobucket.com/albums/bb429/gjsb56/AC-standard-Gary_zpsc62ee776.jpg
http://i1205.photobucket.com/albums/bb429/gjsb56/AC-speculative-Gary_zps0c634c2a.jpg

Interesting that it does suggest a tiny bit of Finnish (up to 0.5%). FTDNA's Population Finder showed "100% European, +/- 0.01%", but suggested a breakdown into "Finnish, Russian, Spanish, Tuscan". I have no Finnish ancestors that I know of, although there might be something via my German ancestry. I'm pretty sure I don't have Russian ancestry.

I am reasonably certain that "Balkan, Italian, and Iberian" all are tied to the same Catalan ancestry from two of my 2nd great grandfathers.

PureEvil
01-11-2013, 08:36 AM
I think it's reasonable to say that the "Speculative Estimate" seems to fit for the most part with my known ancestry as well, being as I'm predominately English and French by ancestry to my knowledge.

Scarlet Ibis
01-13-2013, 02:12 AM
Mine changed today!

On speculative:



http://i50.tinypic.com/25a5a92.jpg

azuly24
01-14-2013, 07:56 PM
I am curious as to which setting people on here find to be the most accurate when it comes to reading your Ancestry Composition results? ''Speculative'', ''Standard'' or ''Conservative''?

Personally when I use the conservative view I get 44% unassigned, in standard view I get 23% unassigned and in speculative only 7% unassigned.

I sent my raw data and got my results from Doug McDonald and they are more similar to the speculative view.

J Man
01-15-2013, 05:56 AM
I have another question for those on here who are more knowledgeable about this than I am. Ancestry Composition results are based on IBS and not IBD segments correct?

AppalachianGumbo
01-20-2013, 01:49 AM
Here is that of my father
http://i1087.photobucket.com/albums/j470/RebeccaSummey/DadSpecAP2_zps6473ab0b.jpg

Here is mine.
http://i1087.photobucket.com/albums/j470/RebeccaSummey/BeckySpecAp2_zpsaf6973a3.jpg

Here is my *Split View* with one parent connected, my father.
http://i1087.photobucket.com/albums/j470/RebeccaSummey/Split_zpsfd013542.jpg

Scarlet Ibis
01-20-2013, 05:18 AM
^ How do you do the split view?


Edit: Nevermind, I found it. Never knew about this feature!

AppalachianGumbo
01-20-2013, 04:37 PM
^ How do you do the split view?


Edit: Nevermind, I found it. Never knew about this feature!

Okay. Just takes a day or so to compute.

Onetalllnerd
01-21-2013, 04:46 PM
Here's my results again and my maternal grandparents. :) http://i.imgur.com/dY8r3VF.png http://i.imgur.com/QdGolA8.png http://i.imgur.com/fEL1a9a.png

Scarlet Ibis
01-27-2013, 03:24 AM
23andme confirmed on their homepage that they have a temporary bug, where many customers' Sub-Saharan African results are being elevated to absurd amounts. I first noticed it on my dad's account, so I started looking around, and some of the stuff is hilarious.

Some of the most extreme examples I've seen on my share list:
-A 100% Slovenian woman who's now at 66.6% SSA
-A 100% Austrian woman who's now at 67.6% SSA
-A Colombian went all the way up to 94.8% SSA

AJL
01-27-2013, 04:42 AM
Well, that would account for how my mother went from 100% European to 17% Subsaharan African overnight without passing any on to me!

Scarlet Ibis
01-27-2013, 05:31 AM
Well, that would account for how my mother went from 100% European to 17% Subsaharan African overnight without passing any on to me!

lol Yeah, I noticed my dad's shot up from 0% to 6%, but mine is still 0%. At first, I was surprised, but when I started looking around on my share list, I realized that the 6% explosion is NOTHING compared to what I'm seeing on some.

I mean the one Colombian I saw had a score that exploded from 0% to 94.8%??? That's crazy.

AJL
01-27-2013, 04:19 PM
Maybe they've got Jane Elliott doing an experiment with Ancestry Composition? :)

http://www.janeelliott.com/

Scarlet Ibis
01-27-2013, 05:16 PM
lol Saw a documentary about her recently. In one part of the film, someone with blue eyes in the audience asked her why she was so hateful against blue-eyed people if she has blue eyes herself. Jane replied with something like, "Because I know how to behave like a brown-eyed person, and I have a brown-eyed husband." Classic. :lol:

Veridicus
01-27-2013, 05:40 PM
Maybe they've got Jane Elliott doing an experiment with Ancestry Composition? :)

http://www.janeelliott.com/

I saw her on Oprah way back and thought how can this audience be so stupid to fall for her crap. I would have just laughed at her.

geebee
01-27-2013, 05:52 PM
I've very glad I happened to see this thread before I looked at Ancestry Composition tonight! My father and my wife -- both previously 100% European -- became mysteriously sub-Saharan African. In fact, all the way up to 39% for my dad (from 0, or at least <0.1%). Funny thing, none of his three tested kids (out of six total) inherited any of this. :)

Is it just 100% Europeans this has happened to? My brother, my sister, and I -- plus my daughter -- all have a small percentage of Native American ancestry, and there was no change in sub-Saharan. But maybe that's just coincidence?

EDIT: If I didn't already know it's a glitch, chromosome view would have told me that. In some instances, dad is shown as having sub-Saharan ancestry in the same places on both chromosomes -- including all of 2, 5, 7, 6, and 18. That means my siblings and I could hardly help but have sub-Saharan ancestry on those same chromosomes, if it were real.

Scarlet Ibis
01-27-2013, 06:46 PM
I've very glad I happened to see this thread before I looked at Ancestry Composition tonight! My father and my wife -- both previously 100% European -- became mysteriously sub-Saharan African. In fact, all the way up to 39% for my dad (from 0, or at least <0.1%). Funny thing, none of his three tested kids (out of six total) inherited any of this. :)

Is it just 100% Europeans this has happened to? My brother, my sister, and I -- plus my daughter -- all have a small percentage of Native American ancestry, and there was no change in sub-Saharan. But maybe that's just coincidence?

EDIT: If I didn't already know it's a glitch, chromosome view would have told me that. In some instances, dad is shown as having sub-Saharan ancestry in the same places on both chromosomes -- including all of 2, 5, 7, 6, and 18. That means my siblings and I could hardly help but have sub-Saharan ancestry on those same chromosomes, if it were real.

Yeah, same here. My dad's AC shot up to 6% SSA on the same places on both chromosomes, but I'm still at 0%.

As for whether or not it just affects Europeans, I'm not sure....it seems to be worse with some Latin Americans (mixed European and Native American) who have shot up to 80-95% SSA.

Scarlet Ibis
01-27-2013, 06:56 PM
Update: Looks like they're in the process of fixing the bug. Or at least it's been fixed on my dad's profile.

AJL
01-27-2013, 07:13 PM
Is it just 100% Europeans this has happened to? My brother, my sister, and I -- plus my daughter -- all have a small percentage of Native American ancestry, and there was no change in sub-Saharan. But maybe that's just coincidence?

What was interesting to me is my grandfather, who does have a small SSA amount (pre-glitch), didn't see any increase in his.

PureEvil
01-28-2013, 02:39 AM
Mine hasn't changed at all and I have a very minor amount of SSA as well.

geebee
01-28-2013, 06:50 AM
Interesting. The "sub-Saharan African" problem seems to be fixed for my dad and my wife now. They're back to showing 100% European. And, all of my brother's and my sister's Native American once again are showing as maternal in split view. However, mine had been showing as entirely maternal, but now it's divided between my parents.

Here's how it breaks down:

For me, map view (speculative estimate) says 2.0% Native American/East Asian for me, all actually NA. Also <0.1% sub-Saharan African and 0.6% unassigned. The remainder is Euro. For my brother, the estimate is 1.8% EA/NA, with 1.7% NA and 0.1% nonspecific EA/NA; 0.1% unassigned. Our sister has 2.5% EA/NA, of which 2.3% is NA, 0.1% is EA, and 0.2% is nonspecific EA/NA. Another 0.1% is sub-Saharan African, less than 0.1% is Oceanian, and 0.6% is unassigned.

Now, split view says that from my mother I got 1.1% EA/NA, all NA. It also says less than 0.1% sub-Saharan African and 0.6% unassigned. In addition, it shows 0.9% EA/NA -- 0.8% NA and 0.1% nonspecific NA/EA -- from my father; and less than 0.1% unassigned. For my siblings, our mother is identified as the sole source of their Native American ancestry. Well, for my sister it actually shows less than 0.1% EA/NA from our father, all of that nonspecific.

But, this is a very recent change -- same time as the "fix". The siblings went from having their EA/NA from both parents to only having it from our mother, which is likely correct; while I went the opposite direction. The overall percentage still seems about right for me, but it should all be maternal. (I might also note that there are no chromosomes that show EA/NA on both strands in the same location. Likely, it's simply being miss-assigned, but I don't know why.)

AJL
01-28-2013, 04:34 PM
I still have no Subsaharan, my grandfather's is unchanged, and my mother's is gone, but there seem to be problems with the phasing engine because the Split View is very wonky. The Map View seems less precise than before but also includes new populations I'm not surprised to see based on some of my genetic cousins and history (Iberia, Italy, Scandinavia). This may only be an issue where the split view is extrapolated from one parent: I'm not sure.

geebee
01-29-2013, 07:23 AM
As Alice would say, "Curiouser and curiouser." Of the 2.0% NA that Ancestry Composition reports for me -- which is a figure that stands in agreement with Dr. McDonald, by the way -- I passed on over half to my daughter. Her percentage is reported as 1.5. Now, split view claims that only 1.1% came to me from my mother, and 0.9% from my father.

The curious parts are: (1) My father's own NA percentage is 0; and (2) all of my daughter's NA is paternal (from me), but none of that is in regions where she shares DNA with my father. So how could my mother pass on only 1.1% to me, but 1.5% to her granddaughter?

Of course, especially given that the NA/EA for my siblings is 2.5% for my sister and 1.7% for my brother -- and all of it maternal for both -- the obvious answer is that the phasing is incorrect. This is even more likely given that, until the SSA glitch, all three siblings were shown as having inherited all our NA ancestry from our mother.

AJL
01-29-2013, 06:27 PM
the obvious answer is that the phasing is incorrect.

Yup! That's the only way my mother could have passed on to me 17% Ashkenazi when she has 0%... Also, The British and Irish should all be coming from her, not my father, while the East Euro should be coming from him rather than her.

J Man
02-01-2013, 10:08 PM
I still don't know about all these Ancestry Composition results. I have noticed some strange things in my own and my family member's results. For example there are spot's on my chromosomes where I have large Finnish segments which I should have inherited from my father but in these same spots he shows either not as much Finnish (the segment is not as long as mine) or it is broken up into numerous smaller segments for him and not me. Its is strange how the segments can show up on different sides of the chromosomes too compared to your parents.

geebee
02-02-2013, 11:54 AM
I noticed just this evening that my split view has now become unavailable. Right now I get this message: " ... Split View results are almost ready and should be available within a few days." The same holds true for my brother and my sister, but not my daughter. In my daughter's case, both of her parents have been tested; whereas for my siblings and me, only our father has been tested. So it looks to me as if some sort of recalculation (or something) may be in the works to improve split view's accuracy. At least, so I hope.

It makes sense that there would be more difficulty with only one parent tested. If both parent and child are heterozygous for a given SNP -- say, AC and AC -- there's no way to know whether the child and parent have the A or the C in common. However, with two tested parents you only encounter this problem when all three are heterozygous. If either the child or either one of the parents is homozygous, then phasing for the child is easy.

Now, in a case like mine where you only have one parent but also multiple siblings tested, you could use the siblings' data to help with the phasing. So if both my dad and I are AC, but my brother is AA, then you can infer our mother to be either AA or AC, but not CC. And if my sister happens to be CC in the same location, then our mother must be AC.

I'd imagine that in a case like my father's -- with 10 full brothers and sisters -- if all of them were tested (unfortunately not the case) you could do a pretty decent job of phasing, even without either parent being tested. So all I'd have to do is buy 10 more kits and convince all my uncles and aunts on my father's side to do a little spitting for their "favorite nephew". (They sure wouldn't buy the kits themselves, even at $99.)

I think I'll do this after I test my remaining three untested siblings, and if I win big in the lottery. Not too likely, since I've never bought a ticket. ;)

AppalachianGumbo
02-03-2013, 03:20 PM
I noticed just this evening that my split view has now become unavailable. Right now I get this message: " ... Split View results are almost ready and should be available within a few days." The same holds true for my brother and my sister, but not my daughter. In my daughter's case, both of her parents have been tested; whereas for my siblings and me, only our father has been tested. So it looks to me as if some sort of recalculation (or something) may be in the works to improve split view's accuracy. At least, so I hope.

It makes sense that there would be more difficulty with only one parent tested. If both parent and child are heterozygous for a given SNP -- say, AC and AC -- there's no way to know whether the child and parent have the A or the C in common. However, with two tested parents you only encounter this problem when all three are heterozygous. If either the child or either one of the parents is homozygous, then phasing for the child is easy.

Now, in a case like mine where you only have one parent but also multiple siblings tested, you could use the siblings' data to help with the phasing. So if both my dad and I are AC, but my brother is AA, then you can infer our mother to be either AA or AC, but not CC. And if my sister happens to be CC in the same location, then our mother must be AC.

I'd imagine that in a case like my father's -- with 10 full brothers and sisters -- if all of them were tested (unfortunately not the case) you could do a pretty decent job of phasing, even without either parent being tested. So all I'd have to do is buy 10 more kits and convince all my uncles and aunts on my father's side to do a little spitting for their "favorite nephew". (They sure wouldn't buy the kits themselves, even at $99.)

I think I'll do this after I test my remaining three untested siblings, and if I win big in the lottery. Not too likely, since I've never bought a ticket. ;)

Yesterday, my split view is off and on. The day before yesterday it was being calculated but later in the day was up. It's working today.

ADW_1981
02-04-2013, 03:25 AM
After the Ancestry Composition 'fix', my mother's became a little more interesting. My father's profile is still a reference sample which throws my own results off.


British/Irish 47.3%
French/German 8.4%
Scandinavian 2.2%
Non-Specific North Euro 28.9%

1.2% Italian
0.4 % Iberian
4.0% Non-Specific South Euro

0.6 East Euro

4.6% Non-Specific European

1.2 East Asian
<0.1 Non-Specific East Asian

0.8 Sub-Saharan African

0.1 South Asian

0.4 Unassigned

Interestingly enough, she has no relatives on the East Asian or African segments. Lots of matches on the British/English segments with other people from UK and many colonial Americans.

J Man
02-08-2013, 03:47 PM
How far back in time do the Ancestry Composition results go? 500 years or so? Or longer?

geebee
02-08-2013, 04:17 PM
How far back in time do the Ancestry Composition results go? 500 years or so? Or longer?

I'm no expert, to be able to give you a definitive answer, but I suppose in one sense you can argue that Ancestry Composition is only really reliable to the number of generations involved in the reference samples. With the ones that 23andMe selected from "in house", I understand they were looking for people with all four grandparents from the same place. They excluded folks whose grandparents were from a colony, although that did not include folks whose grandparents were Native American. And they also excluded folks who proved to be "outliers" -- which I would guess means anyone whose results were significantly away from the norm for their population.

So I suppose part of the answer to your question would be, how stable were the populations of the grandparents' home countries? I mean, prior to the grandparents' own generation. (I'd imagine that this was one of the reasons for excluding outliers, because that could reflect recent "input" from some other region.)

Hopefully, someone more knowledgeable than I will either correct or expand on this answer.

geebee
02-08-2013, 05:25 PM
Okay, now here's something maybe I can get folks to comment on as well:

I have no known Finnish ancestry, and I was rather puzzled by what FTDNA decided my ancestry was when I took the Family Finder test. They came up with "100% European, ±0.01%, which was in contrast with 23andMe's Ancestry Painting. That estimated 98% European, 2% Asian.

The Population Finder map was a puzzle as well, because on paper about half my known ancestry is German (and a tiny bit of Dutch). Another quarter or so is British, an eighth is Spanish, and the remaining eighth is a mixed up blend of French, Irish, Swiss, and Native American. But the countries shown on the map did not include central Europe at all, and the populations were identified as "Finnish, Spanish, Tuscan, and Russian". Plus, the map was a uniform dark blue, which I think means that they could not estimate percentages for any of the populations.

However, I have NO known Finnish, Tuscan, or Russian ancestry -- and even my Spanish ancestry is definitely a minority one. The only thing I could figure was that the Population Finder algorithm first determined that I was European, and couldn't see enough non-European DNA to add a second continent. Once it limited itself to Europe, it was forced to match me as well as possible to ONLY European populations. The result was that it included populations which were the closest available ones to account for what 23andMe considered "Asian".

To make things even more peculiar, when they opened Family Finder to people with test results from 23andMe, I submitted the files for my brother and my sister. My brother's results were similar to mine in that he showed as fully European, except that his "European" was less generic than mine. He was listed as "72.44% Western European, ±8.99%" and "27.56% European, ±8.99%". This time, the Western European populations were French and Spanish instead of Spanish and Tuscan, though for "Europe" they still had "Finnish and Russian". There was, however, a color difference. The Western European part was still dark blue, but the generic Europe part was shown as orange. My sister's results were even more peculiar. She was listed as "97.33% Western European, ±0.26%" (French only), and "2.67% East Asian, ±0.26%" (Han, Japanese, Lahu, Malay, Mongolian, Yakut, Cambodian). Once again, the Western European portion was shown as dark blue. The Asian portion was shown as orange.

Later, I took the AncestryDNA test. The "Genetic Ethnicity Summary" had a breakdown of Central European 48%, Scandinavian 23%, Eastern European 16%, Finnish/Volga-Ural 6%, and Uncertain 7%. This seemed closer to my paper trail, although there still seemed to be an unexpected Finnish/Russian result. The "Scandinavian" also seemed a bit odd, but the vast majority of my "British" ancestry is actually Scottish and Scots-Irish, so I presumed that some or all of it might be showing as "Scandinavian". No Spanish was showing, but Eastern European seemed to include Greek, and some part of the Spanish ancestry -- which is actually Mediterranean ancestry, from the island of Menorca -- might possibly appear as Greek. Also, the "unknown" could account for some of it, as well as possibly accounting for the Native American ancestry.

Now Ancestry Composition is out, which should be a great deal more reliable than any of the previous tests -- if only because of improved population sets. I've posted my results elsewhere, but I'll mention them again here to try to bring all of this together. I'll go with the "speculative estimate", because it really does seem to be the closest to the paper trail of all the analyses I've seen so far. This has me as 97.3% European and 2.0% East Asian and Native American. There's also less than 0.1% sub-Saharan African, and 0.6% unassigned.

The "European" part is mostly northern: 36.1% British and Irish, 9.1% French and German, 0.5% Finnish, and 38.8% nonspecific northern European. In addition, there's a southern European component. That's broken down into 2.2% Italian, 1.3% Balkan, 1.0% Iberian, and 4.0% nonspecific southern European. All of these could have a common Mediterranean connection, and be related to my Menorcan ancestry, although they still total less than the paper trail 12.5%. However, there is also a "nonspecific European" part that totals 3.4%. If that were somehow associated with the Menorcan ancestry, it would bring the total to 11.9. (I might be tempted to throw in the 0.7% "Eastern European", as well, to end up with 12.6% -- but I think that would be stretching things.)

The 2.0% East Asian and Native American is broken down into 2.0% Native American and 0.1% nonspecific East Asian and Native American, which really doesn't quite add up; but essentially it seems to mean all of this is probably Native American.

Notice that there's still a tiny Finnish portion, though certainly not enough to warrant inclusion of "Finnish" if I'm limited to just four components. However, one of the things I really like about Ancestry Composition is the different "views" available. One of them, of course, is "chromosome view". In that view I can see that my "Finnish" ancestry is shown in the form of two segments. One of them is located on a chromosome shown almost entirely as "British and Irish", except for the Finnish segment. I'm aware that some "British" DNA might be seen as "Scandinavian", but could some also be seen as "Finnish"? (Yes, I'm aware that Scandinavian and Finnish are not the same thing.)

[The second "Finnish" segment is located near, but not touching, a Native American segment; and the rest of the chromosome is all "nonspecific northern European".]

Somewhere I think I have a question here, but I don't think I've articulated it very well -- if at all. Maybe I've just been thinking out loud? But I'd appreciate comments (I think).

AJL
02-08-2013, 08:39 PM
This is an interesting point. My Population Finder is somewhat similar to yours

Europe: Finnish, French, Orcadian, Romanian, Russian 40.69% ±26.29%
Europe (Southern European): Sardinian, Tuscan 59.31% ±26.29%

while my Ancestry Composition has smaller components of Eastern European (0.9%) and Italian (1.2%) and larger bits of French and German, British/Irish, and Ashkenazi. In effect, the emphasis sometimes shifts to one's minority-component populations for reasons I don't fully understand, but which might have to do with the probability algorithm. I still have no evidence of Finnish, but with one fairly close Finnish cousin, it's certainly possible.

As an anecdotal answer to your question, J Man, I have some populations represented that appear to go at least as far back as 1500-1700, and some (such as the Scandinavian) might conceivably go back to the years 1200-1500.

J Man
02-08-2013, 09:55 PM
Hmmm okay thank you for the answers and help both GB and AJL. I am starting to take more of an interest in Ancestry Composition now since they seem to be fixing it. It is actually starting to look like a pretty useful tool at least for us of European ancestry.

Maybe the size of the segments indicates age as well? But then these segments do not just appear at a certain time. They came down to us from ancestors who lived a very long time ago as well I suppose. The longer they are and the more of them that you have mean that you have quite a bit of ancestry from a certain group I would say. This is probably all very obvious stuff and I am just rambling on lol.

geebee
02-09-2013, 04:08 AM
Well, both segment size and number of segments do relate to the amount. In addition, I'd look at how many chromosomes included a particular ancestry. "More" would, I think, indicate more recent ancestry. For example, if one parent were East Asian and the other European, obviously you should expect to find both ancestries on all chromosomes -- even if neither parent were tested. A grandparent of a given ethnicity might still appear on most chromosomes, but the number may drop off pretty quickly going back.

From my southern European forebears, which should on paper constitute an eighth of my ancestry, it appears I only inherited about nine segments. Some of these segments seem to be quite large. (Of course, what looks like a single large segment could be multiple segments that happen to touch.)

As it happens, two out of my mother's four great grandfathers were both from the Spanish island of Menorca. One of them married an Alsatian immigrant, while the other married a woman of Irish, French, Swiss, and Native American ancestry. This means the two children who became my grandmother's father and mother should each have had a paternal set of fully Spanish chromosomes, although it's doubtful they would actually appear that way at 23andMe. But let's say they at least could be identified as "southern European". One of them would also have a complete maternal set of Alsatian chromosomes, perhaps identified either as "French and German", or at least as "northern European".

Now, because my grandmother had two half-Spanish grandparents, she herself would have been half-Spanish. Only, whereas both of them would have a full set of Spanish chromosomes, because of recombination my grandmother would most likely have had only a few "fully-Spanish" chromosomes, at most. She could conceivably have had none, but given the way recombination works, she might well have gotten one or more "unrecombined" Spanish chromosomes -- from either parent, or both. It's even possible that one or more apparently unrecombined, or minimally recombined, "Spanish chromosomes" could have been passed on from my grandmother to my mother, and from my mother to me. In fact, that seems possibly to have happened with my maternal chromosome 3. Except for a tiny portion at the very beginning, it's shown as completely southern European. Of course, it could be partly from both Spanish 2nd great grandfathers -- recombined by the time it reached their granddaughter. Interestingly, a sizable portion of this chromosome is shown as "Italian", even though my ancestors were not Italian -- as far as I know. But while I say they were "Spanish", it's possible that the composition of Menorcan DNA is somewhat different from that of the mainland.

Additionally, I can't be certain that the fact that it appears not to have a contribution from the non-Spanish 2nd great grandmothers (unless that's what the tiny part at the beginning is), indicates that it came from just one of the 2nd great grandfathers. However, that a single chromosome might almost entirely from just two of 16 2nd great grandparents is amazing to me. But the underlying point is really that if you know the ethnicities of your ancestors -- and if they came from multiple regions -- you can not only consider segments individually, but in relation to what other segments may appear on the same chromosome. This could be useful in ferreting out the possible origin(s) of that chromosome.

AJL
02-09-2013, 03:51 PM
337


However, that a single chromosome might almost entirely from just two of 16 2nd great grandparents is amazing to me.

Because of the higher rate of meiotic recombination in females, I suspect this is slightly more likely to happen if you're following a mainly or completely male path. If not probable, it does seem possible. Attached is myself as compared to my paternal grandfather (blue) and his paternal first cousin (green) -- so, my first cousin twice removed. The green clearly represents all the common DNA from no more than two great-great grandparents; however, since more or less my entire chromosome 6 comes from my grandfather, it's possible that the recombination happened on the cousin's side, and my whole chr 6 is from the same great-great grandparent(s).

geebee
02-10-2013, 10:19 AM
^Well, in a similar vein I've discovered something interesting -- if true -- about my X chromosome.

In "Chromosome View/Conservative Estimate", all but perhaps the first sixth of my X chromosome is identified as "nonspecific northern European". That first sixth is still European, but "nonspecific European". Now, this in itself wouldn't be unexpected, since my European ancestry is shown here as 94.8%.

However, it's in switching to the "Speculative Estimate" that things get interesting. All but a very tiny portion at the beginning of the X becomes "British and Irish". Still not so strange, perhaps, since I have both British and Irish ancestry. But here's why I find it "interesting".

My X, of course, is 100% from my mother. So the signification "British and Irish" contribution from my father is meaningless. On paper (and in AC), I do have some "British/Irish" input from my mother, but much less. Now, continuing to following the X path back, I have to consider both of her parents. But, I can basically just skip from my maternal grandfather to his mother. She was, as far as I can tell, 100% German on her father's side. And, continuing up the line on her mother's side, her mother's mother was also (to my knowledge) fully German. However, her mother's father's mother was likely British or Irish.

Now, on my mother's maternal side it's actually a bit easier. My grandmother's father's mother was an Alsatian immigrant. And, I know that her family had been in the Alsace-Lorraine area for generations. So I would expect an X contribution from her to perhaps look "French and German", or possibly "nonspecific northern European". Continuing on to my grandmother's mother's parents, her father's mother (in fact both his parents, but only one is relevant for current purposes) was a Catalan immigrant, specifically from the island of Menorca. I'd therefore expect an X contribution from this ancestor to perhaps be labeled as Iberian, or at least as "nonspecific southern European".

If you're still with me, I think we're at my grandmother's mother's mother's parents. The father's mother was paternally French, and maternally French and Native American. The mother's father was maternally Swiss and Native American; and the mother's mother was paternally French (on both sides, as far as I know), and the mother's mother was paternally pure French -- again, as far as I know -- and maternally French and Swiss. I should note that more than one pair of my maternal grandmother's maternal grandmother's ancestors married cousins. (I think the closest pair were first cousins, once removed.)

I know that was long and involved, but what I was attempting to say is that I don't seem to have many paths to a fully British and Irish X chromosome. It could be that the estimate is simply wrong; but if it is not, it almost seems as if the X would have to trace back more-or-less in its entirety to just one 4th great grandmother, Elizabeth Mills. Does that even seem possible?

EDIT: Looking at my siblings, I think it likely is more of an issue of the accuracy of the estimate. I have some of my X chromosome in common with my sister, yet no part of either of her Xs is identified as British and Irish -- even using the Speculative Estimate. In my brother's case, part of his X is identified as British and Irish -- including both part that we share, and part that we don't -- but the rest is only identified as "nonspecific northern European". But again that includes both part that we share, and part that we don't; and DNA identified as British in me is not identified the same way in him.

Don Felipe
02-13-2013, 11:11 AM
I'm half Cape Verdean, half Dutch. Quite pleased with AC, although it did take them very long to update. Also still eagerly looking forward to SSA and Asian components being subregionally specified. I'm relieved the fake Asian %'s showing up for Afrodescendants have finally disappeared and also the SSA is no longer being underestimated because of newly added reference populations. On the very first Ancestry Painting of 23&me my results were:

European 80.20%
African 16.47%
Asian 3.33%


Now they are much more in line with other Admixture tools i've tried :biggrin1:


Standard

http://i989.photobucket.com/albums/af18/oditous2/AncestryCompositionStandard.jpg



Probably just a coincidence but the 50,9 % North Euro I get on speculative estimate corresponds perfectly with my half Dutch background B) I'm not really surprised about the high non-specific %'s and minor UK, Scando i get, i've also seen them on Dutch cousins who didn't fill in the Ancestry Survey. Would be nice if eventually the results get more specific, but I'm wondering if it ever will be possible to completely & reliably delineate ethnicities who are closely related genetically? I thought the whole "overfitting" issue was quite comical, of course there's an inherent risk when using reference populations being composed of people with self reported ethnic backgrounds.


http://i989.photobucket.com/albums/af18/oditous2/AncestryCompositionSpeculative2.jpg

http://i989.photobucket.com/albums/af18/oditous2/AncestryCompositionSpeculative3.jpg



With other Admixture tools i also got some minor but significant MENA results which most likely derives from historically documented Sephardic ancestry running high among Cape Verdeans and possibly also from some North African geneflow into selected Senegambian ethnicities. It's not showing up right now, but apparently it's "hidden" in the Italian or unspecified South Euro category. Several people have already been clamouring about it, hopefully 23&me will fix this soon. Even so i do also have several Italian cousins in RF so my very minor "Italian" and "Sardinian" %'s might be real afterall.

The biggest surprise with AC sofar has been the minor South Asian showing up. Could it be above noise level? It's also showing up on conservative estimate. Plus the segments identified as being South Asian also appear on exactly the same spot in the chromosomepainting by Dr. Mc Donald.


http://i989.photobucket.com/albums/af18/oditous2/Chromosomepaintingspeculative.jpg

http://i989.photobucket.com/albums/af18/oditous2/BGA2.jpg

I've run Harappaworld to get some more details. Overall I only get 5.8% Baloch and no S-Indian %.

# Population Percent
1 NE-Euro 31.75
2 W-African 24.29
3 Mediterranean 19.98
4 Caucasian 10.73
5 Baloch 5.85
6 SW-Asian 4.27
7 E-African 2.23
8 San 0.91


However when selecting Admixture Proportions by Chromosome i do get some S-Indian %'s, most elevated at chromosomes 3 and 19 which also showed up on 23&me and Dr. Mc Donald. On chromosome 22 i seem to have an elevated Baloch affinity.


http://i989.photobucket.com/albums/af18/oditous2/Harappachromosomepercentage.jpg

AJL
02-13-2013, 04:42 PM
I'm half Cape Verdean, half Dutch.

{SNIP}

With other Admixture tools i also got some minor but significant MENA results which most likely derives from historically documented Sephardic ancestry running high among Cape Verdeans and possibly also from some North African geneflow into selected Senegambian ethnicities.

Welcome to Anthrogenica!

I have partial Sephardi ancestry, and an Ancestry Finder cousin from Cabo Verde who I suspect must be related through the Sephardi side.

There is no "Sephardi detector" with 23andme the way there is for Ashkenazi, but I do have non-trivial levels of both Iberian and Italian, which (since neither my mother nor my paternal grandfather has Iberian) must come from my paternal grandmother's Sephardic side. Thus it's completely credible that some of your Iberian and possibly Italian and Sardinian are Sephardi, consistent with McDonald's finding of Middle East/North African. Of course, some of that might also be gene flow from Africa into Southern Europe, as you note.

My mother has a little Dutch ancestry and gets fairly high Nonspecific Northern European. I don't think there are too many Dutch samples at 23andme, but we do have a few Dutch and Flemish Relative Finder cousins. She also has a little South Asian, which, since some of her ancestors came from areas of the Scottish Borders with significant Roma presence, is credible. I've also seen low percentages of various populations in other people that are much harder to explain, though.

Hanna
02-13-2013, 05:08 PM
I was definitely used as a reference sample:

http://img849.imageshack.us/img849/6246/map23andme.png
http://img694.imageshack.us/img694/9170/chrm23andme.png

geebee
02-13-2013, 07:14 PM
My father on paper should be almost equal proportions of French & German (really, all German) and British & Irish (nearly all Scot, or Scots-Irish). But even the speculative estimate only identifies 36.1% British & Irish and 21.6% French & German. There's also 1.5% Scandinavian, which I think more likely to be British than German, because as far as I know the Germans were nearly all southwestern. I'm not sure what the significance of the 0.4% Finnish is.

Curiously, he also shows 3.8% nonspecific southern European. For whatever it might be worth -- which is probably not much since it was based on average affinity and not admixture -- in Global Similarity he actually falls into three boxes: French, German, and Austrian. In addition, the German form of our family name may be Bavarian, but there have been suggestions it may be Austrian. But the time of emigration from Germany, however, our ancestor was living in the somewhat vaguely-defined "Palatinate". Of course, the surname ancestor is only one among many.

The remainder of the estimate consists of 33.9% "nonspecific northern European" and 2.7% nonspecific European.

AJL
02-13-2013, 07:29 PM
I'm not sure what the significance of the 0.4% Finnish is.

That could be from the Scots. There were a number of Scottish merchants, military, clergy, etc., in Finland.

Don Felipe
02-13-2013, 09:56 PM
Welcome to Anthrogenica!

I have partial Sephardi ancestry, and an Ancestry Finder cousin from Cabo Verde who I suspect must be related through the Sephardi side.


Thanks for the welcome! There's some Cape Verdeans with confirmed Sephardic ancestry dating from rather recent migrations in the late 1800′s from Morocco and Gibraltar. Here's a very interesting website (http://capeverdejewishheritage.org/about-2/meet-some-prominent-cape-verdean-jewish-families/) on them. Probably most Sephardic ancestry in Cape Verdeans dates from much earlier times though when the islands were getting settled in the 1500's and attracted many Converso's.



My mother has a little Dutch ancestry and gets fairly high Nonspecific Northern European. I don't think there are too many Dutch samples at 23andme, but we do have a few Dutch and Flemish Relative Finder cousins. She also has a little South Asian, which, since some of her ancestors came from areas of the Scottish Borders with significant Roma presence, is credible. I've also seen low percentages of various populations in other people that are much harder to explain, though.

I'm sharing with 12 Dutch cousins on 23&me. I'm pretty sure 7 of them have been used as samples for the French/German/Benelux category. They score 99%-100% for it.

This guy also has very high French/German, he probably did fill in the ancestry survey as well, but even so AC also picked up on his Ashkenazi ancestry and even some very minor SSA which might be noise though cause it doesn't show up in conservative mode.

http://i989.photobucket.com/albums/af18/oditous2/NL1.jpg

These Dutch persons most likely didn't fill in the ancestry survey. It's interesting that besides French/German being their primary component they also get hefty amounts of Scandinavian and British/Irish. The first two ones also show some very minor South Asian (also when set at conservative). I figure if my South Asian is indeed genuine I could have gotten it either through some distant gypsy blood (Dutch people sometimes jokingly refer to it when trying to explain darkhaired/exotic looking Dutch types) or through some distant South Asian ancestry from one of the former Portuguese settlements in India (Goa) that probably through an admixed descendant somehow ended up in Cape Verde probably with a detour via Portugal.


http://i989.photobucket.com/albums/af18/oditous2/NL5.jpg



http://i989.photobucket.com/albums/af18/oditous2/NL2.jpg



http://i989.photobucket.com/albums/af18/oditous2/NL3.jpg



http://i989.photobucket.com/albums/af18/oditous2/NL4.jpg







I was definitely used as a reference sample

Did you get any East Asian with other Admixture tools?

J Man
02-13-2013, 11:48 PM
That could be from the Scots. There were a number of Scottish merchants, military, clergy, etc., in Finland.

Yes this certainly is true. My mother who is of all Irish/British Isles ancestry shares one segment with a Finn on chromosome 6. She also has one single Finnish segment in Ancestry Composition also on chromosome 6. The Finn that she shares this segment with on Relative Finder has some known old Scottish or West European ancestry.

J Man
02-13-2013, 11:50 PM
It is interesting to see such high Scandinavian scores for those Dutch individuals you are sharing with Oditous. Maybe related to trade links between the Netherlands/Frisia and Scandinavia or maybe even some older Viking type connections.

Hanna
02-14-2013, 02:20 PM
Did you get any East Asian with other Admixture tools?

With other admixture tools I did score some Asian and European but my main components are West Asian/Caucasus and Mediterranean however, 23andme points me out as 100% Middle Eastern. Even the naming of the components is completely flawed, Middle Eastern appears to be West Asian/Caucasus and North African seems to be Southwest Asian.

Don Felipe
02-14-2013, 07:56 PM
It is interesting to see such high Scandinavian scores for those Dutch individuals you are sharing with Oditous. Maybe related to trade links between the Netherlands/Frisia and Scandinavia or maybe even some older Viking type connections.

Yes, I imagine so, I have a couple of Swedish and Danish cousins in RF & AF. There has been some low level migration between The Netherlands and Scandinavia in the historical era, maybe more significant in some selected places and during some particular time periods. Norwegian migration to Amsterdam during its Golden Age is pretty well known as is Dutch migration to the Swedish city of Goetheborg also in the 17th century. Besides that there's also the older Viking era connection, i suspect even before the Vikings actually the Frisians already had trading relations with Scandinavia.

I'm curious to know how Germans and Britons end up in AC when they haven't been used as reference samples. I'm only sharing with one German cousin, this is his AC set at speculative. I haven't seen any Dutch person sofar with that amount of East Euro. I suppose it depends a lot from which region they're from, both Germany and Great Britain being much larger than the Netherlands.

http://i989.photobucket.com/albums/af18/oditous2/De.jpg










With other admixture tools I did score some Asian and European but my main components are West Asian/Caucasus and Mediterranean however, 23andme points me out as 100% Middle Eastern.

I've seen your Mc Donald results, your minor East Asian components should normally show up, if even 0.1% is being detected in Europeans, same goes for your other non-"Middle Eastern" components. Hopefully this "overfitting" issue will be resolved soon.

J Man
02-15-2013, 02:22 AM
^From what I have seen so far at least some British and Irish people score very high in the ''Nonspecific Northern European'' category when they are not included in the reference samples. My great uncle is 100% North Irish Protestant and here are his results below.

3.8% British and Irish
70.6% Nonspecific Northern European
24.4% Nonspecific European

I have seen other British results similar to his as well. Here is one from an English fellow I am sharing with.

0.6% British and Irish
61.1% Nonspecific Northern European
36.2% Nonspecific European

A number of others are similar to this that I share with from Britain.

Don Felipe
02-15-2013, 07:08 PM
^From what I have seen so far at least some British and Irish people score very high in the ''Nonspecific Northern European'' category when they are not included in the reference samples.

That's weird, i guess it means that some Brits can be very diverse in their North Euro affinities if they show that little overlap with the Britsh/Irish category on 23&me which is based on 932 samples. Or perhaps those samples just happen to be quite homogeneous.

J Man
02-16-2013, 03:48 PM
^I'm not sure really. All I know is that many Brits seem to score very high in the Nonspecific North European category.

Don Felipe
02-17-2013, 12:02 PM
I guess we'll have to wait for the next update to see if it will remain so. Maybe AC needs some more tweaking. I'm curious how the "overfitted" samples will turn out. Just had a new cousin from the UK accept my sharing invite, she's mostly Cornish on her mother's side. This is her AC at speculative:

http://i989.photobucket.com/albums/af18/oditous2/UK-1.jpg

J Man
02-18-2013, 03:36 AM
Oh I forgot to mention those are my great uncle's results on the Conservative setting. He scores much more British and Irish on Speculative.

Kyte
03-15-2013, 06:26 PM
Can anyone tell me whether features in 23andMe's AC could be under represented? For example, I show a lot more East Asian on all of the Gedmatch analyses, and, on 23andMe's AC I show 18% Middle Eastern (which looks about right), and 67% European (which also seems fair), of which only 21.7% is labelled as 'Balkans', the rest being 'non-specific Southern European' and 'Non-specific European'. With 3 grandparents being from the Balkans for as many generations as are known, it seems odd that I don't have a higher Balkan percentage, in contrast to Romanians and Bulgarians who show much higher Balkans overall. Anyone know if it's possible that my atypical, mixed DNA is leading to the AC not being able to distinguish between much of my DNA? I have almost 14% of my DNA as 'unassigned'.

AJL
03-15-2013, 08:19 PM
Can anyone tell me whether features in 23andMe's AC could be under represented?

Yes, this is quite possible.

Some users still have only a few matches with RF and AF. To compare my own kit to that of Oditous, I have 0.1% coverage from Cape Verde, while he has 0.3% -- and I have no documented ancestors from there, it's probably rather an old Sephardi link from 500 years ago! In other words the common segment is probably not even from an ancestor from CV, but rather from one in Portugal.

So if you're from a highly local population, or if not many people from your exact area have tested, it's quite possible you will have lots of generic "Southern European" segments for example. (Even my mother has lots of generic "Northern European," which is the same phenomenon my cousin J Man noticed.)

Jean M
03-24-2013, 12:58 PM
I think it's more likely your sample was used in the population sample. If you answered your AF questionnaire then most like it was and then of course your DNA is a high match to your own DNA.


Yes I came to that conclusion after the reply on this thread from AJL. If it is any consolation to anyone, the lineages of my four grandparents do trace back to Britain as far as I can go. Admittedly that is not all that far on most lines - 18th-19th century. The surnames so far uncovered are all British except for one Norman, which is likely to be pretty standard. I have no known Welsh lines, which could be a problem for some, if others in the population sample are also short on Welsh.

I have just realised that I do have a Welsh surname among my ancestors. I overlooked it as I did not actually trace the line to Wales. So stupid of me. That makes me even more of a standard Briton.

Bakhash_Aydan
04-26-2013, 12:41 AM
Hi,

My father is an Iraqi jew (Bakhash-Einy), and my mother's ancestry is from the Jews of Djerba (Huri-Aydan). However, my mom's grandpa's earlier ancestry is unknown.

Here are the results:

Standart estimate:

59.9%
Middle Eastern & North African

52.3%
Middle Eastern
3.3%
North African
4.4%
Nonspecific Middle Eastern & North African

15.8%
European

Southern European
11.5%
Italian
1.8%
Nonspecific Southern European
0.6%
Ashkenazi
2.0%
Nonspecific European

24.2%
Unassigned

Conservative estimate:

42.7%
Middle Eastern & North African

39.5%
Middle Eastern
1.4%
North African
1.8%
Nonspecific Middle Eastern & North African

11.4%
European

Southern European
3.8%
Italian
4.2%
Nonspecific Southern European
0.3%
Ashkenazi
3.1%
Nonspecific European

45.9%
Unassigned


Speculative estimate:

73.5%
Middle Eastern & North African

63.4%
Middle Eastern
5.7%
North African
4.3%
Nonspecific Middle Eastern & North African

25.8%
European

Southern European
17.8%
Italian
3.7%
Nonspecific Southern European
1.0%
Ashkenazi
3.3%
Nonspecific European

0.7%
Unassigned


I was pretty amazed by two things:

1. 2000 years people of my mother's side lived in north africa - and no significant genes from that area.
2. The italian ancestry. That's probably explains my mom's father ancestry, he came to the island from italy.

AJL
04-26-2013, 12:45 PM
2. The italian ancestry. That's probably explains my mom's father ancestry, he came to the island from italy.

Pleased to meet you! One thing I would say is that your Italian is about even with your Ashkenazi. However, it's not likely you have much Ashkenazi ancestry, and more likely your shared ancestry with Ashkenazim is Sephardi from the time of the Iberian expulsion (many Ashkenazim probably got a littleSephardi ancestry especially through France, the Netherlands, or Bulgaria/Romania). So the Italian could be actual Italian ancestry or also (1) noise, or (2) shared ancestry that is actually something like Italian Jewish or more Sephardi. Don't forget Calabria and Sicily in particular once had very large Sephardi populations.

Bakhash_Aydan
04-26-2013, 04:18 PM
Pleased to meet you! One thing I would say is that your Italian is about even with your Ashkenazi. However, it's not likely you have much Ashkenazi ancestry, and more likely your shared ancestry with Ashkenazim is Sephardi from the time of the Iberian expulsion (many Ashkenazim probably got a littleSephardi ancestry especially through France, the Netherlands, or Bulgaria/Romania). So the Italian could be actual Italian ancestry or also (1) noise, or (2) shared ancestry that is actually something like Italian Jewish or more Sephardi. Don't forget Calabria and Sicily in particular once had very large Sephardi populations.

The numbers are above the headlines. On the standart estimate, I have 11.5 italian, and 0.6 Ashkenazi.
Looking inside the alleles, the minor ashkenazic segments are within the italian / south-european ones.
Also, there are no duplicated european segments. my X chromosom is marked european. therefore, it's from my mom.

I concluded two things:

A. My mothers family are from the Italian / Sicilian expulsion, and not from spain (no Iberian genes at all, but 11.t italian?)
B. My father's genes are all middle-eastern.

AJL
04-26-2013, 05:01 PM
The numbers are above the headlines. On the standart estimate, I have 11.5 italian, and 0.6 Ashkenazi.


Aha, I see that now -- for some reason I was reading down rather than up. Your conclusions seem reasonable.

I am also curious how many matches you have? I have about 5 matches that are certainly or probably from my Sephardi ancestry (which is about 3.125% of my total ancestry). My father has about 5 more, at 6.25% of his ancestry. However, I would be surprised if it were typical for fully Sephardi or Mizrahi people to get as many as 150 matches, though certainly possible. A fully Ashkenazi person might easily get over 1,000.

Bakhash_Aydan
04-27-2013, 06:33 PM
Aha, I see that now -- for some reason I was reading down rather than up. Your conclusions seem reasonable.

I am also curious how many matches you have? I have about 5 matches that are certainly or probably from my Sephardi ancestry (which is about 3.125% of my total ancestry). My father has about 5 more, at 6.25% of his ancestry. However, I would be surprised if it were typical for fully Sephardi or Mizrahi people to get as many as 150 matches, though certainly possible. A fully Ashkenazi person might easily get over 1,000.

Hi,

I have 341 matches, most of them are Iraqi / Iranian Jews. But, indeed, some are Italian / North-African Jews. True Sephardi Jews (with spanish names) i got only as more distant cousins.

J Man
04-27-2013, 07:41 PM
Are these Ancestry Composition results based off of IBS or IBD segments does anyone know?

Scarlet Ibis
04-27-2013, 07:58 PM
^From what I have seen so far at least some British and Irish people score very high in the ''Nonspecific Northern European'' category when they are not included in the reference samples. My great uncle is 100% North Irish Protestant and here are his results below.

3.8% British and Irish
70.6% Nonspecific Northern European
24.4% Nonspecific European



Thanks for posting this. Do you have his results with the sub-regional populations shown?

J Man
04-28-2013, 03:12 PM
Thanks for posting this. Do you have his results with the sub-regional populations shown?

Yes I do.

Speculative Estimate:

100% European

Northern European

64.6% British and Irish
8.5% French and German
3.0% Scandinavian
20.9% Unspecific Northern European

Southern European

0.6% Iberian
0.6% Nonspecific Southern European
<0.1% Asheknazi
1.8% Nonspecific European


Standard Estimate:

99.9% European

Northern European

39.0% British and Irish
1.1% French and German
0.4% Scandinavian
49.3% Nonspecific Northern European

10.1% Nonspecific European

0.1% Unassigned


Conservative Estimate:

98.8% European

Northern European

3.8% British and Irish
70.6% Nonspecific Northern European
24.4% Nonspecific European

1.2% Unassigned


Interesting to see how his British and Irish score turns into Nonspecific Northern European as the settings are changed from Speculative to Conservative.

Solothurn
04-28-2013, 04:31 PM
Mine too :-)

Speculative

100% European
Northern European
99.3% British and Irish
0.4% Scandinavian
0.2% Nonspecific Northern European
< 0.1% Ashkenazi
< 0.1% Nonspecific European


Standard

100% European
Northern European
98.6% British and Irish
1.3% Nonspecific Northern European
< 0.1% Ashkenazi
0.1% Nonspecific European

Conservative

99.8% European
Northern European
91.4% British and Irish
8.0% Nonspecific Northern European
< 0.1% Ashkenazi
0.4% Nonspecific European


I don't know where Ancestry's get their composition prediction from:

British 61%
Scandinavian 27%
East Euro 12%





Here is my ‘Ancestry Composition’. Accurate, but boring. :\

Speculative
99.8% British and Irish
0.1% Finnish
0.1% Nonspecific Northern European

Standard
99.6% British and Irish
0.4% Nonspecific Northern European

Conservative
95.0% British and Irish
3.7% Nonspecific Northern European
1.2% Nonspecific European

Clinton P

Telfermagne
05-21-2013, 07:35 PM
Speculative estimate, V3 chip:

http://i1161.photobucket.com/albums/q506/sar1227/Screenshot2013-05-21at33429PM_zps8defd725.png (http://s1161.photobucket.com/user/sar1227/media/Screenshot2013-05-21at33429PM_zps8defd725.png.html)

geebee
05-22-2013, 10:24 AM
I don't know where Ancestry's get their composition prediction from

That's one of my questions, too.

My paper trail largest percentages should be German -- close to half; British -- about a quarter; and Spanish -- an eighth. The remaining eighth includes Dutch, Swiss, French, Irish, and Native American (generally thought to be Choctaw).

Ancestry Composition (speculative estimate) comes up with:

97.3% European

Northern European
36.1% British and Irish
9.1% French and German
0.5% Finnish
38.8% Nonspecific Northern European

Southern European
2.2% Italian
1.3% Balkan
1.0% Iberian
4.1% Nonspecific Southern European

0.7% Eastern European

3.4% Nonspecific European

2.0% East Asian & Native American
2.0% Native American
0.1% Nonspecific East Asian & Native American

< 0.1% Sub-Saharan African

0.6% Unassigned

Ancestry, on the other hand, gives:

Central European 48%
Scandinavian 23%
Eastern European 16%
Finnish/Volga-Ural 6%
Uncertain 7%

Some of these might not be too unreasonable. Central European, for example. "Scandinavian" here makes absolutely no sense to me, unless it's because my "British" ancestry is actually mostly from eastern Scotland (and hence greatly influenced by the Vikings). "Eastern European" is also a puzzle, as is "Finnish/Volga-Ural". "Uncertain", of course, leaves room for anything.

J Man
06-12-2013, 12:15 AM
Do people here find Ancestry Composition to be a good analysis and of much use at all? The genome bloggers such as Davidski and Dienekes do not seem to like it much.

AJL
06-12-2013, 12:41 AM
Well, it's not perfect but it does have the advantage of dealing with actual current populations instead of imaginary ancestral components, based (mainly? apparently?) on IBD rather than IBS.

J Man
06-12-2013, 01:10 AM
Well, it's not perfect but it does have the advantage of dealing with actual current populations instead of imaginary ancestral components, based (mainly? apparently?) on IBD rather than IBS.

Ahh I see thanks. I tried sending you a PM about this but for some reason it won't let me. Would you say Ancestry Composition is at least fairly accurate?

AJL
06-12-2013, 01:27 AM
It seems reasonably accurate for me and the kits I manage, but it doesn't of course tell you which way influence goes. My mailbox may be full -- I'll check it!

Esmeralda
07-13-2013, 10:16 AM
I got 0.6% Eastern European, I wonder if it is noise or real :\
http://i.imgur.com/llZsW9x.png

jeanL
07-27-2013, 07:54 PM
My mother's Ancestry Composition results in speculative mode:

581582

The small native and Sub-Saharan isn't a surprise as she does have early colonial ancestry via her dad, now the Middle Eastern(namely Iran, Turkey, Druze) is indeed a surprise.

AJL
07-27-2013, 08:38 PM
My mother's Ancestry Composition results in speculative mode:

581582

The small native and Sub-Saharan isn't a surprise as she does have early colonial ancestry via her dad, now the Middle Eastern(namely Iran, Turkey, Druze) is indeed a surprise.

L2a1a is found in lots of Middle Eastern and North African people -- so it could pertain to your Subsaharan ancestry.

jeanL
07-27-2013, 10:30 PM
L2a1a is found in lots of Middle Eastern and North African people -- so it could pertain to your Subsaharan ancestry.

Well, then thing is my MDKA matrilinearly is from Southern Galicia, Spain.

AJL
07-28-2013, 01:59 AM
Well, then thing is my MDKA matrilinearly is from Southern Galicia, Spain.

Right -- but L2a1a is originally an African haplogroup. Spain does have North African influence, and vice versa.

Joe B
08-19-2013, 11:50 PM
633634635
Time to ask for your opinions, which I will value greatly.
First obvious question is what does the Ashkenazi percentage mean? Is it significant and if so how far up the grandparent tree is that?
Is it possible to determine if the Ashkenazi component is maternal, paternal or both? How good is the 23andme algorithm in Chromosome View when assigning to one side of the chromosome or the other?
Great-great-grandparents known heritage
Maternal side
Grandma German-German and North Italian-Unknown(New York cir. 1883)
Grandpa Irish-Irish and German-German
Paternal Side
Oma German-German and German-German
Unknown Opa(Südhessen 1931)
Also I am interested in what role my y-haplogroup, with all 111 strs, could play. #257842
I've posted the speculative results since the Irish percentage matches up pretty well.
Thanks, Joe

AJL
08-20-2013, 12:10 AM
I'd say there's a good chance you do have some Ashkenazi ancestry. 5% ends up being about 3 at the 64-ancestor level, which I think is 4th great-grandparents, so if we take this literally that could mean a 3rd great-grandparent (born about 1750-1800) was part or mainly Ashkenazi. I suspect that could pertain to your German ancestry.

I doubt if this attaches to your Y line though, both because the odds of receiving any autosomes from your Y line that far back are fairly low, and because if your rare Y group was Ashkenazi, one would have thought other examples would have presented themselves by now.

Joe B
08-20-2013, 03:14 AM
I'd say there's a good chance you do have some Ashkenazi ancestry. 5% ends up being about 3 at the 64-ancestor level, which I think is 4th great-grandparents, so if we take this literally that could mean a 3rd great-grandparent (born about 1750-1800) was part or mainly Ashkenazi. I suspect that could pertain to your German ancestry.

I doubt if this attaches to your Y line though, both because the odds of receiving any autosomes from your Y line that far back are fairly low, and because if your rare Y group was Ashkenazi, one would have thought other examples would have presented themselves by now.
I have fond memories discussing your Aleppo cousins.
Thanks for your quick response. I wanted to absorb what you had to say. And try to figure out that 4th great-grandparent thing. Please know that I'm not trying to be "crypto" anything. Just trying to find some truth that my old man was denied.
My German ancestry is very strong. In fact, one set of maternal great-great-great-grandparents were born only 25km from my dad's hometown, seperated by 117 years.
Are you saying that the Ashkenazi autosomes could not have come from the y-line? Or that the Ashkenazi autosomes are coincidental to my rare Y or mtdna group?
I had to ask about that rare Y-group just because it is so out of place. Sort one way and you've got the UK or Polish, sort another and it's Armenian. That's for another thread.

AJL
08-20-2013, 03:43 AM
Joe:

I'd say it is possible that the Ashkenazi could come from your Y line but I would rather expect with the many, many Eastern European Jewish men having tested, and your not having any near matches, that this is not especially likely. There might be some Jewish Z2103* lurking somewhere but I would imagine probably Mizrahi or Sephardi (hardly tested at all) rather than Ashkenazi. As you say, your Y seems to be part of a very early branching and is quite an enigma.

The Ashkenazi influence could be coming from any line, is I guess my point.

geebee
08-20-2013, 12:55 PM
Joe,

If I read the numbers right, 5.5% looks more like a 2nd great grandparent than it does a 4th great grandparent -- especially if you also consider the 1.8% "Eastern European".

You always get 50% from each parent -- well, close enough to 50%, though if you're male you actually inherit a teeny bit more from your mother than from your father. On average, you get 25% from each grandparent, 12.5% from each great grandparent, 6.25% from each 2nd great grandparent, and 3.125% from each 3rd great grandparent.

You can see that 5.5%, if all from one person, is much closer to 6.25% than it is to 3.125%. And I would suggest that your Y line 2nd great grandparent probably did pass along some autosomal DNA to you.

Unfortunately, even if you limit consideration to 2nd great grandparents, there are still four men and four women. Only one of those men is on your Y line, and only one of those women is on your mtDNA line.

EDIT: Okay, so I took a closer look at AJ's earlier reply. I think he's saying 5% is about like saying 3 out of the 64 great grandparents. But a 4th great grandparent contributes an average of 1.625%, which times 3 is 4.875%. Now, I know I'm probably pushing the numbers a little too hard ... but you can see that this is less than 5%, and your number is actually ever higher -- 5.5%.

All I'm really trying to say is that I would put the hypothetical "mainly or partly Ashkenazi" at least one generation closer. That would mean perhaps 1800-1850, or even a bit later. That may not change much, but if you're trying to see what may have been going on in the world at that time, it can be helpful.

AJL
08-20-2013, 03:30 PM
Yes, geebee's scenario is within the realm of possibility too. At least for Eastern Europe this fairly small change in time can make a huge difference as to whether records are obtainable.

Joe B
08-20-2013, 08:20 PM
If I read the numbers right, 5.5% looks more like a 2nd great grandparent than it does a 4th great grandparent -- especially if you also consider the 1.8% "Eastern European".
All I'm really trying to say is that I would put the hypothetical "mainly or partly Ashkenazi" at least one generation closer. That would mean perhaps 1800-1850, or even a bit later. That may not change much, but if you're trying to see what may have been going on in the world at that time, it can be helpful.

Geebee,
Thank you for your insight. The 1.8% Eastern European associated with the 5.5% Ashkenazi is something I've been pondering too. It would make a lot of sense, especially considering the interaction between the southern half of the Rhine-Main area and the Pale.
Looking at conservative, standard and speculative percentages, the Ashkenazi numbers are the most solid. UK/Ireland, Italy and Germany/France vary the most.
Ashkenazi 4.8 - 5.2 - 5.5%
Eastern European null - 0.8 - 1.8%

Some of records go back to the 1500s on the maternal side in Germany and paternal Oma's go back a long time in some of the small towns of the Odenwald. Most likely the AK would be from my father’s father (grandpa) or from the unknown great-great grandpa from New York on the maternal side. Slight, slight possibility would be the Asti Jews of Italy, another group that needs to be studied.

The “what level” grandparent question is important because of the one Jewish grandparent rule, the Nuremberg Laws and the Mischling Test.

geebee
08-21-2013, 08:07 AM
Joe

When you look at Ancestry Composition, be sure to also consider Chromosome View. There are obviously different ways to arrive at a given percentage of any ancestry.

For example, both my maternal grandmother and her father were half Spanish. But he was half Spanish because his father was Spanish and his mother was Alsatian. His daughter was half Spanish because both her parents were half. The difference is that if somehow "Spanish" and "Alsatian" always showed up clearly in Ancestry Composition -- which, unfortunately, they do not -- my great grandfather would therefore show Spanish across the one half of all of his chromosomes in AC, and Alsatian across the other half. My grandmother's chromosome view would show just as much Spanish overall, but in a more broken-up fashion.

Since you're looking at someone probably no closer than a 2nd great grandparent, this may not mean as much. But my expectation is that if AC were more precise, you might be able to distinguish between a single contribution from a 2nd great grandparent versus a double contribution from 3rd great grandparents, and so on. What you'd be considering would be the size of the segments, as well as their number and total percentage.

Of course, the precision of AC is a limiting factoring. How distinct the ancestral populations are from each other is another limiting factor. (So "French and German" might not always be fully discernable from "English and Irish", and eventually just get lumped together as "Northern European".) In addition, I believe some "smoothing" is done, so very small segments will disappear rather quickly.

EDIT: Looking back at an earlier post of yours, I see that you posted Chromosome View and have obviously been considering what it might tell you. I'm afraid I can't tell you how good a job 23andMe does at parsing the chromosomes in the absence of data from the parents.

But again, I want to emphasize the percentages can be quite deceiving. I already mentioned the case of my grandmother and her father, with the same percentage of Spanish ancestry but derived two different ways. My father is about half German and half British -- but so was his father, and his father's father. The end result is that the two ethnicities -- not so dissimilar to start with -- are pretty thoroughly intertwined.

Since, as I understand it, Ancestry Composition looks at a person's DNA results block-by-block, the more distinctive a particular block may be, the more likely to be assigned to a given ethnicity. But a single block that is itself admixed isn't likely to be identified except in generic way (such as "northern European" or "southern European").

Hume
09-03-2013, 10:32 AM
667


This is very surprising to me on several fronts. First, I only was aware of a single, very distant English Jew ancestor from the early 1800's. But it looks like both of my parents are roughly half Jewish, based on the fact that I am half Ashkenazi, and my single X chromosome is about 40% Ashkenazi.

My expected results from family tree research, v. actual

German/French 12.5% v. 22.6%
British 25% v. 1.6%
Eastern Euro 50% (25% Polish/25% Slovak) v. 11.1%
Balkan/Southern Euro (Croatian) 12.5% v. 2.7%
Ashkenazi 0.5% v 50.9%

It is very hard to explain at first glance outside of some affairs or adoptions! I know I have one great grandparent whose parents are both English immigrants, and another great grandparent who is from Canadian with two English-surnamed parents (Ellwood and Carter).

Have one of my parents be conversos is one thing, but both of them apparently are!

I also can't explain why my German % is higher than expected. AFAIK, my german ancestry is confined to one of my grandmothers.

And of the 35+ surnames I have in my family tree, only one, from very far back, is recognizably Jewish!

AJL
09-03-2013, 01:03 PM
It would help to test both your parents as well, though I suspect you should probably have a conversation with them first about your results.

Hume
09-03-2013, 01:10 PM
Looking at my chromosomes a bit more, it looks like it may be more like my mother is ~30% and father is ~70% Ashkenazi. Most of my chromosomes have one mostly jewish strand and one non-Jewish or lightly jewish strand. Also my Y haplogroup is Ashkenazi, g2c*.

The other notable items are that I have 0.0% non-euro, not even a little like so many whites, and my non-specified European is also on the low side. My chart pasted above is speculative, but all three versions are pretty similar.

Overall very surprising! I don't look especially Jewish, more central European features than anything like my Slovak grandfather, though I do have dark hair that is slightly curly.

My Slovak ancestors came to the USA via Ellis Island with $15 in their pockets, were illiterate, worked in rural Pennsylvania coal mines when they arrived, and list their city of origin as a small town in northern Slovakia that currently has 8,000 people. Not really the profile you think of when you think of Jewish central European immigrants from the 1910's, who'd you expect to be from cities, be literate, not be nearly broke, and do white-collar or factory work.

Nor did my ancestors on either side move to any of the traditional heavily Jewish cities when they arrived in the USA, instead having patterns more typical of German/Eastern Euro/late-arriving English, settling in rural Pennsylvania and catholic immigrant neighborhoods of Pittsburgh and Detroit.

Hume
09-03-2013, 01:19 PM
It would help to test both your parents as well, though I suspect you should probably have a conversation with them first about your results.

Yeah I don't know how I'd broach the subject. "Mom it seems like you used a sperm donor, since my genetic test says my father is mostly Jewish, and Dad isn't!" My Dad I understand to be 50% polish, 25% croat, and 25% English/Scots-Irish. It is possible that both the poles and croats were all conversos who changed their names to something that doesn't sound Jewish. Unlike Poland however, Croatia never had a large jewish population, before WWII it was about 0.5% jewish per Wikipedia.

AJL
09-03-2013, 10:47 PM
Yeah I don't know how I'd broach the subject. "Mom it seems like you used a sperm donor, since my genetic test says my father is mostly Jewish, and Dad isn't!" My Dad I understand to be 50% polish, 25% croat, and 25% English/Scots-Irish. It is possible that both the poles and croats were all conversos who changed their names to something that doesn't sound Jewish. Unlike Poland however, Croatia never had a large jewish population, before WWII it was about 0.5% jewish per Wikipedia.

I was thinking along the lines of "Mom, so I took a DNA test and I am having trouble understanding the results." I suppose there is some chance that that may actually elicit some admission from her (lack of forthcomingness about paper trail? adoption? suspected hospital mixup? infidelity?).

It may be that your father's Polish ancestry is entirely Jewish, plus your mother has some Jewish ancestry. Croatia, as you say though, never had much of an Ashkenazi community. My grandfather, who shows up as almost entirely Ashkenazi, does have one match in Croatia, but it is way at the bottom of his CoA list significantly below Uzbekistan, Iran, Iraq, and Syria, none of which every had Ashkenazi settlement. In short, it seems unlikely to me that Croatia has masses of Ashkenazi genetic material floating around in it.

jeanL
09-05-2013, 01:24 PM
Anyone here get Sardinian on their AC? What I found interesting is that my paternal grandfather gets Sardinian in all modes.

Conservative: Sardinian 0.5%, Italian 1.7%

Standard: Sardinian 0.6%, Italian 5.0%

Speculative: Sardinian 0.8%, Italian 9.4%

What I find interesting is that my grandfather does have what appears to be distant Sardinian ancestry. Namely his paternal grandfather's maternal grandmother's surname was Sardiñas, which is of Sardinian origin. So to me it is awesome that 23andme is able to pick that up.

On a different subject matter, anyone knows where in Italy is the surname "Forte" more common?

MitchellSince1893
09-05-2013, 02:14 PM
I have 0% conservative, .2% standard, .6% speculative on Sardinian. No known Sardinian or Italian ancestry.

Dr_McNinja
09-05-2013, 09:21 PM
I had 98% South Asian, 1% European on 23andMe's AC. For comparison, DNA Tribes said 92% South Asian and 7% European using the same 23andMe data with them.

Scarlet Ibis
09-05-2013, 10:22 PM
On a different subject matter, anyone knows where in Italy is the surname "Forte" more common?


According to gens.info, it seems to have a pretty broad distribution.

http://www.gens.info/lib/cog/maps/cognomi/F/FORTE.gif

AJL
09-05-2013, 10:52 PM
Anyone here get Sardinian on their AC? What I found interesting is that my paternal grandfather gets Sardinian in all modes.

Conservative: Sardinian 0.5%, Italian 1.7%

Standard: Sardinian 0.6%, Italian 5.0%

Speculative: Sardinian 0.8%, Italian 9.4%

What I find interesting is that my grandfather does have what appears to be distant Sardinian ancestry. Namely his paternal grandfather's maternal grandmother's surname was Sardiñas, which is of Sardinian origin. So to me it is awesome that 23andme is able to pick that up.

On a different subject matter, anyone knows where in Italy is the surname "Forte" more common?

1.2% Italian (from my mother's side), no Sardinian. Is it possible you have Corsican ancestry too?

This link (http://www.cognomix.it/mappe-dei-cognomi-italiani/) suggests Forte peaks in Campania and Lazio.

{EDIT}
My maternal aunt shows no Italian but does show <0.1% Sardinian. This makes me a little suspicious of those very low level matches.

ADW_1981
09-06-2013, 05:19 PM
Hume, what is your mtDNA? My father's profile has a recent G2c* cousin and I am wondering if it's you.
Thanks

Mac von Frankfurt
09-24-2013, 02:10 PM
My sample has reached the lab. :horn:

It will be interesting to see if they pick up any of the alleged 3/128ths Amerindian.

I'm also curious about the Y-DNA and mtDNA from Frankfurt.

Be back in a month.B)

Mac von Frankfurt
10-15-2013, 02:27 PM
Well the results for my wife came back first.

99.6% European

93.6% Northern European

17.8% British and Irish
2.5% Scandinavian
1.6% French and German
0.9% Finnish

Pretty consistent with her colonial North American ancestry.

Her mtDNA is V3. Which explains her love of all things Christmas. I am checking zoning regulations to see if we can have Reindeer on the property. (Yes I know V3 is believed to have originated on the Iberian peninsula and her mtDNA is likely not from Finland but I can't tell her she is from Dogerland because she is an SF Giants fan).

Scarlet Ibis
10-15-2013, 07:57 PM
Well the results for my wife came back first.

99.6% European

93.6% Northern European

17.8% British and Irish
2.5% Scandinavian
1.6% French and German
0.9% Finnish

Pretty consistent with her colonial North American ancestry.

Her mtDNA is V3. Which explains her love of all things Christmas. I am checking zoning regulations to see if we can have Reindeer on the property. (Yes I know V3 is believed to have originated on the Iberian peninsula and her mtDNA is likely not from Finland but I can't tell her she is from Dogerland because she is an SF Giants fan).

Was that on the "standard" or "speculative" setting?

Mac von Frankfurt
10-16-2013, 04:15 AM
The above post is standard setting. Speculative is:

99.7 European

50.1 British and Irish
14.2 French and German
6.6 Scandinavian
1.1 Finnish
27.0 Nonspecific Northern European
0.2 Nonspecific Southern European
0.6 Nonspecific European
0.1 Sub-Saharan Africa
0.2 Unassigned

jeanL
11-03-2013, 06:08 PM
Could a moderator/administrator look into why I can't edit my post above? I would greatly appreciate it!!

AJL
11-03-2013, 06:17 PM
After 24 hours, posts become locked and you have to PM an admin or moderator, who can edit posts for you.

Isidro
11-03-2013, 08:04 PM
On Conservative, my Chromosome Painting Nonspecific Southern European and Nonspecific European always lays on the edges of each chromosome, is that a common occurrence?.

867

Yudi
11-10-2013, 06:07 AM
Here are my results, typical for Colonial American ancestry. They are phased using one parent. Interestingly, the African segment never showed up until I got my parent genotyped.
890891

AJL
11-10-2013, 04:26 PM
I have to disagree slightly. 0.4% South Asian is not typical of American ancestry without known Roma ancestry (probably in the 18th century or so).

Yudi
11-10-2013, 07:38 PM
Yes, that was a surprise; we have no family story of any connection with the Roma. I inherited that segment from my mother. I want to get her mother (the only one alive in my grandparents' generation) genotyped to see if she has it. My theory is that the segment has traveled along with my mtDNA haplogroup. R0a2 is quite unusual in Europeans, whereas it is common in the Middle East, where the Roma are known to have traveled. Although haplogroup R seems to have come into Europe with the Unetice culture, so it's possible that it's been among us for a long time. It's always so hard to know with haplogroups. In any case, I would love to solve the mystery of our Roma connection, but it will probably be next to impossible.

AJL
11-10-2013, 10:59 PM
I agree -- I don't think with my mother's similar low level of South Asian I will be able to pinpoint an exact ancestor, but I have narrowed it down at least to a specific lineage through comparing multiple cousins who tested. Because our common ancestors along that line were Palatine Germans, it seems most likely that there is a German Sinti connection before 1710.

Dr_McNinja
11-17-2013, 01:44 AM
Did 23andMe change the colors in their Ancestry Composition? What else did they change?

Dr_McNinja
11-19-2013, 02:57 AM
My mom's 23andMe ancestry composition:

97.4% South Asian
< 0.1% European
< 0.1% Nonspecific European
< 0.1% Middle Eastern & North African
< 0.1% Middle Eastern
2.5% Unassigned


And in speculative mode:

98.4% South Asian
0.7% Middle Eastern & North African
0.7% Middle Eastern
0.6% European
0.6% Nonspecific European
0.3% UnassignedThey changed my Ancestry Composition results. It used to be 98.x% South Asian and in speculative mode it was 98.9% South Asian and 1.1% Non-Specific European.

Now it's:

96.6% South Asian
0.1% East Asian & Native American
East Asian
0.1% Nonspecific East Asian
0.1% European
0.1% Nonspecific European
3.1% Unassigned

and in Speculative mode:

98.7% South Asian
0.5% European
Northern European
0.1% Nonspecific Northern European
0.4% Nonspecific European
0.1% East Asian & Native American
East Asian
0.1% Nonspecific East Asian
0.7% Unassigned

My Chromosome 15 used to have a bunch of European marks (and in Gedmatch's one-to-many results, and 23andMe's Countries of Ancestry tool, I get a lot of big chunks matching Europeans or Americans of European descent on Chr 15) and now it's all South Asian even though my mom's Chr 15 is still mostly European/Mideastern. Which is weird I guess.

Here's the split view:

http://i.imgur.com/GvVFlHL.png

Tolan
11-19-2013, 11:38 PM
Here are my results.
All my ancestors, for at least 10 generations, are french, and come from a small area at the borders of Brittany, Maine and Normandy.

A few small surprises:
- In speculative, more British than French / German, maybe a problem for 23andme to differentiate the two? (Standard: 5.6% British 0% French / German)
- Iberia 11% (3.3% in standard). Perhaps the ancient people of my region?
- 1.1% Scandinavia: It is true that the Vikings have settled in the region. (0.7% in standard)
- Italian 0.9% (0% in standard): no explanation!
- 0.1% East asian. Appears only on speculative. Big Interogation! The Uns?

943

krotx
11-20-2013, 01:18 PM
945
Seems accurate enough on speculative, I'm getting around 51% Nonspecific European on standard estimates though.

ADW_1981
11-20-2013, 04:14 PM
Here are my results.
All my ancestors, for at least 10 generations, are french, and come from a small area at the borders of Brittany, Maine and Normandy.

A few small surprises:
- In speculative, more British than French / German, maybe a problem for 23andme to differentiate the two? (Standard: 5.6% British 0% French / German)
- Iberia 11% (3.3% in standard). Perhaps the ancient people of my region?
- 1.1% Scandinavia: It is true that the Vikings have settled in the region. (0.7% in standard)
- Italian 0.9% (0% in standard): no explanation!
- 0.1% East asian. Appears only on speculative. Big Interogation! The Uns?



The combined land mass of France + Germany + Belgium + Netherlands is huge, so I would not expect it to capture all the variation nor should it. Since Britain, and ultimately Ireland were sourced from the region your ancestors came from, I'm not surprised that you are more like them than you are - potentially a northern German or eastern German (who source the French German component as well). The biggest surprise is the 11% Iberian. That's pretty significant and surprising.

Tolan
11-21-2013, 08:45 PM
The combined land mass of France + Germany + Belgium + Netherlands is huge, so I would not expect it to capture all the variation nor should it. Since Britain, and ultimately Ireland were sourced from the region your ancestors came from, I'm not surprised that you are more like them than you are - potentially a northern German or eastern German (who source the French German component as well). The biggest surprise is the 11% Iberian. That's pretty significant and surprising.


Thank you for your comment!

It is likely that the countries of Europe which have a central position in Europe, as France, Germany and other countries of Central Europe, have fewer specific markers, as can have outlying countries countries as British Isles, Scandinavia or Iberian Peninsula...

Dr_McNinja
11-23-2013, 08:32 PM
945
Seems accurate enough on speculative, I'm getting around 51% Nonspecific European on standard estimates though.You are from Uzbekistan I take it? Have you done admixture with the Harappa and Eurogenes calculators?

vettor
11-23-2013, 08:44 PM
what is the " geographical border" between being a Northern-European and a Southern-European in 23andme?

I ask because I am noted as Southern European but my X chromosome is 97% Northern European.

I have 33% southern European of which 31% is Italian, 1% iberian and have 22% of French-German ............I thought ancestry was greatly affected by the X chromosome?

MitchellSince1893
11-23-2013, 08:59 PM
Iberia, Sardinia, Italy, and the Balkans are in Southern Europe.

British Isles, France, Germany, Low Countries, Scandinavia, Finland, Switzerland are Northern Europe.

Poland, Czech, Slovakia, Austria, Former Soviet Republics in Europe are Eastern Europe.

Corsica isn't colored coded for any group.

Joe B
11-23-2013, 10:15 PM
what is the " geographical border" between being a Northern-European and a Southern-European in 23andme?

I ask because I am noted as Southern European but my X chromosome is 97% Northern European.

I have 33% southern European of which 31% is Italian, 1% iberian and have 22% of French-German ............I thought ancestry was greatly affected by the X chromosome?
Noticed something similar with my 23andme X chromosome result. ~97% Northern European with none of the ~12.5% Italian/S. European that is expected. Gedcom gives a variety of Mediterranean possibilities for the X chomosome.
Is the X weaker for admixture composition in general?

Il Papà
11-24-2013, 02:17 AM
Here are my results.
All my ancestors, for at least 10 generations, are french, and come from a small area at the borders of Brittany, Maine and Normandy.

A few small surprises:
- In speculative, more British than French / German, maybe a problem for 23andme to differentiate the two? (Standard: 5.6% British 0% French / German)
- Iberia 11% (3.3% in standard). Perhaps the ancient people of my region?
- 1.1% Scandinavia: It is true that the Vikings have settled in the region. (0.7% in standard)
- Italian 0.9% (0% in standard): no explanation!
- 0.1% East asian. Appears only on speculative. Big Interogation! The Uns?

943
salut :),what's your dodecad k12b results ?

Tolan
11-24-2013, 08:56 AM
salut :),what's your dodecad k12b results ?

I downloaded dodecad K12B today and I could get something (Non sans mal!).

38.26% Atlantic Mediterranean
37.87% North European
13.04% Caucasus
7.8 Gedrosia
2.05% Southwest_Asian
1.48% Northwest_African
0.02% South_Asian

Difference from a French reference: i am
- 6% Atlantic Mediterranean (While my ancestors are in the west of France)
+1,5% North European
+5% Caucasus
= Gedrosia
= Southwest_Asian
+1,5% Northwest_African
= South_Asian


But there is a big difference between the detail of chromosomes!
Nothing similar with 23andme!
I am more confident with 23andMe than dodecad!

Il Papà
11-25-2013, 02:04 PM
I downloaded dodecad K12B today and I could get something (Non sans mal!).

38.26% Atlantic Mediterranean
37.87% North European
13.04% Caucasus
7.8 Gedrosia
2.05% Southwest_Asian
1.48% Northwest_African
0.02% South_Asian

Difference from a French reference: i am
- 6% Atlantic Mediterranean (While my ancestors are in the west of France)
+1,5% North European
+5% Caucasus
= Gedrosia
= Southwest_Asian
+1,5% Northwest_African
= South_Asian


But there is a big difference between the detail of chromosomes!
Nothing similar with 23andme!
I am more confident with 23andMe than dodecad!

Yeah your dodecad result is approximately the same as my grandfather one who is from ardennes,yours are even a bit more southern,you're fully french from northwest France right ?.

The french sample used in dodecad seems biased,they seems too northern influenced as we can see looking at the supposed average french Caucasus score(8% for the dodecad reference) compared to mine, yours and some of my family members that are in between 13-18% .

my grandfather results:
Gedrosia 4.67%
Siberian 0.67%
Northwest_African 1.72%
Southeast_Asian -
Atlantic_Med 39.69%
North_European 38.32%
South_Asian 0.97%
East_African -
Southwest_Asian 0.66%
East_Asian -
Caucasus 13.30%
Sub_Saharan -

AJL
11-25-2013, 03:38 PM
Yeah your dodecad result is approximately the same as my grandfather one who is from ardennes,yours are even a bit more southern,you're fully french from northwest France right ?.

The french sample used in dodecad seems biased,they seems too northern influenced

The French sample is probably based mainly on Québécois, who typically have much of their ancestry from Normandy, Île-de-France, and Charente-Maritime, and who also usually have a bit of Irish, Scots, or English ancestry. This might explain a bit of deviation in the French average.

Il Papà
11-25-2013, 05:03 PM
The French sample is probably based mainly on Québécois, who typically have much of their ancestry from Normandy, Île-de-France, and Charente-Maritime, and who also usually have a bit of Irish, Scots, or English ancestry. This might explain a bit of deviation in the French average.

Well the guy Tolan is also northern french and score "much south" than the Quebecois so I think that it's the quebecoi's Foreign ancestry (Irish,english,etc) that is far more likely to explain the gap between the french reference and the "true french" than the Quebecois being of northern french stock.

ADW_1981
11-25-2013, 06:13 PM
Well the guy Tolan is also northern french and score "much south" than the Quebecois so I think that it's the quebecoi's Foreign ancestry (Irish,english,etc) that is far more likely to explain the gap between the french reference and the "true french" than the Quebecois being of northern french stock.

If anything, Tolan's ancestry seems to be more eastern, towards Italy as per the Caucasus component. A question would be does he have any Italian ancestry? What is "True French"? You will find variation depending on where your ancestors came from within the country.

AJL
11-25-2013, 06:58 PM
If anything, Tolan's ancestry seems to be more eastern, towards Italy as per the Caucasus component. A question would be does he have any Italian ancestry? What is "True French"? You will find variation depending on where your ancestors came from within the country.

Yes, it might work best to run the DNA in mixed mode for 2 populations in this case and see what comes up for French and a second population, in what proportion.

Québécois are not a perfect proxy for all France but neither do they cluster in the UK: it's more that the typical Québécois has one or two lineages that are Irish or Scottish a couple hundred years ago. And then Québécois often have a perceptible trace of Amerinidan ancestry too so they should be east of the average French population in one aspect.

Il Papà
11-25-2013, 08:22 PM
If anything, Tolan's ancestry seems to be more eastern, towards Italy as per the Caucasus component. A question would be does he have any Italian ancestry?

not really.

Here are my results.
All my ancestors, for at least 10 generations, are french, and come from a small area at the borders of Brittany, Maine and Normandy.


[B] What is "True French"? You will find variation depending on where your ancestors came from within the country.

Yeah ofc ,but for someone who is considered northern french, scoring more southern than the average french is weird, if he was southern french ,that would be more logical but in the actual case not really.

AJL
11-25-2013, 08:50 PM
Could be migration from Italy a long time ago. For example I have ancestors near Caen named de Vassal de la Tourette, who some scholars think were orignally Torretta, from Italy. (Incidentally fron this line I am very very distant cousins with Antoine de Saint-Exupéry.)

Tolan
11-25-2013, 09:20 PM
What surprises me in dodecad 12k, is that I am more in the east than in the West, that the French average!
Results with dodecad 12kA, are even more mysterious to me:
45.92% West European
30.94% Mediterranean
11.46% East European
8.71% West Asian
1.92% Southwest Asian
1.59% Northwest African
0.06% South Asia

At 23andMe, 0% East European. I have a large share of Iberian in speculative mode (11%).

I do not have Italian ancestors, far from it! And although I would have an unknown branch, this would represent a small percentage!
I got back all my branches since 1700, and you can see that my ancestors have not moved!:
http://gw.geneanet.org/jerome4?lang=en&carto=1

ADW_1981
11-25-2013, 09:42 PM
not really.




Yeah ofc ,but for someone who is considered northern french, scoring more southern than the average french is weird, if he was southern french ,that would be more logical but in the actual case not really.

I'm skeptical that anyone can trace every ancestor for 10 generations. Scottish and English have had minimal impact on Quebecois. The bottom line is that English and Scottish aren't even Catholic, so the likelihood of them skewing results is unlikely, as marriages between the two Christian groups was minimal, with the exception of maybe Huguenots with the typically Protestant groups. There might be Irish impact on Quebecois though.

Generalissimo
11-25-2013, 11:42 PM
What surprises me in dodecad 12k, is that I am more in the east than in the West, that the French average!

Can you check whether you're also more eastern than the French average in the Eurogenes tests, like the latest one?

http://bga101.blogspot.com.au/2013/11/updated-eurogenes-k13-at-gedmatch.html

Joe B
11-29-2013, 06:07 PM
23andme expanded Ancestry Composition to 31 populations. When did this happen?




Ancestry Composition tells you what percent of your DNA comes from each of 31 populations worldwide. The analysis includes DNA you received from all of your ancestors, on both sides of your family. The results reflect where your ancestors lived 500 years ago, before ocean-crossing ships and airplanes came on the scene.
European

Northern European
French & German
British & Irish
Scandinavian
Finnish
Nonspecific Northern European

Southern European
Italian
Balkan
Sardinian
Iberian
Nonspecific Southern European
Ashkenazi
Eastern European
Nonspecific European

East Asian & Native American
East Asian
Japanese
Korean
Yakut
Mongolian
Chinese
Nonspecific East Asian
Southeast Asian
Native American
Nonspecific East Asian & Native American

Middle Eastern & North African
Middle Eastern
North African
Nonspecific Middle Eastern & North African

Sub-Saharan African
West African
East African
Central & South African
Nonspecific Sub-Saharan African

South Asian

Oceanian

Unassigned

vettor
11-29-2013, 11:36 PM
23andme expanded Ancestry Composition to 31 populations. When did this happen?

986987988

i don't get the name as per your links...............where is this setting?

what does the 500 years mean............the data as at the year 1500/1510?

AJL
11-30-2013, 12:22 AM
23andme expanded Ancestry Composition to 31 populations. When did this happen?

986987988

Not sure, mine haven't changed much in a while.

Joe B
11-30-2013, 12:26 AM
Ancestry Composition Updates (https://customercare.23andme.com/entries/23244235-Ancestry-Composition-Updates-November-2013): November 2013 Receiving updated results


In order to receive results that include the new populations, profiles will be enqueued for computation on the updated version of Ancestry Composition. These computations will occur on a rolling basis, so there is no specific timeline for receiving these updated results. Some customers will see these results before others.

Until these computes are run, the new populations will appear in your list with a question mark icon indicating that your profile is "not yet computed" for this population:
https://customercare.23andme.com/attachments/token/aaltaqk5qaubdbi/?name=Screen+Shot+2013-11-15+at+11.21.55+AM.png

23andMe Makes More Improvements to Ancestry Features
(http://blog.23andme.com/ancestry/23andme-makes-more-improvements-to-ancestry-features/) November 19, 2013

Customers should start seeing the update to their Ancestry Composition over the next several weeks. It will offer people with African or Asian ancestry greater detail than they had before. We added several additional reference populations — for a total of 31.
http://blog.23andme.com/wp-content/uploads/2013/11/Asian-Ancestry.jpghttp://blog.23andme.com/wp-content/uploads/2013/11/African-American-Example1.jpg


i don't get the name as per your links...............where is this setting?

what does the 500 years mean............the data as at the year 1500/1510?Those links are just the three resolution levels found in the 23andme Ancestral Composition page. Standard is the default level. View choice of map or chromosome.
The 500 years ago thing is just to get peoples minds off of modern borders. Maybe somebody else has a better answer for that.

Tolan
11-30-2013, 07:55 AM
i don't get the name as per your links...............where is this setting?

what does the 500 years mean............the data as at the year 1500/1510?

I think that 23andMe, not searching the ethnicity of Europeans, unlike, for example, genographic, which will give for a Britishman: 50% Northern European, 33% Mediterranean and 17% Southwest asian.
23andMe does not seek the origin of the british, but will send the result to the client: 100% British for a British man

Generalissimo
11-30-2013, 09:53 AM
23andMe does not seek the origin of the british, but will send the result to the client: 100% British for a British man

This is only possible if this person is overfitted based on his reported ancestry, but that problem has been corrected, and new results will be sent out over the next few weeks. So no one will score 100% British at 23andMe once that happens.

vettor
12-01-2013, 09:54 AM
Ancestry Composition Updates (https://customercare.23andme.com/entries/23244235-Ancestry-Composition-Updates-November-2013): November 2013 Receiving updated results


23andMe Makes More Improvements to Ancestry Features
(http://blog.23andme.com/ancestry/23andme-makes-more-improvements-to-ancestry-features/) November 19, 2013


Those links are just the three resolution levels found in the 23andme Ancestral Composition page. Standard is the default level. View choice of map or chromosome.
The 500 years ago thing is just to get peoples minds off of modern borders. Maybe somebody else has a better answer for that.

It will not do me any good, I am nearly 100% european...the Finnish is already in for me................the africans and asians numbers...good for others

Telfermagne
12-05-2013, 03:54 AM
Got updated results and they are similar to what was yielded by the 4-gp Oracle (both Dodecad & Eurogenes) - the main difference being the expansion beyond merely 4 populations, so a more refined result. It took them long enough. :P

Dodecad: Argyll + Argyll + North_Italian + Swedish @ 1.682.
Eurogenes: German + Irish + Irish + Southwest_French @ 1.804

http://i1161.photobucket.com/albums/q506/sar1227/Screenshot2013-12-04at104805PM_zps6268c4ee.png (http://s1161.photobucket.com/user/sar1227/media/Screenshot2013-12-04at104805PM_zps6268c4ee.png.html)

vettor
12-05-2013, 08:54 AM
the new composition got worse, they butchered Italian group and gave the bulk to middle-east group.
they butchered french/german group and made it non-specific

this is stated in many forums

ADW_1981
12-05-2013, 03:03 PM
the new composition got worse, they butchered Italian group and gave the bulk to middle-east group.
they butchered french/german group and made it non-specific

this is stated in many forums

Because you don't personally agree with the results, does not mean they are incorrect. Italians do have a Middle-East portion that north Europeans don't have. Deal with it.

newtoboard
12-05-2013, 03:09 PM
Because you don't personally agree with the results, does not mean they are incorrect. Italians do have a Middle-East portion that north Europeans don't have. Deal with it.

So Northern Europeans are free of Middle Eastern components in your opinion?

ADW_1981
12-05-2013, 03:13 PM
So Northern Europeans are free of Middle Eastern components in your opinion?

Nah, didn't say that.
Italians are about 2x the West Asian that north Europeans are though (see the old Dodecad runs), and have chunks of SW Asian that simply don't exist in northern Europe in some areas. It's the reality of the situation, it doesn't make it better or worse.

So if his new Ancestry Composition shows Middle-East, even 2-3%, it's probably real and those mutations likely don't exist in Northern Europe for the most part.

newtoboard
12-05-2013, 03:24 PM
Nah, didn't say that.
Italians are about 2x the West Asian that north Europeans are though (see the old Dodecad runs), and have chunks of SW Asian that simply don't exist in northern Europe in some areas. It's the reality of the situation, it doesn't make it better or worse.

So if his new Ancestry Composition shows Middle-East, even 2-3%, it's probably real and those mutations likely don't exist in Northern Europe for the most part.

Yes that all seems correct. Do you think this is because the Middle Eastern influence on Northern Europeans is actually via Anatolia, the Caucasus or Central Asia (if the theories on R1a and R1b being born there are correct) while Southern Europeans have some chunks of Levantine ancestry as shown by the SW Asian component you mentioned? Or because Southern Europeans received these components first (people have speculated on West Asia having a recent shift to being more Gedrosia and Caucasus like)?

ADW_1981
12-05-2013, 03:36 PM
Yes that all seems correct. Do you think this is because the Middle Eastern influence on Northern Europeans is actually via Anatolia, the Caucasus or Central Asia (if the theories on R1a and R1b being born there are correct) while Southern Europeans have some chunks of Levantine ancestry as shown by the SW Asian component you mentioned? Or because Southern Europeans received these components first (people have speculated on West Asia having a recent shift to being more Gedrosia and Caucasus like)?

One of the possibilities is that farming essentially collapsed in central Europe and locals may have carried on the traditions. If that's the case then many of those West Asian genes wouldn't have been carried forward to subsquent generations. It seems farmers were at bare minimum G2a and subclades and this seems to have arrived from Anatolia if we corroborate the evidence. If farmers weren't "dying" so to speak in Italy, that means they would have carried forward some additional West Asian input.

vettor
12-05-2013, 05:35 PM
Nah, didn't say that.
Italians are about 2x the West Asian that north Europeans are though (see the old Dodecad runs), and have chunks of SW Asian that simply don't exist in northern Europe in some areas. It's the reality of the situation, it doesn't make it better or worse.

So if his new Ancestry Composition shows Middle-East, even 2-3%, it's probably real and those mutations likely don't exist in Northern Europe for the most part.

your wrong,

my Italian went from 31% to 9%, my german from 22% to 2 %, my ME from 0% to 1%, my non-specific from 29% to 77% , my african and asian stayed at 0%...................is it improved!

All I read from majority of people is that :
1- 23andme numbers for this composition tests barely grew
2 - Italian group was butchered and given to ME
3- Iberian group butchered and given to north aftrican

I can deal with it and have zero issue with what I am, but I cannot deal with a system which says its better and then gives NON results!

And Lastly, 23andme state the results are only UP to 500 years old, so basically from the year 1500AD to present ..............did this ME migration happen in bulk in these last 500 years!

Tolan
12-05-2013, 06:07 PM
your wrong,


And Lastly, 23andme state the results are only UP to 500 years old, so basically from the year 1500AD to present ..............did this ME migration happen in bulk in these last 500 years!

Yes, that is (or was) makes the results of 23andMe interesting and different from other...

ADW_1981
12-05-2013, 06:07 PM
your wrong,

my Italian went from 31% to 9%, my german from 22% to 2 %, my ME from 0% to 1%, my non-specific from 29% to 77% , my african and asian stayed at 0%...................is it improved!

All I read from majority of people is that :
1- 23andme numbers for this composition tests barely grew
2 - Italian group was butchered and given to ME
3- Iberian group butchered and given to north aftrican

I can deal with it and have zero issue with what I am, but I cannot deal with a system which says its better and then gives NON results!

And Lastly, 23andme state the results are only UP to 500 years old, so basically from the year 1500AD to present ..............did this ME migration happen in bulk in these last 500 years!

Well no, I'm not wrong. My point is that the Middle-Eastern part is likely correct. Have you put your settings to Speculative? It's a known fact that Standard and Conservative aren't giving great results, and large chunks of ancestry are aligning to non-specific categories.

vettor
12-05-2013, 06:29 PM
Well no, I'm not wrong. My point is that the Middle-Eastern part is likely correct. Have you put your settings to Speculative? It's a known fact that Standard and Conservative aren't giving great results, and large chunks of ancestry are aligning to non-specific categories.

I will try what you say............but you are confusing because in other posts you deal with farmers and hunters from very ancient times..............the 23andme information is only from 500 years ago

Ancestry Composition tells you what percent of your DNA comes from each of 31 populations worldwide. The analysis includes DNA you received from all of your ancestors, on both sides of your family. The results reflect where your ancestors lived 500 years ago, before ocean-crossing ships and airplanes came on the scene.

99.4% European ------SPECULATIVE results

Southern European
26.5% Italian
3.4% Balkan
3.2% Iberian
0.4% Sardinian
18.6% Nonspecific Southern European

Northern European
20.1% French & German
1.6% British & Irish
16.8% Nonspecific Northern European

0.8% Ashkenazi
0.1% Eastern European
7.9% Nonspecific European

0.5% Middle Eastern & North African
0.3% North African
0.1% Nonspecific Middle Eastern & North African

< 0.1% East Asian & Native American
< 0.1% Nonspecific East Asian & Native American
0.1% Unassigned


somewhat better, but still 44% on non-specific for last 500 Years ( maybe some alpine -refuge area !!!!:amen:)

Where is it documented that I need to use speculative or is this your opinion?

MfA
12-05-2013, 06:50 PM
Do the math and see how much autosomal DNA you can inherit from >500 years ago.. 500/25=20 generations, 100%/2^20=0,000095367431640625%..

multifariousm
12-05-2013, 07:07 PM
As an adoptee, I have almost no family information. How sound does the following look? I was initially skeptical about the Ashkenazi, but it mapped to self declared Ashkenazi 3rd-distant cousins.


Standard view:

27.0% British & Irish
13.1% French & German
0.6% Scandinavian
25.5% Nonspecific Northern European
26.5% Ashkenazi
2.5% Nonspecific Southern European
4.5% Nonspecific European
0.1% East Asian & Native American

evon
12-05-2013, 07:11 PM
I think the new AC is better then the last in being more correct, but also less specific. So basically it was a step backwards from the last AC due to mistakes made by over-fitting, but this is better overall, because many whom used to be 100% are now more divided among the various components, such as my uncle whom used to be 98% Scandinavian is now around 22% Scandinavian, and i used to be 86% Scandinavian and now i am 55%..in reality the numbers should even be lower then that, and the replacement for this is nonspecific Northern European, which is more correct.

But the 500 year label should be dropped, as its not realistic given what we see, most people show connections that are much older then this (such as the common Yakut connection in Europeans, which is likely 1000's of years old), my own family show ancient ties to India which is via Roma migrations around 1000CE, so its well over the 500 mark..

vettor
12-05-2013, 08:11 PM
But the 500 year label should be dropped, as its not realistic given what we see, most people show connections that are much older then this (such as the common Yakut connection in Europeans, which is likely 1000's of years old), my own family show ancient ties to India which is via Roma migrations around 1000CE, so its well over the 500 mark..

after re-reading the text from 23andme , it now means IMO from a point in time 500 years ago, so its the 1500 to 1510.

for my part of ancestry it could reflect the Italian wars which lasted 55 years , from 1494 onwards. with armies coming from aragon, catalan, france, swiss, bavaria, tyrol, hapsburg austrian races, and all of italy. throw in the baggage train with its camp followers of women and children ..........and ..........it could be nonspecific

ADW_1981
12-05-2013, 08:13 PM
Do the math and see how much autosomal DNA you can inherit from >500 years ago.. 500/25=20 generations, 100%/2^20=0,000095367431640625%..

That's when you're looking at the contribution from a single ancestor 500 years ago. Of course, this far back you have 1000's of contributors to your DNA. If the bulk of those are Italians, you will have a greater likelihood of inheriting West Asian chunks (aka Middle Eastern)

0.5% Middle Eastern & North African
0.3% North African
0.1% Nonspecific Middle Eastern & North African

This seems consistent of a north Italian result. It would be odd to find a Scottish person to have a 0.9% contribution in this category. So yes, we can say there is a division.

ADW_1981
12-05-2013, 08:16 PM
after re-reading the text from 23andme , it now means IMO from a point in time 500 years ago, so its the 1500 to 1510.

for my part of ancestry it could reflect the Italian wars which lasted 55 years , from 1494 onwards. with armies coming from aragon, catalan, france, swiss, bavaria, tyrol, hapsburg austrian races, and all of italy. throw in the baggage train with its camp followers of women and children ..........and ..........it could be nonspecific

On Conservative my family has only 0.4% British/Irish, the rest non-specific, so yes there is a problem with the algorithm. Moving up to standard it's only about 10% British/Irish, the rest non-specific. From what I recall, the French/German component has a low recall rate in the old version, and I think some of the people used as reference samples are still getting a lot of non-specific results.

Joe B
12-05-2013, 11:29 PM
Here are the before and after Chromosome Speculative Views at the sub-regional resolution.
Old AC 1014 New AC 1015
The conservative view is too conservative. Ashkenazi is highest at 0.6% in conservative view compared to 4.7% and 6.1% in standard and speculative views respectively. Looks like some of the Irish/UK was made Scandinavian. Iberian is new. It and the Nonspecific Southern European increased at the Italian composition's expense.

Dr_McNinja
12-06-2013, 03:04 AM
They changed my results again:


Speculative

98.4% South Asian
1.1% Middle Eastern
0.2% East Asian and Native American
0.1% European
0.2% Unassigned

---

Standard

95.8% South Asian
0.6% Middle Eastern
3.6% Unassigned

-------------------------

Mother @ Speculative (seems unchanged)

98.4% South Asian
0.7% Middle Eastern
0.6% European
0.3% UnassignedMy original AC was 98.7% South Asian, 1.1% European or something like that. I know my Chr 15 has European on it because of GEDmatch and 23andMe's own Countries of Ancestry tool, and it was listed as European in the first AC and in my mom's AC, but now they switched the whole thing to Mideastern which is just weird. I don't think their AC is worth anything, it seems like they alter it with no methodology, just to fudge things over so they seem more sensible to casual users under the pretense of "last 500 years".

Tolan
12-06-2013, 04:06 AM
Well no, I'm not wrong. My point is that the Middle-Eastern part is likely correct. Have you put your settings to Speculative? It's a known fact that Standard and Conservative aren't giving great results, and large chunks of ancestry are aligning to non-specific categories.

Maybe, it is because 23 andme are more honest than the others!:)

Many SNP are not specific to a particular region of Europe. It is not because they are more common in a country, that necessarily your ancestors come from this one! 23 andme propose three levels of certainty, corresponds in probability (more 90 %, more 75 % and more 50 %)

AJL
12-06-2013, 04:27 AM
This is very interesting. Ancestry Composition now has a small Yakut as well as nonspecific Asian percentage for me, which is inline with Northeast Asian influence showing in other analyses but all trace of which was previously missing at 23andme.

My grandfather now also has a little Finnish as well as Yakut; my British & Irish have gone up to more-or-less expected levels of 36.1% from a startingly low 10-ish; and my aunt's very questionable Sardinian has disappeared and become nonspecific Southern European. I wouldn't say it's perfect but it still seems quite a good tool and even excellent when compared to Population Finder, say.

AJL
12-06-2013, 04:57 AM
As an adoptee, I have almost no family information. How sound does the following look?

Looks like one parent was colonial Caucasian American, the other was half Ashkenazi and possibly half Dutch (or something like that).

Dr_McNinja
12-06-2013, 05:01 AM
I wouldn't say it's perfect but it still seems quite a good tool and even excellent when compared to Population Finder, say.It depends I guess. Population Finder had a very reasonable result for me, more in line with the other admixture calculators as well as Dr. McDonald's analysis which is probably the best. 23andMe's is way off all of them for South Asians.

AJL
12-06-2013, 05:23 AM
It depends I guess. Population Finder had a very reasonable result for me, more in line with the other admixture calculators as well as Dr. McDonald's analysis which is probably the best. 23andMe's is way off all of them for South Asians.

Possibly so -- the database is tilted heavily away from South Asia. Which is unfortunate for me as well because my mother and several of her maternal cousins all have a little South Asian (we think it is a German Sinti ancestor but are not sure).

Scarlet Ibis
12-06-2013, 07:43 AM
This is very interesting. Ancestry Composition now has a small Yakut as well as nonspecific Asian percentage for me, which is inline with Northeast Asian influence showing in other analyses but all trace of which was previously missing at 23andme.

My grandfather now also has a little Finnish as well as Yakut; my British & Irish have gone up to more-or-less expected levels of 36.1% from a startingly low 10-ish; and my aunt's very questionable Sardinian has disappeared and become nonspecific Southern European. I wouldn't say it's perfect but it still seems quite a good tool and even excellent when compared to Population Finder, say.

Were you on the V3 chip?

evon
12-06-2013, 10:12 AM
after re-reading the text from 23andme , it now means IMO from a point in time 500 years ago, so its the 1500 to 1510.

for my part of ancestry it could reflect the Italian wars which lasted 55 years , from 1494 onwards. with armies coming from aragon, catalan, france, swiss, bavaria, tyrol, hapsburg austrian races, and all of italy. throw in the baggage train with its camp followers of women and children ..........and ..........it could be nonspecific

Unless you have decent proof to back that up, that is just wild speculation, and in my experience people tend to overemphasis "cool" ancestry, such as those associated with wars and known historical battles and ethnicity, which is usually false and based on the same biased that was rolled out from Historians such as Leopond Von Ranke...

evon
12-06-2013, 10:14 AM
Possibly so -- the database is tilted heavily away from South Asia. Which is unfortunate for me as well because my mother and several of her maternal cousins all have a little South Asian (we think it is a German Sinti ancestor but are not sure).

If you have Roma ancestry its an easy way to check it, basically look for matches with known Roma, as they share a very small genepool, so you should match at least one other with similar ancestry.

Táltos
12-06-2013, 03:03 PM
I have to agree with AJL that 23andme's AC is much better than Population Finder. However I am not impressed with this upgrade. I tested my mother here, and really I felt that her initial results were not too bad from them. I do like the other tools they offer though like countries of ancestry and global similarity map. Anyway here are her original results.

Conservative:
Italian-7.7%
Balkan-0.7%
Nonspecific Southern Euro-10.3%
Nonspecific Northern Euro-6.0%
Nonspecific European-69.4%
Unassigned-6.0%(If I didn't already think the 69.4% above this one was unacceptable, I found this one even worse.)

I won't bother with Standard to save space.

Old Speculative:
Italian-36.4%
Balkan-5.4%
Iberian-1.0%
Nonspecific Southern Euro-17.9%
French and German-9.6%
British and Irish-4.9%
Nonspecific Northern Euro-14.1%
Eastern European-2.0%
Nonspecific European-8.7%
(No Unassigned here!)

New AC Conservative
Italian-2.3%
Nonspecific Southern Euro-9.3%
Nonspecific Northern Euro-1.2%
Nonspecific Euro-80.5%
Unassigned-6.7%
Clearly this is worse than the first conservative mode result. Once again I will skip Standard to save space. One thing I will add on Wed night when I noticed the change my Mom did have East Asian/Native American listed all of a sudden in Standard, but on Thursday it disappeared from the Standard mode.

New Speculative AC
Italian-31.9%
Balkan-15.2%
Iberian-0.7%
Sardinian-0.1%
Nonspecific Southern Euro-17.0%
British and Irish-8.1%
French and German-2.2%(What happened here!)
Nonspecific Northern Euro-15.9%
Nonspecific European-8.7%
East Asian and Native American-0.1%
Unassigned-0.3%>:(

And this new East Asian/Native American shows on three of my Mother's chromosomes one is identified as Yakut (which I was reading on the 23andme forum seems to be the typical Asian group that Europeans are getting). On another chromosome it is listed as Nonspecific East Asian, and on the third one that it shows it is Nonspecific Native American. Really though we do have autosomal matches with Native American ancestry, both showing in 23andme before and after the upgrade, and also having paper trail to Wyandot, I don't believe that this is real. On Gedmatch my Mom will show noisy percents of Siberian or Amerindian at say .20 off the top of my head. My data and my paternal phased data would show higher amounts than that. My father was very Eastern European.

Anyway for those who may recognize me from another forum sorry for this rehash, but those who don't here is my Mom's known ancestry. Half Arbëresh (Albanian/Southern Italian) from her father, on her mother's side she is Colonial, German and Dutch. There are lines that went back into France and Switzerland that are connected to her extensive German lines. Also the other thing that really bothers me with this new update is her X chromosome painting. The whole bottom half is Italian, Almost 75% of the upper half is Balkan! There is a little bit that is Nonspecific European. Her mother's contribution to her X looks almost nonexistent. Either I am really way off on knowing what my mother's mother line is (at least more recently) or this AC painting is way off.

AJL
12-06-2013, 04:01 PM
Were you on the V3 chip?

V2 + V3 combined.

AJL
12-06-2013, 04:02 PM
If you have Roma ancestry its an easy way to check it, basically look for matches with known Roma, as they share a very small genepool, so you should match at least one other with similar ancestry.

In fact my mother does match to a person with Hungarian Roma ancestry but they also have German ancestry so it is not conclusive.

evon
12-06-2013, 06:26 PM
In fact my mother does match to a person with Hungarian Roma ancestry but they also have German ancestry so it is not conclusive.

If you want i can likely solve it for you, i share with a good number of Roma over at 23andme, some whom i am related to, but also others...let me know if you want to share...

evon
12-06-2013, 06:27 PM
My grandmothers came in, now i am hoping the split-view for my uncle will update his results more, her results seem more true then last time, and as i though, she lost the east Asian in speculative and gained more south Asian:

Standard:
http://imageshack.com/a/img202/2998/lhdy.jpg

Speculative:
http://imageshack.com/a/img809/6328/wt8a.jpg

Contrasted with her Population finder result:

http://imageshack.com/a/img9/3309/0sci.jpg

Now i am just waiting for my aunt..

AJL
12-06-2013, 06:38 PM
If you want i can likely solve it for you, i share with a good number of Roma over at 23andme, some whom i am related to, but also others...let me know if you want to share...

Thank you, PM sent.

ADW_1981
12-06-2013, 07:55 PM
My profile data hasn't phased yet properly but both the v2 and v3 AC versions have come in for my parents. This is both on Speculative mode, and still have a lot of non-specific north European ancestry. Any other mode gives very little information.

Father's v3
48.1%British & Irish
6.3%Scandinavian
5.8%French & German
38.5%Nonspecific Northern European
0.4%Nonspecific Southern European
0.9%Nonspecific European

Father's v2
49.3%British & Irish
9.7%French & German
6.2%Scandinavian
33.4%Nonspecific Northern European
0.5%Nonspecific Southern European
0.9%Nonspecific European

--------------------
Father's top 5 AF 4gp

UK 3.8%
Netherlands 1.2%
Germany 0.8%
Ireland 0.7%
Italy 0.4%


Mother's v2

48.7%British & Irish
15.1%French & German
6.1%Scandinavian
27.7%Nonspecific Northern European
0.2% Iberian
< 0.1%Nonspecific Southern European
1.4%Nonspecific European
0.7%West African
0.1%Nonspecific Sub-Saharan African
< 0.1%Unassigned

Mother's v3
48.5%British & Irish
16.3%French & German
5.9%Scandinavian
26.5%Nonspecific Northern European
0.2%Iberian
1.0%Nonspecific Southern European
0.8%Nonspecific European
0.8%West African
< 0.1%Nonspecific Sub-Saharan African
< 0.1%Unassigned
------------------
Mother's top 5 AF 4gp

UK 5.0%
Ireland 2.1%
Germany 1.4%
Norway 0.9%
Sweden 0.8%

MitchellSince1893
12-06-2013, 10:30 PM
On 23andme's ancestry composition I've become:

More British and Irish and Scandinavian.

Less French and German, Southern European, and South Asian.

The changes were most dramatic in the Scandinavian (4 time more) and Southern European (83% less). But both are very small overall percentages.

My ancestry composition mode old and new in speculative mode:
British & Irish: From 52.6% to 64.3%
French & German: From 6.3% to 4.1%
Scandinavian: From .6% to 2.6%
Finnish: From 1.1% to 1.0%
Non specific N. Euro: From 37% to 26.1%
Southern Euro: From 1.7% to .2%
Non Specific Euro: From 1.2% to .4%
South Asian: From .5% to .3%

jeanL
12-07-2013, 12:12 AM
All of the people in my family updated already, and while my split view is showing my Dad's isn't showing yet, so I don't know what gives??

Things that I found inconsistent:

AC-2 gives me 0.1% South Asian which according to the slip view I get it from my mother, yet in her chromosome 7 there isn’t any South Asian, nor is there any unassigned, but then again 23andme says anything under 1% could be noise.

AC-2 gives me a huge chunk of Middle Eastern which accounts for 0.9% of my genome, this is supposed to be coming from my dad, yet my dad chromosome 8 shows only a bit of Nonspecific North Africa/Middle Eastern, which even if added the chunks of unassigned next to it, doesn’t add up anywhere to the length of my segment. It seems the algorithm is assigning some of what assigned in my dad as nonspecific Southern Europe as Middle Eastern in me. There are also a couple of chunks(<0.1%) of Nonspecific East Asian/Native built in there which are supposed to be coming from my Dad, yet in that region he only shows Italian, and Nonspecific Southern European. There also seems to be a chunk(~0.1%) of SSA assigned to my chromosome 12 which is supposed to be coming from my father's side, yet he has no SSA in his chromosome 12.

Overall there seems to be some bugs that I wish 23andme will look into, and I have written to them many times about it, unfortunately, when one add the parents a lot of these inconsistencies arise, at least that has been my experience.

I have a feeling my AC will likely change after my dad's parents get readded, so I'll wait for it before I post it.

MitchellSince1893
12-07-2013, 12:46 AM
My split view isn't working either.

For my father's AC: His Middle East/N African ancestry is now gone and his South Ancestry has increase slightly. This may be a more accurate reading as I know he has India ancestry and am unaware of any ME or NA ancestry.

Isidro
12-07-2013, 01:34 AM
I just got mine today, definitely a big change, it resembles a few of my Gedmatch tests but I get lot of unassigned which is new for me.

AJL
12-07-2013, 03:18 AM
My split view isn't working either.


My split view has never worked properly, either before or after.

Boudicca
12-07-2013, 10:27 PM
Does anyone know how AC can distinguish a difference between Austria/Switzerland in the French/German category and Hungary/Czech Republic in the Eastern European category? My Mum still gets no Eastern European but her French/German has increased despite getting many matches over the 7cM threshold to Hungary :\

Wulf Talented
12-07-2013, 11:29 PM
Since this is the official thread, i'll post mine here as well as in my other thread.

Speculative,
1024

Standard
1025

I was quite interested that 23andme do Ashkenazi ancestry, having a Great Grandfather ancestor from Poland I was expecting to see something, but clearly not. I do know that he was a catholic, so not declared Jewish but still could have had ancestors.

Interesting about the 0.1% North African, on conservative view it goes from to >0.1%. Is that noise at all?

Joe B
12-08-2013, 03:07 AM
Does anyone know how AC can distinguish a difference between Austria/Switzerland in the French/German category and Hungary/Czech Republic in the Eastern European category? My Mum still gets no Eastern European but her French/German has increased despite getting many matches over the 7cM threshold to Hungary :\
Somehow they figure it out from the reference populations. Is your Mum Hungarian German or Danube Swabian? That might confuse the 23andme algorithm?

Next time you are in the Ancestry Composition page, put your cursor over French & German. Click that little arrow that points right. Then click details. The reference populations should appear.
:beerchug:

French & German


Connected to the British Isles, Scandinavia, southern Europe and eastern Europe, France and Germany have seen myriad peoples come and go over the last ten thousand years. Genetically and geographically the French and Germans are at the heart of Europe.
Population Source Sample Size
Germany 23andMe 367
Netherlands 23andMe 207
France 23andMe 200
Switzerland 23andMe 87
Belgium 23andMe 80
Austria 23andMe 54
France HGDP 29
Eastern European


Eastern Europe, represented by people of Ukraine, Russia, Poland and Hungary, is bordered on the east by the Ural Mountains. Although there are no such geographic borders to the west, eastern Europe has been distinct from European countries to the west in terms of cultural and linguistic affiliations.
Population Source Sample Size
Russia 23andMe 253
Poland 23andMe 240
Hungary 23andMe 86
Ukraine 23andMe 82
Slovakia 23andMe 45
Czech Republic 23andMe 44
Belarus 23andMe 31
Russian HGDP 25
Slovenia 23andMe 20

geebee
12-08-2013, 05:09 AM
My father is now getting 2.9% "Southern European", which is more than any of my siblings or I do. The reason that seems a bit odd is that he has no known southern European ancestry, but only German and British, while my mother was 1/4 Spanish. Both her mother's father and her mother's mother had fathers who immigrated directly from the Spanish island of Menorca.

Or to put it differently, my siblings and I should have the equivalent of one Spanish great grandparent - though it's actually two Spanish 2nd great grandparents. So we should have an average of 12.5% if all this ancestry showed as "southern European".

Curiously, in speculative my father shows 7.3% total southern European, which is more than the equivalent of having a southern European 2nd great grandparent, I only show 5.2%, which is less than 2/3 of what I used to show. My father's "southern European" is broken down in this estimate as including 2.7% Iberian and 1.8% Italian. Mine includes only 1.4% Iberian and 1.0% Italian.

But here'a an interesting detail. Probably over 2/3 of my father's "Italian" is on his X chromosome. In fact, it seems to occupy a bit more than half the chromosome. So this would tell us it's an inheritance from his mother -- though from their it could trace to either of her parents.

Here's the thing, though. I have two sisters who have now been tested. They are my father's daughters, so one of the X chromosomes of each of them should be identical to our father's. While both sisters have both Iberian and Italian in their Ancestry Composition (speculative), neither one shows Italian on the X chromosome.

RGM
12-08-2013, 08:00 AM
I have tested both parents, both grandmothers, and an uncle from each side. Results for my mom's mother and brother only came through recently. I haven't been following these AC updates, but I just noticed that the split view for myself, my dad, and my dad's brother no longer work. However my mom's now works and her profile has clearly been updated.

I'll have to go back and find the old numbers, but it appears that in standard mode, nonspecific values dropped significantly throughout my family. My heritage is Irish across the board, with a hint of Scottish on my mom's side, and it looks like everyone's "British and Irish" standard results jumped 10-20 percent. I think my standard value used to be 59 percent, now it's 79.

It all looks about right, nearly all British and Irish with trace Scandinavian, French and German, Iberian. The only strange thing is that in speculative mode only, my grandmother and uncle now show 0.1 percent Native American. That segment is unassigned in standard. My mom also shows 0.1 percent Finnish in speculative mode, is that a new category?

Here are the current standard values for tested members of my family:

Me

100% European

79.5% British & Irish
15.3% Nonspecific Northern European
5.3% Nonspecific European
< 0.1% Unassigned

Mom - All lines traced back to Ireland or Scotland

100% European

76.9% British & Irish
2.6% French & German
18.5% Nonspecific Northern European
0.1% Sardinian
0.1% Nonspecific Southern European
1.7% Nonspecific European
< 0.1% Unassigned

Dad - both parents born in Ireland

99.9% European

84.9% British & Irish
13.6% Nonspecific Northern European
1.5% Nonspecific European
0.1% Unassigned

Maternal Grandmother

99.8% European

80.2% British & Irish
17.7% Nonspecific Northern European
1.9% Nonspecific European
0.2% Unassigned

Paternal Grandmother

99.8% European

84.3% British & Irish
0.2% Scandinavian
14.4% Nonspecific Northern European
1.0% Nonspecific European
< 0.1% Middle Eastern & North African
< 0.1% North African
0.1% Unassigned

Maternal Uncle

99.8% European

77.3% British & Irish
18.8% Nonspecific Northern European
0.9% Iberian
0.3% Nonspecific Southern European
2.4% Nonspecific European
0.2% Unassigned

Paternal Uncle

99.8% European

89.4% British & Irish
9.3% Nonspecific Northern European
1.1% Nonspecific European
0.1% Middle Eastern & North African
0.1% North African
0.2% Unassigned

geebee
12-08-2013, 01:00 PM
Here are the old and new "Southern European" estimates for me:

http://i1205.photobucket.com/albums/bb429/gjsb56/AC-spec-SouthernEuro-G_zps0caa514d.jpg (http://s1205.photobucket.com/user/gjsb56/media/AC-spec-SouthernEuro-G_zps0caa514d.jpg.html)http://i1205.photobucket.com/albums/bb429/gjsb56/SouthernEuro-Gary-spec_zpsa868b989.jpg (http://s1205.photobucket.com/user/gjsb56/media/SouthernEuro-Gary-spec_zpsa868b989.jpg.html)

You might note that chromosome 3 has gone from having been identified as almost completely Southern European, to not being Southern European at all. There is a tiny segment remaining, but that is represented as being from the opposite chromosome. Split view is not currently working, but the bottom half of my chromosomes previously had reflected inheritance from my father.

Chromosome 4 is not changed as much -- but now includes a region between the two segments previously identified as "Southern European".

Chromosome 8 was likewise previously mostly SE, but has nearly vanished -- again, the little now being reported in on the opposite chromosome.

Chromosome 11 has increased its SE identification, but this is again depicted on the opposite chromosome. Unless my father were to have recent "southern European" ancestry, you would not expect long segments. Long segments show less recombination, which might be believable from, say, a grandmother to grandson, but not for more generations. My maternal grandmother was half Spanish. None of my father's grandparents nor great grandparents was Southern European of any kind, so any such ancestry on his side would be much further back than for me.

Chomosome 13 has picked up some Southern European, which is actually reasonable. It's located in proximity to Native American ancestry, and both of these can be expected to have been passed down to my grandmother from her mother, and from my grandmother through my mother, to me.

The same thing applies to chromosome 15, which has also picked up Southern European (in association with NA, which was already previously identified).

Chromosome 16, on the other hand, was previously mostly Southern European and now has no SE at all.

Chromosome 17 has picked up Southern European, but all on the opposite chromosome.

Chromosome 18 has lost all that it previously had on one chromosome -- presumably the one from my 1/4 Spanish mother -- but picked up a little bit on the other chromosome -- the one from my German/Scots-Irish father.

A segment of Southern European ancestry has also disappeared from chromosome 22.

My guess would be that AC is now picking something up as "Southern European" that is actually perhaps a component of "British" ancestry, while at the same time discounting what was genuinely southern European. I fail to see how this is an "improvement".

Táltos
12-08-2013, 03:37 PM
@geebee yes I am most disappointed in these new results, and my mother's X chromosome is not accurate either. I have to say I have always liked 23andme better for ancestry as my kit at FTDNA leaves a lot to be desired in way of Ancestry tools. I also have to admit to myself now, for as much as I did not like FTDNA not having chromosome painting provided. Before this update, and now after, I have NEVER been able to match up an ancestry painting for my Mom and her matches. For example she will have three matches in the same spot of a chromosome, they will all be matches to each other. They would all show a different ancestry in these spots on their painting. One would be Nonspecific European, one would be British and Irish, another would be French and German, and my Mom in that spot would be Italian! At least when Dr. McDonald did my chromosome painting (which thank goodness he was able to provide this for so many) my chromosome 7 was painted half Middle Eastern, and this is where all my Jewish matches matched me. Oh well let's see if they try to improve what they did. I'm realizing that their chromosome painting before and after this actually left a lot of question marks for me.

Isidro
12-08-2013, 04:00 PM
The change in my results by themselves could be credible, although after seeing my son and daughter's I say they are not accurate, not even close by any stretch of the imagination.

Mine on standard went from 99.9% Iberian to:

42% Iberian
14.2% nonspecific Southern Euro
6.9% nonspecific Northern Euro
36.5% nonspecific Euro
0.4% Unassigned
Total 100%

These results could make sense for me, AFAIK to the 1700' ; all my ancestors are from Spain. Looking at my son and daughter's results makes me wonder though.

Daughter is 8.95% Iberian and son is 2.1% Iberian.

geebee
12-08-2013, 04:02 PM
Among other things it should be interesting once they have split view working again. Right now I just get a message saying that Split View isn't available for me, because "To enable Split View, you must have at least one genotyped parent connected to your Family Tree. Because it requires recomputation, Split View will become active within a few days once a parent is connected to your tree." I DO have a genotyped parent connected to my Family Tree, and have had for a couple of years: my father.

But my daughter gets the same message, and both her mother and I have been genotyped at 23andMe and are connected to her tree (and have been), which is why I'm assuming Split View just hasn't been recalculated yet.

When it is, I think it will confirm that I apparently inherited almost no Southern European DNA from my 1/4 Spanish mother, and most of what I DID inherit is from my German-British father. This, of course, makes no sense, but may illustrate that one can "overcorrect", which is what I think 23andMe has done here.

It's too bad I can't have DNA testing done for either my mother or maternal grandmother. My mother would have been fascinated by the technology ... though perhaps vaguely grossed out by the whole "spitting thing". :-) Ironically, I think her mother would not have had that reaction, but would only be amused to learn the things a person can discover from a tube of saliva.

It was interesting to watch them interact, though also sometimes painful. My mother married as a teenager, in part to get away from home. And she never seemed to be able to quite get past feeling like a teenager when around her mother.

For someone not taller than maybe 4" 10", let's just say my grandmother was ... "strong willed".

Hanna
12-08-2013, 06:08 PM
My results really changed, before I was 100% Middle eastern and now I am scoring 9.2% European as well as 1.2% Asian.

My updated AC

http://imageshack.com/a/img706/445/7jd4.png

Old AC

http://img233.imageshack.us/img233/6246/map23andme.png

My aunt's updated AC

http://imageshack.com/a/img51/2962/aux0.png

My aunt's old AC

http://img844.imageshack.us/img844/9297/28943708.png

My grandfather's updated AC

http://imageshack.com/a/img801/3093/wuhm.png

My grandfather's old AC

https://imageshack.com/a/img43/4372/c4a.png

Dr_McNinja
12-08-2013, 11:07 PM
They changed my results again:

Speculative

98.4% South Asian
1.1% Middle Eastern
0.2% East Asian and Native American
0.1% European
0.2% Unassigned

---

Standard

95.8% South Asian
0.6% Middle Eastern
3.6% Unassigned

-------------------------

Mother @ Speculative (seems unchanged)

98.4% South Asian
0.7% Middle Eastern
0.6% European
0.3% Unassigned
My original AC was 98.7% South Asian, 1.1% European or something like that. I know my Chr 15 has European on it because of GEDmatch and 23andMe's own Countries of Ancestry tool, and it was listed as European in the first AC and in my mom's AC, but now they switched the whole thing to Mideastern which is just weird. I don't think their AC is worth anything, it seems like they alter it with no methodology, just to fudge things over so they seem more sensible to casual users under the pretense of "last 500 years".They changed my mother's:


99.1%
South Asian
0.7%
European
0.5%
Northern European
0.2%
Nonspecific European
0.1%
East Asian & Native American
0.1%
East Asian
< 0.1%
Nonspecific East Asian & Native American
0.1%
Unassigned... Yeah, makes no sense compared to the changes they made to mine. They changed her Chr 15 to all South Asian this time.

Scarlet Ibis
12-09-2013, 04:22 AM
To be 100% honest, it seems like it was a bit of a downgrade for many people on my share list who are of 1 ethnicity. It pushed a great bit of their ancestry back into the "nonspecific" categories, even on the speculative setting. And I know they WEREN'T used as reference samples before, because they were using pseudonyms, and never filled out their "where are you from?" questions.

Still, though, it's not terrible. I applaud their adding more specific populations for non-europeans. Plus, it's good on a regional scale, which is all I've come to expect of the genetic testing companies. It will take them a while to catch up with the specificity of user-made tools by Polako, Dienekes, Magnus Ducatus, etc. And I suspect many will be pleased with the new results, because seeing something new pop up is more interesting than a vanilla result, like 98% British & Irish, 98% East Asian, or whatever else they used to have.

Anyway, here are mine.


http://oi40.tinypic.com/e654r4.jpg


Just for fun, here's what I got from AncestryDNA.

http://oi44.tinypic.com/2gxe23r.jpg

Scarlet Ibis
12-09-2013, 04:26 AM
Also, what's funny (to me, anyway) is that my mom consistently gets a small bit of 0.1% noise in her 23andme results that always showed up as "Middle Eastern" before. Now, under the new Ancestry Composition, it shows up on her, and my results as Ashkenazi.

I know it's just noise, and there's no reason to suspect any recent Middle Eastern or Ashkenazi movements into Korea, but it still amuses me, and makes my imagination run wild sometimes when I see it.