PDA

View Full Version : Genetic Ethnicity



Ian B
01-30-2013, 10:03 AM
I saw a British TV program once in which a persons YDNA was broken down into countries of origin, e.g. Italy 5%, Greece 20%, Polish 15%, English 30% etc., etc.
Was this correct? If so, how does one establish proportionate DNA Ethnicity?

DMXX
01-30-2013, 11:58 AM
DNA ethnicity - as you've termed it - tends to be approximated through autosomal DNA (chromosomes 1-22, which you inherited from both parents roughly equally).

The same cannot be done with Y-DNA and/or mtDNA because (depending on the haplogroup or subclade) it originated from one person at one point in time from one place.

The presenter might have meant the frequencies of that person's Y-DNA across Europe rather than proportioned areas of origin?

geebee
01-30-2013, 03:03 PM
Yes, you really can't estimate ancestry from Y-DNA is most cases. You may be able to say, to some degree, that this haplogroup is more associated with this part of the world than with that, especially as you consider an ever-increasing number of markers. For example, my haplogroup of R1a1a1b1a1b (or R-L1029) is typically associated with Central Europe. But that doesn't necessarily mean that's where it originated. As DMXX said, a SNP defining a particular haplogroup -- in this case L1029 -- originated in one man at a particular time and place. For L1029, possibly what is now Poland; or possibly what's now the Czech Republic or Slovakia. Who knows for sure? [But I am happy to note that it was in my own "Walk Through the Y" that this particular marker was discovered.:) ]

In a few cases you might be able to approximate a person's ancestry if a marker is very associated with a given locale, and if the person's overall ancestry -- not just the "Y line" -- is all from that place. Otherwise, someone who is 99% European, for example, might nevertheless end up with a "Native American" Y haplogroup, for example.

Ian B
01-31-2013, 06:40 AM
geebee: The explanation of your ancestry contains the tpe of information that I'm trying to find - i.e. "... my ancestry is German-British-Catalan, but it actually includes smaller amounts of French, Irish, Dutch, Swiss, and Choctaw (and possibly also Mohawk" )-How did you find this out?

AJL
02-01-2013, 02:46 AM
Good question: you have to look at the autosomes for this, which you get as a mixture from both parents. This is something you can probably best find out at the moment through a mix of Ancestry Composition (at 23andme), and, if you know it, your paper trail. For people with a recent adoption in the family, it is obviously much harder.

My own Y line traces most likely to southern Russia. My mtDNA line traces to Central Europe. My Ancestry Composition only reflects Eastern Europe, not Germany:

42.7% Ashkenazi (consistent with my father being ~7/8 Ashkenazi)
39.7% Nonspecific European
8.7% Nonspecific Northern European
5.9% British and Irish
1.5% Nonspecific Southern European
0.7% Eastern European
0.4% Iberian
0.3% Scandinavian
0.1% Unassigned

You will note there's no "French + German" -- their category -- but my mother has a whole bunch of French and German (about 5%) in her Ancestry Composition, much of which (about 3.5%) showed up in mine until a few days ago. I suspect the phasing engine is broken at the moment.

Additionally, my paternal grandfather has a small amount of Subsaharan African, while my mother has a small amount of South Asian. This might come from a Romany ancestor, probably from the England/Scotland border area.

In any case, even small numbers like this can make sense with some genealogical probing. A friend of mine who is French Canadian ends up being 0.3% Native American, which is consistent with the paper trail but not the haplogroups.

My point is, your Ancestry Composition might not match your haplogroup, but both are picking up on real parts of your ancestry. The Y has the advantage that you know which line it's from.

geebee
02-01-2013, 01:50 PM
Well, most of it isn't from DNA. Unfortunately, genetic genealogy isn't quite as precise as all that -- although it's come a long way.

Much of my information is from records, or copies of records. For example, in the case of the ancestor who brought a form of my surname to America, I even know the name of the ship he arrived on (Edinburgh) and the date on which he arrived (15 Sep 1749). Unfortunately, I don't have the name of the village in or near which he lived, only that the passengers aboard the ship were identified as from "the Palatinate". That was a broader term then than now, but roughly it means southwestern Germany. I also have seen records pertaining to him and his family since that time, including some that document the transition of my family name from its German to Americanized form.

The same thing is true for many other ancestors, although I am finding with some of the Germans that even in 17th-18th century central Europe there was a lot of movement. I was able to trace one ancestor back to Germany, only to find that her family actually seems to have come first from France. (Apparently they were Huegenots.) Further, I have seen documentation for French lines from French Canada; and French, Swiss, and Irish lines from the Mississippi and Louisiana Gulf Coast region. The "Choctaw" ancestry has fairly good documentation at least as Native American -- but in most of the documents I've seen the reference is usually just to her being "une indienne". (The time and place make Choctaw the most likely possibility, and there's some anecdotal support for her being Choctaw.)

Some of my estimate of my ancestry is based on the earliest forms (that I've seen) of various family surnames -- before they were as likely to have been transformed into something completely different. For example, I have 18th century Keith and Gregg (Grieg) lines that point back to Scotland, or to Scotland by way of Ireland. Of course, with a name like Weaver -- my mother's maiden name -- it would be impossible to know the origin without additional information. But the family tradition was that they were German, and I've since been able to find records showing that this was the case. The original form of the surname was Weber, or according to one source it was possibly even Weberling (which would mean "little weaver"). However, I've also seen the name spelled as Wever in one 18th century record. That could just be a phonetic spelling, but Wever is also the Dutch form (or so I understand). I nearly as I can tell, my Weaver ancestors were already in America by 1700, or not much after. And this line appears to early have mixed with French Huguenots, Dutch, and Swiss. Some researchers have traced the Dutch part back to New Holland, but I haven't independently verified it. Oh, and the same research suggests the "Mohawk", as well; but again, I haven't verified it.

The Catalan ancestry is the most recent. Both of my maternal grandmother's grandfathers immigrated from the Spanish island of Menorca, in the Mediterranean. They arrived separately. (One by himself, the other with both his parents and possibly a sister.) I've seen immigration documents for both, plus a copy of the naturalization document for one of them. (In becoming a citizen, he had to not only declare allegiance to the U.S., but he had to renounce allegiance to Spain. The language of the document includes: "... and especially to Queen Isabela." I guess so there'd be no confusion about what was expected.) One of my grandmother's grandmothers was born in Alsace-Lorraine but immigrated to the U.S. with her family at a young age. I usually include her among the German ancestors, but I could probably as easy count her among the French.

Anyway, I don't know whether that helps you. I suppose it would be nice if I were able to say that this is what genetic genealogy has been able to show me, but it's come more from years of following this lead or that. However, I should note that 23andMe has pretty consistently reported me as being between 98 and 98.5 percent European, with the remainder being "Asian". The new Ancestry Composition is better able to distinguish between Asian and Native American (actually, it couldn't at all), and it reports all of my "Asian" ancestry as Native American. Doug McDonald also did an analysis of my 23andMe data a while back and determined that probably 2% was Native American. However, he also estimated a much higher proportion of "Irish" ancestry for me than my paper trail suggests (something around 69 percent), about 6.5% Spanish ancestry, and suggested I also probably had 22% Jewish ancestry. No other research or DNA testing or analysis has suggested Jewish ancestry, and I suspect that his program was actually interpreting something else as Jewish. (Of course, if I can find and document any Jewish ancestry, I'll be quite happy to do so. Or Berber ancestry, or anything else. About the only thing I ask is that it at least be human. So far no little green men -- or women -- but I'm still waiting.)

Finally, if what you're really wanting to know is just what can DNA testing tell you, I think AJL's answer should be helpful. The results are not always terribly precise, or 100% accurate, but things like 23andMe's new "Ancestry Composition" are becoming much more sophisticated.

My results in AC (using the "speculative estimate", which is actually about the closest match to my paper trail; and of course the estimate is based on autosomal testing and matching -- not on Y DNA or mtDNA):

97.3% European
Northern European
36.1% British and Irish
9.1% French and German
0.5% Finnish
38.8% Nonspecific Northern Euro…
Southern European
2.2% Italian
1.3% Balkan
1.0% Iberian
4.1% Nonspecific Southern Eur…
0.7% Eastern European
3.4% Nonspecific European
2.0% East Asian & Native American
2.0% Native American
0.1% Nonspecific East Asian & Nativ…
< 0.1% Sub-Saharan African
0.6% Unassigned

You can compare this to the results of FTDNA's Family Finder test, which says simply "100% European, ±0.01%". It breaks that down into "Finnish, Russian, Spanish, Tuscan", but gives no percentages for any of those. You'll note that 23andMe says nothing of Russian ancestry, though it does come up with a tiny amount of Finnish (0.5%), as well as with a bit of Spanish (1.0%) and Italian (2.2%). But all of these percentages together only add up to 3.7. My guess it that FTDNA's analysis actually does "see" the Native American part, but it's not enough for it to give me a second continent. So it then has to try to fit everything as well as it can into Europe, which it can only do by deciding I'm more Russian and Finnish (which presumably have higher percentages of "Asian" DNA.)

Or you can compare my 23andMe results to those of AncestryDNA. That test (and interpretation) came up with Central European 48% -- which if you think mostly German, sounds about right -- Scandinavian 23%, Eastern European 16%, Finnish/Volga-Ural 6%, and Uncertain 7%. Since much of my British ancestry is actually Scottish, maybe that partly accounts for the "Scandinavian". The "Eastern European" seems high, but could partly be from the Germany ancestry -- as perhaps could be the "Finnish/Volga-Ural" (which does seem somewhat like FTDNA's Finnish/Russian combination, but still only a relatively small amount). Of course, the 7% "uncertain" could be anything.

I guess the point -- if I have one -- is that the analyses can actually vary quite a bit, even when they're looking at more-or-less the same thing. Much of that depends on what reference populations are used, and 23andMe uses not only published references but also data from some of its own customers. Ultimately, I think this is going to be the way to go, and a block-by-block comparison (especially for admixed folk) is going to be more useful than just looking at individual SNPs.

geebee
02-01-2013, 01:50 PM
Well, most of it isn't from DNA. Unfortunately, genetic genealogy isn't quite as precise as all that -- although it's come a long way.

Much of my information is from records, or copies of records. For example, in the case of the ancestor who brought a form of my surname to America, I even know the name of the ship he arrived on (Edinburgh) and the date on which he arrived (15 Sep 1749). Unfortunately, I don't have the name of the village in or near which he lived, only that the passengers aboard the ship were identified as from "the Palatinate". That was a broader term then than now, but roughly it means southwestern Germany. I also have seen records pertaining to him and his family since that time, including some that document the transition of my family name from its German to Americanized form.

The same thing is true for many other ancestors, although I am finding with some of the Germans that even in 17th-18th century central Europe there was a lot of movement. I was able to trace one ancestor back to Germany, only to find that her family actually seems to have come first from France. (Apparently they were Huegenots.) Further, I have seen documentation for French lines from French Canada; and French, Swiss, and Irish lines from the Mississippi and Louisiana Gulf Coast region. The "Choctaw" ancestry has fairly good documentation at least as Native American -- but in most of the documents I've seen the reference is usually just to her being "une indienne". (The time and place make Choctaw the most likely possibility, and there's some anecdotal support for her being Choctaw.)

Some of my estimate of my ancestry is based on the earliest forms (that I've seen) of various family surnames -- before they were as likely to have been transformed into something completely different. For example, I have 18th century Keith and Gregg (Grieg) lines that point back to Scotland, or to Scotland by way of Ireland. Of course, with a name like Weaver -- my mother's maiden name -- it would be impossible to know the origin without additional information. But the family tradition was that they were German, and I've since been able to find records showing that this was the case. The original form of the surname was Weber, or according to one source it was possibly even Weberling (which would mean "little weaver"). However, I've also seen the name spelled as Wever in one 18th century record. That could just be a phonetic spelling, but Wever is also the Dutch form (or so I understand). I nearly as I can tell, my Weaver ancestors were already in America by 1700, or not much after. And this line appears to early have mixed with French Huguenots, Dutch, and Swiss. Some researchers have traced the Dutch part back to New Holland, but I haven't independently verified it. Oh, and the same research suggests the "Mohawk", as well; but again, I haven't verified it.

The Catalan ancestry is the most recent. Both of my maternal grandmother's grandfathers immigrated from the Spanish island of Menorca, in the Mediterranean. They arrived separately. (One by himself, the other with both his parents and possibly a sister.) I've seen immigration documents for both, plus a copy of the naturalization document for one of them. (In becoming a citizen, he had to not only declare allegiance to the U.S., but he had to renounce allegiance to Spain. The language of the document includes: "... and especially to Queen Isabela." I guess so there'd be no confusion about what was expected.) One of my grandmother's grandmothers was born in Alsace-Lorraine but immigrated to the U.S. with her family at a young age. I usually include her among the German ancestors, but I could probably as easy count her among the French.

Anyway, I don't know whether that helps you. I suppose it would be nice if I were able to say that this is what genetic genealogy has been able to show me, but it's come more from years of following this lead or that. However, I should note that 23andMe has pretty consistently reported me as being between 98 and 98.5 percent European, with the remainder being "Asian". The new Ancestry Composition is better able to distinguish between Asian and Native American (actually, it couldn't at all), and it reports all of my "Asian" ancestry as Native American. Doug McDonald also did an analysis of my 23andMe data a while back and determined that probably 2% was Native American. However, he also estimated a much higher proportion of "Irish" ancestry for me than my paper trail suggests (something around 69 percent), about 6.5% Spanish ancestry, and suggested I also probably had 22% Jewish ancestry. No other research or DNA testing or analysis has suggested Jewish ancestry, and I suspect that his program was actually interpreting something else as Jewish. (Of course, if I can find and document any Jewish ancestry, I'll be quite happy to do so. Or Berber ancestry, or anything else. About the only thing I ask is that it at least be human. So far no little green men -- or women -- but I'm still waiting.)

Finally, if what you're really wanting to know is just what can DNA testing tell you, I think AJL's answer should be helpful. The results are not always terribly precise, or 100% accurate, but things like 23andMe's new "Ancestry Composition" are becoming much more sophisticated.

My results in AC (using the "speculative estimate", which is actually about the closest match to my paper trail; and of course the estimate is based on autosomal testing and matching -- not on Y DNA or mtDNA):

97.3% European
Northern European
36.1% British and Irish
9.1% French and German
0.5% Finnish
38.8% Nonspecific Northern Euro…
Southern European
2.2% Italian
1.3% Balkan
1.0% Iberian
4.1% Nonspecific Southern Eur…
0.7% Eastern European
3.4% Nonspecific European
2.0% East Asian & Native American
2.0% Native American
0.1% Nonspecific East Asian & Nativ…
< 0.1% Sub-Saharan African
0.6% Unassigned

You can compare this to the results of FTDNA's Family Finder test, which says simply "100% European, ±0.01%". It breaks that down into "Finnish, Russian, Spanish, Tuscan", but gives no percentages for any of those. You'll note that 23andMe says nothing of Russian ancestry, though it does come up with a tiny amount of Finnish (0.5%), as well as with a bit of Spanish (1.0%) and Italian (2.2%). But all of these percentages together only add up to 3.7. My guess it that FTDNA's analysis actually does "see" the Native American part, but it's not enough for it to give me a second continent. So it then has to try to fit everything as well as it can into Europe, which it can only do by deciding I'm more Russian and Finnish (which presumably have higher percentages of "Asian" DNA.)

Or you can compare my 23andMe results to those of AncestryDNA. That test (and interpretation) came up with Central European 48% -- which if you think mostly German, sounds about right -- Scandinavian 23%, Eastern European 16%, Finnish/Volga-Ural 6%, and Uncertain 7%. Since much of my British ancestry is actually Scottish, maybe that partly accounts for the "Scandinavian". The "Eastern European" seems high, but could partly be from the Germany ancestry -- as perhaps could be the "Finnish/Volga-Ural" (which does seem somewhat like FTDNA's Finnish/Russian combination, but still only a relatively small amount). Of course, the 7% "uncertain" could be anything.

I guess the point -- if I have one -- is that the analyses can actually vary quite a bit, even when they're looking at more-or-less the same thing. Much of that depends on what reference populations are used, and 23andMe uses not only published references but also data from some of its own customers. Ultimately, I think this is going to be the way to go, and a block-by-block comparison (especially for admixed folk) is going to be more useful than just looking at individual SNPs.

Grossvater
02-01-2013, 07:19 PM
Geebee...I have been reading your posts with interest as I, too, descend from Palatine Germans named Weaver (originally Weber). In fact, my mitochondrial DNA comes from Catherine Weaver Whitaker (1774-1852), the daughter of Nicholas Weaver (1736-1824) and Catherine Franke (we think). I believe they lived in the Mohawk Valley of New York State. Are we related?

authun
11-10-2013, 08:20 PM
It always depends on how much dna you look at. Look at too little and you merely identify a subject as 'human'. Look at too much and you identify the subject as an individual and different from both siblings and parents.

Your genes are older than ethnic groups and no genes, as far as I know, are exclusive to any particular geographic region. They are mostly spread out over areas. They were, at some time in the past, specific to families, clans and tribe but even early man understood the danger of in breeding. They do cluster together geographically but interpretation depends on context. In the UK R1a is often seen as evidence of vikings. In eastern europe R1a is seen as slavic yet the largest R1a population in the world is in India. Poles, Greeks, Italians etc are citizens of modern nation states and there are big variations within each of those.