PDA

View Full Version : Geographic Imbalances in FTDNA Database: Some Comparisons



GoldenHind
02-08-2016, 08:25 PM
The geographic bias toward the British Isles and Ireland in the FTDNA database has been mentioned frequently, but I suspect many fail to appreciate just how large it is. That failure has led to some people drawing conclusions that I believe are largely due to the geographic bias in question. I think many will be surprised by the following comparisons.

I have taken the number of people who have taken at least the basic 12 marker YDNA test (primarily because these figues are readily available), along with their indicated country of origin for their male ancestry, and compared them against the current population figures for the countries in question.. I have broadly rounded off all the figures, so all figures should be considered as approximate rather than exact.

The undoubted leader in terms of FTDNA samples per national population is Ireland. The population of Ireland is around 4.8 million, but there are some 22,500 samples in the FTDNA database who indicate Ireland as their YDNA country of origin.

Following some distance behind is Scotland (pop. 5.3m) with 17,000 entries. However the figures for Britain are complicated by the fact FTDNA lists the United Kingdom as a separate category in addition to England, Scotland and Wales. While my guess is that most of those who specify the United Kingdom as their male ancestral country of origin are actually of English descent, it is impossible to accurately divide them (around 13,500) into the component parts. So I have added the total sum for all the various UK categories: around 70,000. The total UK population is 64.5 million. While 70,000 seems like a lot, to have the same proportionate rate as Ireland, the combined total of samples for all the UK categories would have to increase to over 300,000.

Germany appears to have the highest number of samples at FTDNA of any country on the continent: 18,500. Again this may seem like a lot. However Germany has a population of 81 million. To have the same proportionate rate as the Untied Kingdom total, one would need to add almost another 70,000 German samples, to a total of 88,000. To equal the rate for Ireland, it would have to increase to around 380,000.

It is well known that France (pop. 64.5m) is under represented in the FTDNA database: the total is about 5000. However the situation with Italy (60m) is nearly as bad, with about the same number of samples (5000) in the FTDNA database. To equal the Irish rate, France would need to increase to over 300,000 and Italy to some 280,000.

The Scandinavian countries are also extremely under represented. To equal the rate for Ireland:
Sweden (almost exactly double the population of Ireland at 9.7m)) would need to increase from 3500 to 45,500.
Denmark from 1200 to over 26,000.
Norway from 2700 to to 23,500.

Spain (47.8m) would need to increase from 4750 to some 224,000 to match the rate for Ireland.

The situation is much the same for the other continental European countries.

Obviously one needs to keep these imbalances in mind when looking at raw numbers from the FTDNA database.

leonardo
02-08-2016, 09:04 PM
U.K. Could just as well represent Cornish or Welsh. I have been told that my great great great grandfather came from Wales. My matches that indicate European ancestry are somewhat vague though, likely because theirs are distant too and they are likewise unsure U. K. Is a safe bet.

evon
02-08-2016, 09:12 PM
You see the same problem with 23andme, as these are primarily Anglo-American oriented companies one would expect as much. Most of our matches are American and British/Irish, and our top surnames are Smith etc.

jbarry6899
02-08-2016, 09:54 PM
Not a big surprise, since most testers are American and Germany and Ireland are the largest sources for European-American ancestrh:

http://www.infoplease.com/ipa/A0762137.html

MitchellSince1893
02-08-2016, 10:16 PM
If it wasn't for the French Canadians those French numbers would be much smaller...not the same bias against dna tests as France itself has.

leonardo
02-08-2016, 10:26 PM
Not a big surprise, since most testers are American and Germany and Ireland are the largest sources for European-American ancestrh:

http://www.infoplease.com/ipa/A0762137.html

The undersampling by those with Italian ancestry is interesting.

razyn
02-08-2016, 10:32 PM
The sources of the present American gene pool, and modern stats on how people self-identify their origins on censuses, are very different sorts of data.

People whose ancestors came from somewhere in Europe since about 1848 (European famines and labor unrest, steamships, American railroads and the Homestead Act were factors, among others) are much more likely to have a record of their place of origin than those who were already here for a good many generations before the government much cared about that.

Huntergatherer1066
02-08-2016, 10:58 PM
If you have a friend with well-documented continental European ancestry, get them a nice birthday present.

GoldenHind
02-09-2016, 12:25 AM
Not a big surprise, since most testers are American and Germany and Ireland are the largest sources for European-American ancestrh:

http://www.infoplease.com/ipa/A0762137.html

Note however that Germany is very under represented when compared to both the UK totals and especially Ireland.

jdean
02-09-2016, 01:00 AM
The geographic bias toward the British Isles and Ireland in the FTDNA database has been mentioned frequently, but I suspect many fail to appreciate just how large it is. That failure has led to some people drawing conclusions that I believe are largely due to the geographic bias in question. I think many will be surprised by the following comparisons.

I have taken the number of people who have taken at least the basic 12 marker YDNA test (primarily because these figues are readily available), along with their indicated country of origin for their male ancestry, and compared them against the current population figures for the countries in question.. I have broadly rounded off all the figures, so all figures should be considered as approximate rather than exact.

The undoubted leader in terms of FTDNA samples per national population is Ireland. The population of Ireland is around 4.8 million, but there are some 22,500 samples in the FTDNA database who indicate Ireland as their YDNA country of origin.

Following some distance behind is Scotland (pop. 5.3m) with 17,000 entries. However the figures for Britain are complicated by the fact FTDNA lists the United Kingdom as a separate category in addition to England, Scotland and Wales. While my guess is that most of those who specify the United Kingdom as their male ancestral country of origin are actually of English descent, it is impossible to accurately divide them (around 13,500) into the component parts. So I have added the total sum for all the various UK categories: around 70,000. The total UK population is 64.5 million. While 70,000 seems like a lot, to have the same proportionate rate as Ireland, the combined total of samples for all the UK categories would have to increase to over 300,000.

Germany appears to have the highest number of samples at FTDNA of any country on the continent: 18,500. Again this may seem like a lot. However Germany has a population of 81 million. To have the same proportionate rate as the Untied Kingdom total, one would need to add almost another 70,000 German samples, to a total of 88,000. To equal the rate for Ireland, it would have to increase to around 380,000.

It is well known that France (pop. 64.5m) is under represented in the FTDNA database: the total is about 5000. However the situation with Italy (60m) is nearly as bad, with about the same number of samples (5000) in the FTDNA database. To equal the Irish rate, France would need to increase to over 300,000 and Italy to some 280,000.

The Scandinavian countries are also extremely under represented. To equal the rate for Ireland:
Sweden (almost exactly double the population of Ireland at 9.7m)) would need to increase from 3500 to 45,500.
Denmark from 1200 to over 26,000.
Norway from 2700 to to 23,500.

Spain (47.8m) would need to increase from 4750 to some 224,000 to match the rate for Ireland.

The situation is much the same for the other continental European countries.

Obviously one needs to keep these imbalances in mind when looking at raw numbers from the FTDNA database.

This post should to be pinned !!

leonardo
02-09-2016, 02:09 AM
Note however that Germany is very under represented when compared to both the UK totals and especially Ireland.

I thought I once read that dna test are discouraged there?

Dubhthach
02-09-2016, 04:51 PM
Just a note (I agree with point though about over representation of Isles population vis-a-vis continental ones)

With regards to Ireland vs. Britain, it should be noted that historically the ratio of population in Ireland to Britain was roughly 1:3 -> 1:2, so in the early 1840's you had 8-8.5 million people in Ireland and circa 15.9 million in England & Wales and 2.6 million in Scotland (as I said roughly 1:2 ratio). Ireland is unique in western Europe in having a smaller population today than in 1841.

I should also caution that the population of Island of Ireland today is circa 6.3 million the lower figure quoted is purely for 26 county state which only came into existence in 1922.

Anyways reason why I mention this is that most people who put their MDKA down as Ireland actually left Ireland before 1850. This is particulary case with people who identify as "Scots-Irish".

MitchellSince1893
02-09-2016, 05:39 PM
...So I have added the total sum for all the various UK categories: around 70,000. The total UK population is 64.5 million...

It is well known that France (pop. 64.5m) is under represented in the FTDNA database: the total is about 5000. However the situation with Italy (60m) is nearly as bad, with about the same number of samples (5000) in the FTDNA database. To equal the Irish rate, France would need to increase to over 300,000 and Italy to some 280,000.
...Spain (47.8m) would need to increase from 4750 to some 224,000 to match the rate for Ireland...

Just to put this in perspective:

Using my own U152>Z49>Z142>Z150 branch which is based mostly on FTDNA data, it's currently
51% British (English, Scottish, and Welsh)
16% French,
16% Italian,
7% Spanish,
7% Dutch, and
3% Swiss
http://www.anthrogenica.com/showthread.php?2341-R-Z150-anyone-can-tell-me-more-)&p=134394&viewfull=1#post134394

If I apply these ratios to the raw totals for each country, it would be around
7% British
31% French
29% Italian
13% Spanish
13% Dutch
7% Swiss

Yfull estimates TMRCA of Z150 to be 1900ybp so it's possibly a late Bell Beaker formation in Eastern France/Central Europe, and spread during the Tumulus and/or Urnfield Culture?

GoldenHind
02-09-2016, 10:15 PM
Just a note (I agree with point though about over representation of Isles population vis-a-vis continental ones)

With regards to Ireland vs. Britain, it should be noted that historically the ratio of population in Ireland to Britain was roughly 1:3 -> 1:2, so in the early 1840's you had 8-8.5 million people in Ireland and circa 15.9 million in England & Wales and 2.6 million in Scotland (as I said roughly 1:2 ratio). Ireland is unique in western Europe in having a smaller population today than in 1841.

I should also caution that the population of Island of Ireland today is circa 6.3 million the lower figure quoted is purely for 26 county state which only came into existence in 1922.

Anyways reason why I mention this is that most people who put their MDKA down as Ireland actually left Ireland before 1850. This is particulary case with people who identify as "Scots-Irish".

I was not aware the population of Ireland had dropped since the 1840s. That might change the ratios, but not the fact that Ireland has a much greater proportional repsresentation than that of any other country in the database. Even if I were to bump up the Irish population to 8.5 million, we still would have 22,500 samples at FTDNA for Ireland compared to the larger Sweden (9.7m) with only 3500 samples, and so on.

Nor does that have any effect on the over weighting of samples from the combined components of the UK. As I said, Germany, which appears to have the highest number of FTDNA samples of any continental country, would have to add almost the entire UK total of some 70,000 to its current 18,500 to equal the proportional representation for the UK.

I only wish the descendants of the continental European countries were as enthusiastic about DNA testing as the sons of Erin.

Dubhthach
02-10-2016, 10:38 AM
I was not aware the population of Ireland had dropped since the 1840s. That might change the ratios, but not the fact that Ireland has a much greater proportional repsresentation than that of any other country in the database. Even if I were to bump up the Irish population to 8.5 million, we still would have 22,500 samples at FTDNA for Ireland compared to the larger Sweden (9.7m) with only 3500 samples, and so on.

Nor does that have any effect on the over weighting of samples from the combined components of the UK. As I said, Germany, which appears to have the highest number of FTDNA samples of any continental country, would have to add almost the entire UK total of some 70,000 to its current 18,500 to equal the proportional representation for the UK.

I only wish the descendants of the continental European countries were as enthusiastic about DNA testing as the sons of Erin.

Indeed well that's why I caveat'd my post saying I agree about continental countries been under-represented. But I should note that Ireland had two and half times the population of Sweden in 1841. If we look at US historic census up to 1890 there was on average 10x the number of Irish born living in US than Swedish born (Well post 1860 -- when there was 86x number of Irish-born than Swedish-born), 1890 saw massive increase in number of Swedes, even still there was nearly always 2x number of Ireland-born than Swedish born in US up to 1930.

https://www.census.gov/population/www/documentation/twps0029/tab04.html

People of Irish background have just been showing up in North America for longer in consistently larger numbers over time. With almost 2 million Irish arriving in North America (US & Canada) from 1840-1855.

I've lots members in Ireland project whose MDKA in Ireland is back in the 1720's! (Interesting a cohort of those with earliest MDKA think of themselves as Scots-Irish even though they bear Gaelic Irish surnames)

The population of Ireland is currently at it's highest level since 1861. In 1841 Ireland had about 1/4th the population of Germany, if ye mapped that through to today you be looking at circa 20million, if we were to go with more historic figures for comparison of Ireland and Britain (for last 500 years) you are probably looking at a level of 12-16 million as the population of island of Ireland. (I'm giving it a range of 5:1 -> 4:1)

Part of issue in general though is that FTDNA database is basically entirely of Diaspora origin, I mean I'd regard myself as a rarity here in Ireland to have ever done Y-DNA testing, I'd imagine of those who have Irish MDKA in FTDNA (ballpark back-of-envelope figures here) that 99%+ were Diaspora origin and of that most (say over 95% were specifically Irish-American)

jbarry6899
02-10-2016, 01:03 PM
Another factor that may skew the results is that genealogy as a hobby/obsession in America has long been dominated by two groups: those with Colonial (mainly English) ancestry and members of the LDS church. I would expect that those groups represent a disproportionate share of early adopters of genetic genealogy.

jbarry6899
02-10-2016, 02:39 PM
n as the population of island of Ireland. (I'm giving it a range of 5:1 -> 4:1)

Part of issue in general though is that FTDNA database is basically entirely of Diaspora origin, I mean I'd regard myself as a rarity here in Ireland to have ever done Y-DNA testing, I'd imagine of those who have Irish MDKA in FTDNA (ballpark back-of-envelope figures here) that 99%+ were Diaspora origin and of that most (say over 95% were specifically Irish-American)

In our Anglo-Norman Irish surname project we have YDNA results for 117 men, of whom 6 were born in Ireland and 2 were born elsewhere to Irish parents or grandparents.

AJL
02-10-2016, 03:05 PM
This post should to be pinned !!

Good idea, I've stickied it.

corner
02-10-2016, 03:22 PM
Part of issue in general though is that FTDNA database is basically entirely of Diaspora origin, I mean I'd regard myself as a rarity here in Ireland to have ever done Y-DNA testing, I'd imagine of those who have Irish MDKA in FTDNA (ballpark back-of-envelope figures here) that 99%+ were Diaspora origin and of that most (say over 95% were specifically Irish-American)That is also what I find when mapping subclade distribution using FTDNA data. Like yourself in Ireland, I'm a rarity being born and having early roots in Yorkshire. However, unlike Ireland, any 'outside UK' FTDNA results with Yorkshire origins are a rarity too.

Any homemade subclade distribution map is not really accurate because people emigrated from some regions of Britain and Ireland in much greater numbers than they did from others. A map using FTDNA data does not represent the actual density or frequency of subclade distribution across the Isles, let alone Europe. In my experience, such subclade maps more accurately represent where emigrants to North America left (when known) one, two or three hundred years ago. They don't so much represent where the subclade was frequent or densely/sparsely concentrated this side of the Atlantic in earlier times.

Don't get me wrong, it is great to have so many North American results, there would only be a small handful of British Isles/Ireland matches to our subclade on the map otherwise. However, I think the real density/frequency of subclades would only be seen with a comprehensive Europe-wide NGS study that samples all countries of Europe equally. If our relatively young DF27>ZZ12+ subclade had no North American-born testers, by far the region of highest frequency of European-born FTDNA matches would be from Northern Sweden. An interesting anomaly considering the above-mentioned low proportion of tests from Sweden in FTDNA databases in comparison to North Americans of British/Irish origin.

firemonkey
02-11-2016, 02:32 PM
Using the surname search box for family finder I get Ireland 151 Sweden 15 Norway 9 Denmark 7. On face value it seems like few Scandinavian.
However would I be right in thinking it's quite a lot allowing for ratio of Sweden/Norway/Denmark samples to Irish?

miiser
03-14-2016, 12:36 AM
I've seen the sampling bias used in these forums to dismiss haplogroup concentration information for the Continent and comparisons to the Isles. So it probably bears mentioning that meaningful information can still be obtained in spite of the sampling bias. Many readers probably known this already, but two points need to be addressed. Specifically, I've seen it argued that:

1) Comparisons of haplogroup concentrations between locations are not valid because of the biased sampling rate. This is false. Haplogroup concentration (such as those seen on distribution maps) for a specific location is independent of the varying sampling rate for different locations. For example, L21 concentration in France is independent of the sampling rate of Ireland. L21 concentration in France depends only on the number of L21 samples from France out of the total number of samples from France.

2) The sample size of many Continental locations is too low to make an accurate assessment of haplogroup concentrations. This is false. The thousands of samples in France, Germany, etc. are more than sufficient to make statistical statements with high confidence.

The sampling bias is easily dealt with in analyses. One just needs to keep in mind that it is the percentage of samples out of a particular population that is most relevant, not the number of samples.

AJL
03-14-2016, 12:46 AM
The sampling bias is easily dealt with in analyses. One just needs to keep in mind that it is the percentage of samples out of a particular population that is most relevant, not the number of samples.

^ It's really only where we have very few examples of a haplogroup overall that sampling bias could enter into things, and even then it's relatively easy to make an informed guess based on phylogeny and sampling rates. (E.g. we have about as many cases of R1a-F1345 in Scotland as in the Caucasus, but at several thousand times the sampling rate, Scotland is much more likely to be the destination and the Caucasus the source.)

MitchellSince1893
03-14-2016, 03:14 AM
As I discovered in doing my FTDNA England Y-dna study, there is another bias to consider:


Non R haplotypes are 11% more likely to have a confirmed SNP test compared to R haplotype, and thus will be over represented in the FTDNA data.http://www.anthrogenica.com/showthread.php?6154-Y-DNA-Haplogroup-Percentages-and-maps-for-England-Source-FTDNA-Y-DNA-projects&p=131977&viewfull=1#post131977

This results in an under representation of R1b subclades compared to other haplogroups in FTDNA projects...at least as it pertains to the British Isles.

wombatofthenorth
04-26-2016, 01:06 AM
You see the same problem with 23andme, as these are primarily Anglo-American oriented companies one would expect as much. Most of our matches are American and British/Irish, and our top surnames are Smith etc.

Yeah it's so bad at 23 that on the new site and the new way they present your most common surname match, 99.9% of all people on 23 get Smith as their #1 choice. Doesn't matter what their background hah!
I think I checked and not a single one of my top 10 or even 20 surnames was even associated with any of my countries of main ancestry hah.

TigerMW
05-13-2016, 11:57 AM
Indeed well that's why I caveat'd my post saying I agree about continental countries been under-represented. But I should note that Ireland had two and half times the population of Sweden in 1841. If we look at US historic census up to 1890 there was on average 10x the number of Irish born living in US than Swedish born (Well post 1860 -- when there was 86x number of Irish-born than Swedish-born), 1890 saw massive increase in number of Swedes, even still there was nearly always 2x number of Ireland-born than Swedish born in US up to 1930.

https://www.census.gov/population/www/documentation/twps0029/tab04.html

People of Irish background have just been showing up in North America for longer in consistently larger numbers over time. With almost 2 million Irish arriving in North America (US & Canada) from 1840-1855.

I've lots members in Ireland project whose MDKA in Ireland is back in the 1720's! (Interesting a cohort of those with earliest MDKA think of themselves as Scots-Irish even though they bear Gaelic Irish surnames)

The population of Ireland is currently at it's highest level since 1861. In 1841 Ireland had about 1/4th the population of Germany, if ye mapped that through to today you be looking at circa 20million, if we were to go with more historic figures for comparison of Ireland and Britain (for last 500 years) you are probably looking at a level of 12-16 million as the population of island of Ireland. (I'm giving it a range of 5:1 -> 4:1)

Part of issue in general though is that FTDNA database is basically entirely of Diaspora origin, I mean I'd regard myself as a rarity here in Ireland to have ever done Y-DNA testing, I'd imagine of those who have Irish MDKA in FTDNA (ballpark back-of-envelope figures here) that 99%+ were Diaspora origin and of that most (say over 95% were specifically Irish-American)

In another conversation, Paul C commented, "You're forgetting that the total Irish population - both in Ireland + worldwide diaspora - is between 80-90 million. So Ireland is NOT as overrepresented as you state, nor are some countries as underrepresented."

It's true in some respects but I think it is important to review the representativeness of the data in light of an intended use.

For instance, in a discussion of an origin point for a haplogroup modern frequencies really have not much to do with anything. If Irish populations expanded faster in recent times (due to say Roman Catholic practices?) that is not relevant to ancient Irish branching.

On the other, there must be some methodology and criteria to seeking a representative sampling in science papers as it relates to ancestral origin evaluations. For example, how can we say Germany has enough NGS tests to conclude we understand haplogroups and branching in Germany? I'm afraid the answer is we are a long way off, period. ... and nearly hopeless so modern peoples' data is near useless other than define a bare bones branching skeleton of the tree.

Our my fears correct?

Here is a hypothetical example for a fast growing, successful haplogroup. The actual intended use would be the P311 and P312 families of early branches.

Let's expand on that and use an ancient example of a person who is believed to be a L21+. I'm not saying what he is, but at least we have an example of Irish high king.

"Keating credits Niall with two wives: Inne, daughter of Lugaid, who bore him one son, Fiachu; and Rignach, who bore him seven sons, Lóegaire, Éndae, Maine, Eógan, Conall Gulban, Conall Cremthainne and Coirpre. These sons are the eponymous ancestors of the various Uí Néill dynasties...."
https://en.wikipedia.org/wiki/Niall_of_the_Nine_Hostages

Niall may have been more of a polygamist than we know, but at least we think he had 8 sons. If this rate of male descendants continued for just a couple of generations look at what we would like get.

8 son x 8 = 64 grandsons

64 grandsons x 8 = 512 g-grandsons

Within just three generations there could be 584 father-son transmission events from Niall on down. (584 = 8 + 64 + 512)

The rate of SNPs on the Y per father-son transmissions is about one every three to four.

We could expect 167 SNPs from the hypothetical Niall to g-grandsons example (167 = 584 / 3.5).

When we are looking at your top-layer L21 list from the NGS tree of DF63, A5846, A7900 and DF13; we can see we could easily be missing a couple of hundred SNPs.

Were they there and where did they all go?

If we are missing hundreds of early branches are just evaluating a handful do we have any hopes of using modern people for ancestral origins? I'm afraid there won't even be much discussion on this.

Perhaps the answer is we can't identify actual origins but potentially major launch points for expansions and migrations.

Dubhthach
05-13-2016, 12:55 PM
"Keating credits Niall with two wives: Inne, daughter of Lugaid, who bore him one son, Fiachu; and Rignach, who bore him seven sons, Lóegaire, Éndae, Maine, Eógan, Conall Gulban, Conall Cremthainne and Coirpre. These sons are the eponymous ancestors of the various Uí Néill dynasties...."
https://en.wikipedia.org/wiki/Niall_of_the_Nine_Hostages

Niall may have been more of a polygamist than we know, but at least we think he had 8 sons. If this rate of male descendants continued for just a couple of generations look at what we would like get.

8 son x 8 = 64 grandsons

64 grandsons x 8 = 512 g-grandsons

Within just three generations there could be 584 father-son transmission events from Niall on down. (584 = 8 + 64 + 512)

The rate of SNPs on the Y per father-son transmissions is about one every three to four.

We could expect 167 SNPs from the hypothetical Niall to g-grandsons example (167 = 584 / 3.5).

When we are looking at your top-layer L21 list from the NGS tree of DF63, A5846, A7900 and DF13; we can see we could easily be missing a couple of hundred SNPs.

Were they there and where did they all go?

If we are missing hundreds of early branches are just evaluating a handful do we have any hopes of using modern people for ancestral origins? I'm afraid there won't even be much discussion on this.

Perhaps the answer is we can't identify actual origins but potentially major launch points for expansions and migrations.

Given that the Irish were hardly Catholic (in a Tridentine sense) before the Famine I'm not sure what your point is, the massed expansion of Irish population in early 19th century was driven by cheap food supply (eg. potato) and demand driven for argricultural products by the likes of the Napolenic war. Of course in the proceeding 300 or so years there had been a number of other times when Irish population had more than doubled before hitting a "population crash" (Cromwellian war been prime one)

Leaving that aside don't forget that Niall had 4 brothers as well
http://compsoc.nuigalway.ie/~dubhthach/DNA/eochaid-connachta.png

9272

What should be noted is that the likes of Céithinn (Keating's) Forsa Feasa ar Éirinn is a work from 17th century that is written in a specific context (eg. the Tudor/Stewart conquest of Ireland) ergo it's not a dispassionate document. That and it inherits a long history of other such texts right back to works of the histories of the 8th century who crafted a common narrative in early Christian period for history/pre-history of Ireland

A New History of Ireland: Prehistoric and early Ireland -- Chapter VII Ireland, 400-800
https://books.google.ie/books?id=2rwRDAAAQBAJ&lpg=PA391&dq=synchronistic%20history%20ireland%208th%20centu ry&pg=PA182#v=onepage&q&f=false

A more specific read about the Uí Néill been this chapter from TM. Charles Edwards "Early Christian Ireland"

http://compsoc.nuigalway.ie/~dubhthach/DNA/ui_neill_01.jpeg

http://compsoc.nuigalway.ie/~dubhthach/DNA/eci_ui_neill.html

More of this book can be read here:
https://books.google.ie/books?id=g6yq2sKLlFkC&printsec=frontcover&dq=early+christian+ireland&hl=en&sa=X&redir_esc=y#v=onepage&q&f=false

Niall if he did actually exist (some historians would argue he's an allegorical figure) was probably no more than a king of a Tuatha (petty kingdom), the importance perhaps of him been driven by conquest of Meath by two of his son's (Fiachu and Coirpre) elevating this lineage within context of wider Dál Cuinn.

Leaving that aside the expansion (some would say explosion) of the Dál Cuinn (including the Uí Néill) is part of wider picture that we see in Ireland in period 400-600AD, on a smaller scale it's a situation repeated time again throughout the history of Gaelic Ireland given societal structure which allowed for fairly rampant level of divorce and had no concept of illegitimacy (son of a concubine was just as legit as son of a wife). The rapid expansion of the Maguires from 1300 until 1600 is a case in point.

Anyways if we assume M222 is linked to the Dál Cuinn, than yes we don't have an idea of it's origin (as we are missing early branches etc.) but it's rapid expansion around 1500 years ago (based on TMRCA calculations for major branches) is possibly due to rapid expansion of lineages claiming to be part of the Dál Cuinn (which in 200 year period go from nobodies to ruling directly close on 50% of island and claiming a "titular" High-Kingship over rest)

TigerMW
05-13-2016, 06:56 PM
Given that the Irish were hardly Catholic (in a Tridentine sense) before the Famine I'm not sure what your point is, the massed expansion of Irish population in early 19th century was driven by cheap food supply (eg. potato) and demand driven for argricultural products by the likes of the Napolenic war. Of course in the proceeding 300 or so years there had been a number of other times when Irish population had more than doubled before hitting a "population crash" (Cromwellian war been prime one)

In another discussion I was following Goldenhind's theme about how biased the FTDNA database is so you'd have to multiply results X times to compensate for lack of testing in many European countries and try to have a straight comparison with the UK.

One counter-argument (from Paul C), which is a fair point, is that we should add the Irish-Americans to the current Irish population to calibrate the X times factor. The net is there are so many Irish-Americans it would balance things out quite a bit.

The whole thing just causes me to reconsider how valuable modern population frequencies are in terms of ancient, ancient ancestral origins. The answer is probably "not very" but it does matter the context.

If the context is do we have the ancient early branching covered then a representative survey may not be needed. I was trying to use the Niall example as a possible L21, DF13, DF63, etc. scenario. A very prolific family would have created a couple hundred SNPs in just a few generations. If we contrast that with the handful in the top layer of L21, DF13, DF63, etc. the challenge is there must be multitudes of ancient branching missing. What happened? Probably most died out but there might some in places like Benelux or Germany or France but that goes back to the original concern which is have we done enough NGS testing of those countries?


Leaving that aside don't forget that Niall had 4 brothers as well
http://compsoc.nuigalway.ie/~dubhthach/DNA/eochaid-connachta.png
This really wasn't about Niall or his brothers or M222. This was just about an ancient Celtic elite paternal lineage possibility.

TigerMW
05-13-2016, 07:04 PM
... What happened? Probably most died out but there might some in places like Benelux or Germany or France but that goes back to the original concern which is have we done enough NGS testing of those countries?
Paul D, what is the total # of Big Ys in the Ireland Heritage project on the statistics page?

I can count the German Language Area project. It's 345

The Irish project must have about 3-4 times of that.

Dubhthach
05-13-2016, 08:15 PM
This really wasn't about Niall or his brothers or M222. This was just about an ancient Celtic elite paternal lineage possibility.

The thing is you can't take early medieval Irish lineage dynamics and apply it to wider Celtic speaking Europe, the rapid expansion we see in lineages in Ireland appears to be result of turmoil/changeover in the immediate post christian period. As some have pointed out, Ireland moved from a tribal society to one based on lineage (one could say obsessed) in a relatively short period.

Dubhthach
05-13-2016, 08:19 PM
Paul D, what is the total # of Big Ys in the Ireland Heritage project on the statistics page?

I can count the German Language Area project. It's 345

The Irish project must have about 3-4 times of that.

671 or about ~10% of project members which today is at 6,859 (6,814 have posted Y-DNa results), the figure of about 10% is fairly equivalent for the German project, which apepars to be about half size of Ireland one.

TigerMW
05-13-2016, 11:46 PM
671 or about ~10% of project members which today is at 6,859 (6,814 have posted Y-DNa results), the figure of about 10% is fairly equivalent for the German project, which apepars to be about half size of Ireland one.
Yes, this is proportionally representative per project size, but is this representative of the SNP branching discovery work needed to uncover the tree, at least for R1b-P311 or more specifically for R1b-L21?

I think the answer is we don't know until we've done it and we face the law of diminishing returns where new ancient branches are no longer discovered as new results come in.

I'm using the German Language project as an example partially because southern Germany, Switzerland over to the Czech Republic is the traditional "Proto-Celtic" zone.

I disagree with the legacy this was an Iron Age phenomenon as I think it was much earlier, even to preceed the Celtic from the West theory, but the location is right.

Judith
09-30-2016, 02:20 PM
Reply to goldenhind about the geographical inbalances.
I would like to reply to your pm reply but I am sorry that I am not able to because your "Inbox" is full. I hope to communicate later.

HansDB
04-04-2017, 04:57 PM
Does anybody know where I can find the breakdown of the FTDNA database per country?

mountainguy
07-12-2017, 06:10 AM
This is a bit off topic.... but in reading that picture of what you posted I can see now where J.R. Tolkien got his inspiration for his languages and names and places in his writings.

RobertCasey
07-13-2017, 03:06 AM
There are not only geographic biases - primarily due to the ethnic makeup of the America's appetite to connect to their homelands - but there are other biases:

1) Recent haplogroups that are more prolific tend to draw more testing since better sample sizes leads to better results. Also, having a predictable single signature YSNP haplogroup helps as well as having genetic isolation from other haplogroups. Prolific lines are usually associated with strong leadership and wealth - M222 and L226 due to both uniting / conquering all of Ireland helped with these haplogroups enjoy prolific growth.

2) Even within these prolific haplogroups - there are many who have genetic testing challenges and less initial success tend and these testers to lose interest. For L226, we have hit a testing bottleneck with the current state of NGS testing, SNP pack testing and YSEQ private YSNP testing. The percentage that I can chart with signatures has now peaked at around 80 %. There are four factors that prevent this number growing to 90 to 95 %:

a) Technology limits of YDNA testing. With 80 NGS tests and 80 L226 SNP packs, new branches are just not being discovered with every Big Y these days. So until the long read FGS/WGS yields 2X coverage of the Big Y for not much more cost, until the L226 SNP pack doubles in numbers of private YSNPs included, until all 111 markers can be read with longer read lengths in NGS tests or until YSEQ testing of private YSNPs becomes more widespread - the normal rate of progress will continue to drop off until the next technology improvement and charting will stuck at 80 %.

b) There are lot of smaller genetic clusters of related individuals where there just is less interest and less leadership within the cluster - so progress is very slow when leadership in these clusters are not present to encourage YSNP testing of smaller clusters as well as "growing the genetic YSTR cluster."

c) There are those that are barely part of L226 with very high genetic distance from the L226 signature, high genetic distance with matches and the few matches that they do have - never have the same surname since they belong to very old branches that are not prolific. These people are less interested in testing for the future when they do not experience the same success of more recent and more prolific branches (or well tested branches).

d) L226 (and M222 has this as well) has around 10 % where the genetic distance is only two or three from the L226 signature that is 1,500 years old (this ignores the CDY markers that are just too volatile to use for charting). We actually have one tester that has a genetic distance of zero after 1,500 years and has 400 "FTDNA matches" at 67 markers. His matches include around 25 branches that are known to be over 1,000 years old. Some of these testers have 80 to 90 % false match rates at 67 markers with the FTDNA matching system. The false hit rate primarily depends on how large the branch is. The only solution for these people will be 400 YSTRs or extensive YSNP testing since YSTRs are not reliable for relatedness.

Unless our L226 project can sponsor some of this testing, we are just are not going to be able to chart more than 80 % of the testers. For every new Big Y and L226 SNP pack ordered that reveals more to chart, you have the same percentage of new 67 marker only testers joining the ranks of L226 as well. Just like was suggested that we should fund the Italians even though we are not Italians, it is the same issue of getting others to pay for the testing of the needs of the project (in this case - comprehensive and accurate charting) - which is what genealogists really want from their testing. Getting people to fund testing of those that are not closely related is a tough sell.

Rick P
10-10-2018, 05:38 AM
Anyone know which DNA testing firms are used most in Italy?

rkenobi
05-21-2019, 02:42 PM
The attached map illustrates the problem: every dot is a SNP in the complete FTDNA Y tree, located as a weighted average of the user-reported ancestry. I've dithered the dots to show their numbers. Country "centers" are easily seen for Norway, Sweden, Finland, Germany, Poland (weakly), Russia, Spain, and Saudi Arabia. But England, Ireland, Scotland, and the UK (a separate dot between the other 3) overwhelm all of the others.

30546

Dewsloth
05-21-2019, 06:31 PM
It gets confusing!

How to account for all the factors?
For instance, these are the reporting DF19s on FTDNA. The massive over-reporting UK block is in force.
France has nearly 7 times the population of Sweden, and more than 10 times that of Norway, but maybe Sweden reports as a higher percentage of population than France?
Or do the French Canadians help mitigate this?

England 81 20.56%
Germany 74 18.78%
Scotland 61 15.48%
United States 39 9.90%
Ireland 31 7.87%
United Kingdom 21 5.33%
Netherlands 13 3.30%
Belgium 13 3.30%
France 9 2.28%
Northern Ireland 9 2.28%
Sweden 7 1.78%
Canada 6 1.52%
Norway 5 1.27%
Wales 5 1.27%
Poland 4 1.02%
Czech Republic 4 1.02%
United States (Native American) 3 0.76%
Switzerland 2 0.51%
Austria 1 0.25%
Luxembourg 1 0.25%
Portugal 1 0.25%
Guyana 1 0.25%
Denmark 1 0.25%
Italy 1 0.25%
Isle of Man 1 0.25%
Unknown Origin 391 **

ArmandoR1b
05-21-2019, 08:25 PM
The attached map illustrates the problem: every dot is a SNP in the complete FTDNA Y tree, located as a weighted average of the user-reported ancestry. I've dithered the dots to show their numbers. Country "centers" are easily seen for Norway, Sweden, Finland, Germany, Poland (weakly), Russia, Spain, and Saudi Arabia. But England, Ireland, Scotland, and the UK (a separate dot between the other 3) overwhelm all of the others.

30546
Nice job! I wish Spain and Portugal had at least half the dots that the Isles have. France could use some more too.

razyn
05-21-2019, 08:25 PM
Norway, Ireland and Sweden are among the few countries from which around 25 to over 50 percent of the entire population emigrated to the United States, between the late 1840s and early 1920s. Sweden/Finland and (mostly Northern) Ireland also had small but genetically influential populations well established in eastern North America during the colonial period (before US independence). All of that contributes to the weighting of the FTDNA customer base, in ways that aren't reflected by comparing today's population figures for various European countries.