Results 1 to 10 of 10

Thread: Confusing Genetic Community allocations

  1. #1
    Registered Users
    Posts
    2,791
    Ethnicity
    Northern Ireland
    Nationality
    Northern Irish
    Y-DNA (P)
    R1b-L193(BY2634)
    Y-DNA (M)
    R1b-M222

    Northern Ireland Ireland Scotland

    Confusing Genetic Community allocations

    I have noticed in my matches that most of the Genetic Communities align with the ethnicity of the person. However, I have found a few anomalies of matches that don’t align with their ethnicity.

    I am Northern Irish as most of you know, and I will show you some examples of other Northern Irish people’s results to illustrate the disparity that I am referring to:

    A). 71% Ireland and 29% Scotland.

    GC: Donegal.

    B. 69% Ireland and 31% Scotland.

    GC: Scottish lowlands and SW Scotland.

    C). 92% Ireland and 8% Scotland.

    GC: Scottish Lowlands, Donegal, Ulster and SW Scotland.

    Now for the most part in my Northern Irish matches their Genetic Communities tally up with the ethnicity results of the person.

    However, there are some anomalies as above.

    Let me explain:

    A). This person is majority Irish but clearly has a Scottish connection. Why do they not have the Ulster or Scottish Genetic Community?

    B. Why does this person have a similar ethnic make up as person A but only has Scottish Genetic Communities?

    In both of these examples, I am confused at the allocation of the GC’s. I understand that it’s all based on matches but surely if you have Ulster Scots or Ulster Irish dna you would match people from those regions? So why are these people’s GC’s so different despite having almost identical autosomal results?

    C). This person is way left field. How can one be 8% Scottish but have the Scottish Lowlands community AND the SW Scotland community? 8% is a very small number and less than a Great Grandparents genetic contribution. I have found matches as high as 50% Scottish with no Scottish GC.

    So what is happening here? Why are these disparities occurring?

    If a grandparent etc contributes enough Scottish Dna to be present in autosomal results why is it not appearing in GC’s?
    Last edited by Nqp15hhu; 06-23-2021 at 10:06 PM.

  2. The Following User Says Thank You to Nqp15hhu For This Useful Post:

     Mr.G (06-23-2021)

  3. #2
    Registered Users
    Posts
    150
    Sex
    Location
    USA
    Ethnicity
    1/2 Hungarian, 1/2 German
    Nationality
    USA
    Y-DNA (P)
    N-Y17696 YSEQ
    mtDNA (M)
    H1e1a

    .....
    Last edited by Mr.G; 06-23-2021 at 10:34 PM. Reason: Misread the topic

  4. #3
    Registered Users
    Posts
    23
    Sex
    Ethnicity
    Finno-Anglo-Celt (Bitsa)
    Nationality
    Australian

    Australia Finland Cornwall Scotland Northern Ireland United Kingdom
    Hi Nqp15hhu,

    Do you have any paper trail information that details the last few generations for A, B, and C? Do you know whether their ancestries are fairly 'grouped' into distinct local areas, or are their ancestors drawn from a wide range of places in Ireland and/or Scotland? Also, what are the ranges that AncestryDNA provides for each person's ethnicity estimates? I'm wondering if some of those ranges (especially for C) are quite wide.

    As far as the Genetic Communities side of things goes, my sense of things on AncestryDNA is that GCs are a bit hit and miss. My uncle, for example, gets very accurate (based on paper trail) Genetic Communities in Cornwall and East Anglia, but his Ulster Scots branch doesn't show up as a Genetic Community at all (it just shows up as percentages of generic 'Irish' and 'Scottish'). In his case, that's not all that surprising as we have fewer Ulster Scots branches in recent generations than we do Cornish and East Anglian, but it's interesting to note that LivingDNA pinpoints my mum's Ulster Scots branch quite clearly as Northern Ireland / SW Scotland, so there are obviously 'enough genetics' there to play with between the two of them in order to show which cluster of Irish and/or Scottish signatures we're looking at (and it links in perfectly with our paper trail too, which is good).

    In relation to A, B and C, I'm wondering whether the key to this is simply the fact that what constitutes 'Irish' and 'Scottish' in broad genetic terms doesn't always line up perfectly with the geographies in question - in the dendrogram in Figure 2 of Byrne et.al's 'Insular Celtic population structure and genomic footprints of migration' article, for example, the South Scotland (SSC) cluster is closest, genetically, to the Central Leinster (CLN) cluster. If we were grouping the clusters purely along the lines of which genetics look the most 'alike', SSC would therefore fall into the 'Irish' box rather than the 'Scottish' box*, even though people with SSC genetics live mainly in Scotland. My reading of that is that someone from south-western Scotland (or with ancestors who lived in south-western Scotland) may well look quite 'Irish' from a broad ethno-genetic point of view AND still be identifiable as being descended from communities in south-western Scotland, more specifically. That could very well go a long way to explaining the anomalies you've identified and, I think, (C) in particular.

    What are your thoughts?

    * Having said that, Byrne et.al's Figure 3 ChromoPainter diagram illustrates SSC a little differently from Figure 2 by pulling the SSC cluster back onto the island of Britain completely. Even so, the dendrogram still retains the close genetic relationship of SSC and CLN and broadly classifies both in the Irish genetic grouping.


    Quote Originally Posted by Nqp15hhu View Post
    I have noticed in my matches that most of the Genetic Communities align with the ethnicity of the person. However, I have found a few anomalies of matches that don’t align with their ethnicity.

    I am Northern Irish as most of you know, and I will show you some examples of other Northern Irish people’s results to illustrate the disparity that I am referring to:

    A). 71% Ireland and 29% Scotland.

    GC: Donegal.

    B. 69% Ireland and 31% Scotland.

    GC: Scottish lowlands and SW Scotland.

    C). 92% Ireland and 8% Scotland.

    GC: Scottish Lowlands, Donegal, Ulster and SW Scotland.

    Now for the most part in my Northern Irish matches their Genetic Communities tally up with the ethnicity results of the person.

    However, there are some anomalies as above.

    Let me explain:

    A). This person is majority Irish but clearly has a Scottish connection. Why do they not have the Ulster or Scottish Genetic Community?

    B. Why does this person have a similar ethnic make up as person A but only has Scottish Genetic Communities?

    In both of these examples, I am confused at the allocation of the GC’s. I understand that it’s all based on matches but surely if you have Ulster Scots or Ulster Irish dna you would match people from those regions? So why are these people’s GC’s so different despite having almost identical autosomal results?

    C). This person is way left field. How can one be 8% Scottish but have the Scottish Lowlands community AND the SW Scotland community? 8% is a very small number and less than a Great Grandparents genetic contribution. I have found matches as high as 50% Scottish with no Scottish GC.

    So what is happening here? Why are these disparities occurring?

    If a grandparent etc contributes enough Scottish Dna to be present in autosomal results why is it not appearing in GC’s?

  5. #4
    Registered Users
    Posts
    2,791
    Ethnicity
    Northern Ireland
    Nationality
    Northern Irish
    Y-DNA (P)
    R1b-L193(BY2634)
    Y-DNA (M)
    R1b-M222

    Northern Ireland Ireland Scotland
    Quote Originally Posted by Cunobelinus_T View Post
    Hi Nqp15hhu,

    Do you have any paper trail information that details the last few generations for A, B, and C? Do you know whether their ancestries are fairly 'grouped' into distinct local areas, or are their ancestors drawn from a wide range of places in Ireland and/or Scotland? Also, what are the ranges that AncestryDNA provides for each person's ethnicity estimates? I'm wondering if some of those ranges (especially for C) are quite wide.

    As far as the Genetic Communities side of things goes, my sense of things on AncestryDNA is that GCs are a bit hit and miss. My uncle, for example, gets very accurate (based on paper trail) Genetic Communities in Cornwall and East Anglia, but his Ulster Scots branch doesn't show up as a Genetic Community at all (it just shows up as percentages of generic 'Irish' and 'Scottish'). In his case, that's not all that surprising as we have fewer Ulster Scots branches in recent generations than we do Cornish and East Anglian, but it's interesting to note that LivingDNA pinpoints my mum's Ulster Scots branch quite clearly as Northern Ireland / SW Scotland, so there are obviously 'enough genetics' there to play with between the two of them in order to show which cluster of Irish and/or Scottish signatures we're looking at (and it links in perfectly with our paper trail too, which is good).

    In relation to A, B and C, I'm wondering whether the key to this is simply the fact that what constitutes 'Irish' and 'Scottish' in broad genetic terms doesn't always line up perfectly with the geographies in question - in the dendrogram in Figure 2 of Byrne et.al's 'Insular Celtic population structure and genomic footprints of migration' article, for example, the South Scotland (SSC) cluster is closest, genetically, to the Central Leinster (CLN) cluster. If we were grouping the clusters purely along the lines of which genetics look the most 'alike', SSC would therefore fall into the 'Irish' box rather than the 'Scottish' box*, even though people with SSC genetics live mainly in Scotland. My reading of that is that someone from south-western Scotland (or with ancestors who lived in south-western Scotland) may well look quite 'Irish' from a broad ethno-genetic point of view AND still be identifiable as being descended from communities in south-western Scotland, more specifically. That could very well go a long way to explaining the anomalies you've identified and, I think, (C) in particular.

    What are your thoughts?

    * Having said that, Byrne et.al's Figure 3 ChromoPainter diagram illustrates SSC a little differently from Figure 2 by pulling the SSC cluster back onto the island of Britain completely. Even so, the dendrogram still retains the close genetic relationship of SSC and CLN and broadly classifies both in the Irish genetic grouping.
    Hi Cunobelinus_T, this is a very detailed and well thought out answer. I have actually thought about the issues that you discuss in your first paragraph, unfortunately I can't access the ethnic percentages for these people in a detailed manner, I am sure it would be interesting to see that though. I have just gone through their family trees, person A has ancestry primarily from the Inishowen peninsula and person B has ancestry from the Southwest of Ireland and the North coast of Northern Ireland. Person C's family tree is not present.

    I don't know if this answers your question. But you do have a relevant point, and an interesting one. I can see this myself with my father and myself on LivingDNA, his results are mostly Scottish Borders, whilst my results are supposedly from Ireland. I don't know again if this is because of issues in deciphering the differences in the dna between those two regions. It is a little odd that we would have such a huge difference between our respective results.

  6. #5
    Registered Users
    Posts
    2,068
    Sex
    Location
    Chicago, Illinois
    Ethnicity
    mixed European
    Nationality
    USA
    Y-DNA (P)
    Dad: R1b/L21/DF63
    mtDNA (M)
    K2b2
    Y-DNA (M)
    R1b-M269
    mtDNA (P)
    K2b1a1a

    United States of America England Wales Sweden Germany
    Quote Originally Posted by Nqp15hhu View Post
    I have noticed in my matches that most of the Genetic Communities align with the ethnicity of the person. However, I have found a few anomalies of matches that don’t align with their ethnicity.

    I am Northern Irish as most of you know, and I will show you some examples of other Northern Irish people’s results to illustrate the disparity that I am referring to:

    A). 71% Ireland and 29% Scotland.

    GC: Donegal.

    B. 69% Ireland and 31% Scotland.

    GC: Scottish lowlands and SW Scotland.

    C). 92% Ireland and 8% Scotland.

    GC: Scottish Lowlands, Donegal, Ulster and SW Scotland.

    Now for the most part in my Northern Irish matches their Genetic Communities tally up with the ethnicity results of the person.

    However, there are some anomalies as above.

    Let me explain:

    A). This person is majority Irish but clearly has a Scottish connection. Why do they not have the Ulster or Scottish Genetic Community?

    B. Why does this person have a similar ethnic make up as person A but only has Scottish Genetic Communities?

    In both of these examples, I am confused at the allocation of the GC’s. I understand that it’s all based on matches but surely if you have Ulster Scots or Ulster Irish dna you would match people from those regions? So why are these people’s GC’s so different despite having almost identical autosomal results?

    C). This person is way left field. How can one be 8% Scottish but have the Scottish Lowlands community AND the SW Scotland community? 8% is a very small number and less than a Great Grandparents genetic contribution. I have found matches as high as 50% Scottish with no Scottish GC.

    So what is happening here? Why are these disparities occurring?

    If a grandparent etc contributes enough Scottish Dna to be present in autosomal results why is it not appearing in GC’s?
    I do not believe Ancestry has the ability to tell apart Irish and Scottish as their new results suggest (as noted before my Scottish is way overstated, although in my case I think most is really English), and I think their GCs are likely more reliable (as in they are likely right, not that missing one is exclusive).

    I also think sometimes mixed people with insufficient identified cousins don't get GCs they should. Irish are pretty identified in many cases, not always. I have close Swedish and English (and maybe Welsh, depending on what's close) ancestry that doesn't suffice for a GC. So mainly I think GCs are pretty accurate for actual ancestry, but not for the absence of ancestry.

  7. #6
    Registered Users
    Posts
    2,791
    Ethnicity
    Northern Ireland
    Nationality
    Northern Irish
    Y-DNA (P)
    R1b-L193(BY2634)
    Y-DNA (M)
    R1b-M222

    Northern Ireland Ireland Scotland
    Quote Originally Posted by msmarjoribanks View Post
    I do not believe Ancestry has the ability to tell apart Irish and Scottish as their new results suggest (as noted before my Scottish is way overstated, although in my case I think most is really English), and I think their GCs are likely more reliable (as in they are likely right, not that missing one is exclusive).

    I also think sometimes mixed people with insufficient identified cousins don't get GCs they should. Irish are pretty identified in many cases, not always. I have close Swedish and English (and maybe Welsh, depending on what's close) ancestry that doesn't suffice for a GC. So mainly I think GCs are pretty accurate for actual ancestry, but not for the absence of ancestry.
    I'm not sure to be honest. I acknowledge that there are issues with splitting English, or mixed ancestry. However, I have noticed that the results when comparing myself to other Northern Irish people are quite accurate on the basis of their birthplace/religious background. This also tallies up in that most of my father's shared matches are majority Scottish, whilst most of my mothers are majority Irish. At least with respect to Brittonic Celt dna. I do think it's clear that AncestryDNA allocates Northern English and Irish + English as Scottish dna, I have come across a number of people from the Dublin area online recording a sizeable amount of Scottish in their dna, this must be from English or Norman ancestry.

    However, I do agree that the other sites have major issues in doing this, for example, for a lot of my Northern Irish matches FTDNA allocates Central Europe for their results, and then you've got LivingDNA which said I am 1.8% Northern Irish!

    If we look at the Genetic Communities, you do have a point. As an example, my father has the Scottish Lowlands community at 'very likely' for me it is 'possible', whilst my 1st cousins children do not have it. Nor do my American cousins. There must be a sharp cut off point. I wonder if this is why some of my local matches only have one community, such as Donegal.

    There are some oddities though, for example, I record 'central Donegal' but neither of my parents do?
    Last edited by Nqp15hhu; 06-26-2021 at 12:57 PM.

  8. The Following User Says Thank You to Nqp15hhu For This Useful Post:

     msmarjoribanks (06-30-2021)

  9. #7
    Registered Users
    Posts
    1,180
    Sex
    Location
    Brisbane
    Nationality
    Australian
    Y-DNA (P)
    T-P322 (T1a2b1)
    mtDNA (M)
    H6a1

    Australia Cornwall England Scotland Germany Poland
    To be fair, Northern Ireland is fairly much a worst case for Genetic Communities.
    With the amount of interchange between Scotland and Northern Ireland over just the last two millennia about as high as could be found anywhere.
    (And I do have a horse in the race, with one of my lines having matches from both Northern Ireland and Ayrshire - but no GC yet for them.)

    Whereas my Cornish are fairly much poster boys.
    One of my lines stayed within a few miles of where they were in the Domesday Book until the mid 1850s.
    And most of the rest since 1670.
    Really excellent GC in Cornwall. But in the New Land they spread further than currently shown. The original GC was more accurate.

    But my Germans came from a 50 mile radius to a 50 mile radius in the new land.
    They even have reasonably similar ethnicities. (They had lived alongside each other since about 1300 with some intermixing.)
    And lots of them have tested at Ancestry: even more than my Cornish.
    But no Genetic Community.
    Last edited by Saetro; 06-28-2021 at 07:05 PM.

  10. The Following User Says Thank You to Saetro For This Useful Post:

     Cunobelinus_T (07-02-2021)

  11. #8
    Registered Users
    Posts
    2,791
    Ethnicity
    Northern Ireland
    Nationality
    Northern Irish
    Y-DNA (P)
    R1b-L193(BY2634)
    Y-DNA (M)
    R1b-M222

    Northern Ireland Ireland Scotland
    Quote Originally Posted by Saetro View Post
    To be fair, Northern Ireland is fairly much a worst case for Genetic Communities.
    With the amount of interchange between Scotland and Northern Ireland over just the last two millennia about as high as could be found anywhere.
    (And I do have a horse in the race, with one of my lines having matches from both Northern Ireland and Ayrshire - but no GC yet for them.)

    Whereas my Cornish are fairly much poster boys.
    One of my lines stayed within a few miles of where they were in the Domesday Book until the mid 1850s.
    And most of the rest since 1670.
    Really excellent GC in Cornwall. But in the New Land they spread further than currently shown. The original GC was more accurate.

    But my Germans came from a 50 mile radius to a 50 mile radius in the new land.
    They even have reasonably similar ethnicities. (They had lived alongside each other since about 1300 with some intermixing.)
    And lots of them have tested at Ancestry: even more than my Cornish.
    But no Genetic Community.
    Yeah, I suppose you might be right. That is probably why the results are so divergent and variable across Northern Ireland. I hope in the future there is some way to fully separate out our Ulster Scots and Ulster Irish dna, and there is an explanation as to what these genetic communities actually mean.

  12. #9
    Registered Users
    Posts
    2,068
    Sex
    Location
    Chicago, Illinois
    Ethnicity
    mixed European
    Nationality
    USA
    Y-DNA (P)
    Dad: R1b/L21/DF63
    mtDNA (M)
    K2b2
    Y-DNA (M)
    R1b-M269
    mtDNA (P)
    K2b1a1a

    United States of America England Wales Sweden Germany
    Quote Originally Posted by Nqp15hhu View Post
    I have noticed that the results when comparing myself to other Northern Irish people are quite accurate on the basis of their birthplace/religious background. This also tallies up in that most of my father's shared matches are majority Scottish, whilst most of my mothers are majority Irish.
    Shared matches are reliable vs ethnicity results when it comes to close ethnicities like Irish/Scottish, so I would take those seriously. And again getting a GC I think is significant and not getting one is not.

    But also seeing accurate results with some people doesn't mean they are generally accurate. Using another category of results that are often conflated (British Isles/Scandinavian/German), mine are very close vs paper trail (spoiled only by too much of my British Isles being designated Scottish rather than English, but with the German and Swedish good), whereas my sister ends up mostly English. She doesn't have the weird Scottish thing I do, but then she loses her German and gets a bunch of Norwegian rather than Swedish. So if you looked at me I'd say they could tell Swedish from Norwegian and German from English, but if you look at my sister, I'd say they cannot. My results are really good (but for the inflated Scottish), but likely lucky.

    And if you compare to 23andMe, they are largely the same in broad strokes, but 23andMe doesn't try to get so specific. (If you look at 23andMe's efforts to get more specific, they don't do percentages but do break "British & Irish" down to highly likely UK, likely Ireland (consistent with it being mostly English and Welsh, although UK vs Ireland as a DNA group is stupid), they break "France & German" down to Germany "likely match" and Switzerland "possible" (again consistent with the paper trail), and Scandinavian down to Sweden "likely" -- again, consistent.)

    So I don't think it's surprising the Irish/Scottish/English/Wales breakdowns are right for some, but that doesn't mean I think they can consistently tell the difference -- there's too much overlapping. But I do think GCs are worth noting as additional evidence.

  13. The Following 2 Users Say Thank You to msmarjoribanks For This Useful Post:

     Cunobelinus_T (07-02-2021),  Dts (07-02-2021)

  14. #10
    Registered Users
    Posts
    2,791
    Ethnicity
    Northern Ireland
    Nationality
    Northern Irish
    Y-DNA (P)
    R1b-L193(BY2634)
    Y-DNA (M)
    R1b-M222

    Northern Ireland Ireland Scotland
    Quote Originally Posted by msmarjoribanks View Post
    Shared matches are reliable vs ethnicity results when it comes to close ethnicities like Irish/Scottish, so I would take those seriously. And again getting a GC I think is significant and not getting one is not.

    But also seeing accurate results with some people doesn't mean they are generally accurate. Using another category of results that are often conflated (British Isles/Scandinavian/German), mine are very close vs paper trail (spoiled only by too much of my British Isles being designated Scottish rather than English, but with the German and Swedish good), whereas my sister ends up mostly English. She doesn't have the weird Scottish thing I do, but then she loses her German and gets a bunch of Norwegian rather than Swedish. So if you looked at me I'd say they could tell Swedish from Norwegian and German from English, but if you look at my sister, I'd say they cannot. My results are really good (but for the inflated Scottish), but likely lucky.

    And if you compare to 23andMe, they are largely the same in broad strokes, but 23andMe doesn't try to get so specific. (If you look at 23andMe's efforts to get more specific, they don't do percentages but do break "British & Irish" down to highly likely UK, likely Ireland (consistent with it being mostly English and Welsh, although UK vs Ireland as a DNA group is stupid), they break "France & German" down to Germany "likely match" and Switzerland "possible" (again consistent with the paper trail), and Scandinavian down to Sweden "likely" -- again, consistent.)

    So I don't think it's surprising the Irish/Scottish/English/Wales breakdowns are right for some, but that doesn't mean I think they can consistently tell the difference -- there's too much overlapping. But I do think GCs are worth noting as additional evidence.
    Absolutely the shared matches approach works well and helps to identify possible lines that might be attributed to those ethnicities.

    I think you are right about your synopsis over the British isles as a whole. However, with Irish people due to the homogeneity and the lack of movement I think (though I could be wrong) these ethnic groups have more of a link?

    For example, if you look at a native of Strabane or Pomeroy in Tyrone they’re definitely likely to be 90%+ Irish on AncestryDNA in my experience but if you look at a native of Ballintoy or Ballymena they’re likely to be 60%+ Scottish on AncestryDNA (there seems to be variety on the amount of Scottish assigned in Ulster Scots, I guess it depends on mixing.) Why do I say this? Because most people’s ancestors did not move in Ulster and their ethnic results generally reflect their ethnic background or the background of the area their ancestors are from.

    You can see this if you compare the surnames in the 1901 census to the 1740 census they’re mostly the same.

    Although I’m happy to be corrected maybe this is just a coincidence in my matches.
    Last edited by Nqp15hhu; 06-30-2021 at 09:43 AM.

Similar Threads

  1. Why is E-V13 so confusing?
    By Aahmes in forum E1b-M215
    Replies: 301
    Last Post: 04-06-2021, 03:40 PM
  2. Confusing DNA result
    By aminapamina in forum Eastern
    Replies: 0
    Last Post: 02-14-2020, 10:34 AM
  3. results are confusing
    By annalubov in forum FTDNA
    Replies: 5
    Last Post: 01-01-2018, 01:00 PM
  4. "Newfoundland English" Genetic Community
    By kujira692 in forum AncestryDNA
    Replies: 0
    Last Post: 05-10-2017, 06:35 PM
  5. Confusing Y-DNA resuts
    By migoblu in forum FTDNA
    Replies: 4
    Last Post: 10-19-2013, 01:23 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •