PDA

View Full Version : Matches with People with No Overlapping Ancestry. Can Someone Explain?



Jessie
07-12-2020, 04:30 AM
Not having looked at this too closely in the past and I'll look at other places where I've tested as well but can someone explain how this happens?

On MyHeritage the country I have the most matches outside of the UK and Ireland is Norway but I've noticed that for them they are nearly fully Scandinavian and I'm Irish, Scottish and Welsh. Has anyone got an explanation for this? If this is happening to me I'm sure many other people get a similar situation.

Some examples.

https://i.imgur.com/ckE6Uvy.png

https://i.imgur.com/l1DdRSD.png

https://i.imgur.com/3AFKJ9S.png

Loads of them like the above with no overlapping ethnicity. Has anyone looked into this and come up with an explanation?

Velislav
07-12-2020, 07:32 AM
I noticed the same thing with people who are 100% Finnish. I think it is because their software is not really precise (huge differences between people from same ethnicity) and probably misses some small blocks.

Jessie
07-12-2020, 07:46 AM
I noticed the same thing with people who are 100% Finnish. I think it is because their software is not really precise (huge differences between people from same ethnicity) and probably misses some small blocks.

I have some Finnish matches also with the same issue. No overlapping ethnicity.

https://i.imgur.com/RxuJqk1.png

FionnSneachta
07-12-2020, 07:53 AM
Well their ethnicity estimates aren't 100% accurate so that could explain why there are no shared ethnicities.

msmarjoribanks
07-12-2020, 08:17 PM
I have a number of Norwegian and Finnish matches on MyH that I think aren't real matches. Many of them don't match either of my parents. One of the Norwegian ones is a 33 cM match on 2 segments, but only matches my mother on one of them (22 cM). I suppose that one could be real (I have Swedish ancestry, so I show a little "Scandinavian" and that's our only overlapping ancestry, but all of his places are Norwegian and just given the pattern with other such matches I just think MyH shows a reasonably high rate of false matches.) Part of this is that I never get to my lower cM Ancestry matches, so they may be just as likely to be just pile-up zones or what not, and it's harder to tell because Ancestry is mostly Americans.

(If this guy is a real match, he could match my mother less than me since I uploaded Ancestry and her upload was from FTDNA.)

Shadogowah
07-13-2020, 07:51 AM
I have many matches that allegdly share segments 30+ cM. However, they do not match any of my parents.

My explanation is that some recombination of my parents dna happens to match IBDs of other people.

Due to the overrepresentation of Scandinavian and Jewish customers in MyHeritage, I find it normal to have people with these heritages among my "false" matches.

Tolan
07-13-2020, 08:36 AM
Are they IBP segments?

In other words, very common segments of the European population?

You can known it by looking if this segment is triangulated with other people..

If this is the case, it is a segment which is neither Scandinavian nor British, but European in general ...

Stefanie
07-13-2020, 02:54 PM
I have matches like this at Myheritage, some at Ancestry too. I also think it’s because their ancestry composition is not very good. Also could be false matches, pile up region matches, etc. I’d be leery of their triangulation tool though. In my experience, sometimes it works, sometimes it doesn’t. For instance, I have distant/surprising matches who triangulate with known closer matches. But if I click on the known closer matches, there are many more matches who should triangulate with the distant matches, same segment, and they don’t. It makes no sense. I would write off the distant/surprising matches but sometimes I match them other places like Gedmatch or Living dna, etc. It may be a flaw in Myheritage system. Not sure.

mildlycurly
07-15-2020, 03:20 PM
My theory is pred. Irish people often match with Scandinavians due to distant Viking ancestry. I too have a lot of Scandinavian/Finnish matches and the Viking era is the only significant influx of Scandinavians into Ireland I can think of (the English may also have contributed some Scandinavian to the Irish genome but in far smaller quantities).

Regardless, Myheritage ancestry estimates aren't very good at all. The matches give a better insight into your ancestry IMO.

Luiz Fernando
07-16-2020, 02:05 AM
In My Heritage I have dozens of matches with people from Great Britain and Ireland and it is informed that the person and I do not have ethnicities in common. In these cases their ethnicities are only Irish, Scottish, and Welsh and English, ethnicities that I supposedly do not have according to My Heritage. When I look at the segments of DNA I share with them frequently I find the same segments I share with people from Sweden, Norway and Denmark, and in this cases it is informed that I have Scandinavian ethnicity in common with them (ethnicity My Heritage recognizes for me). So in my opinion these people from Great Britain and Ireland have some Scandinavian ancestry and My Heritage does not recognize it for them. Conversily, I probably have some Irish, Scottish, and Welsh and probably English ethnicity and My Heritage does not recognize it for me, and more, I think my scandinavian ethnicity comes from people from Great Britain and Ireland.

Kellebel
07-16-2020, 02:29 AM
I think it's simply because of the way they calculate their ethnicity estimate - it isn't really the most precise out there. My dad and I have 15% of overlapping ancestry, for example - not even counting all of the close matches with whom I share zero, let alone the more distant ones. Northwestern Europeans are very similar genetically - they all have the same ancestries, but in different proportions, so you easily get cases of skipping ancestries, wrong assignments or oversmoothing, especially when the used software isn't all that great to begin with.

Riverman
07-22-2020, 08:15 AM
I think these results are good, because its their ethnicity and not deep ancestry which being measued. So either if the Irish has Viking or the Norwegian has Gaelic ancestry, they are still 100 percent Irish and Norwegian respectively, since the admixture event formed not just their family, but the Irish and Norwegians as a whole, its part of their modern regional ethnicity. Otherwise you would have to look for "deep" and "historical" or even "prehistorical" ancestries, which is a completely different thing. Lucky the algorithm works so well for you, because for undersampled people your segment, which might be equally deeply shared, might just be assigned to your matches ethnicity, even though its actually deep ancestry. That's for sure not better if suddenly someone from a regional ethnicity gets the information of being "mixed" with something actually most of his compatriots have and getting confused. I think those companies should therefore do different tests, one for recent ethnicity and one for deep historical/prehistorical ancestry. Like Lebanese with Indo-European or Crusader ancestry are not recently and ethnially admixed, so they shouldn't be assigned that way, while on the other hand they might want to know their deeper ancestries in a seperate report. These are just two different things.

Joanna_1998
08-14-2020, 08:32 AM
Same happened to me, noticed this with a German guy and someone with Scandinavian ancestry too! (haven't looked deeper into that but these 2 cases surprised me as well^^)

evon
08-14-2020, 06:48 PM
My theory is pred. Irish people often match with Scandinavians due to distant Viking ancestry. I too have a lot of Scandinavian/Finnish matches and the Viking era is the only significant influx of Scandinavians into Ireland I can think of (the English may also have contributed some Scandinavian to the Irish genome but in far smaller quantities).

Regardless, Myheritage ancestry estimates aren't very good at all. The matches give a better insight into your ancestry IMO.

You have to bear in mind that Norway (and other parts of northern Europe, such as Gothenburg in Sweden and Gdansk in Poland etc) saw a huge influx of people from Scotland and the northern isles (Shetland and Faeroe Islands to Norway) during the 1500-1600's, so matches might be due to shared British-Irish ancestry and not shared Scandinavian ancestry.


Are they IBP segments?

In other words, very common segments of the European population?

You can known it by looking if this segment is triangulated with other people..

If this is the case, it is a segment which is neither Scandinavian nor British, but European in general ...

In my experience there are few, if any, European pile-up segments, but there are many that appear to be regional (north, south ect)..


I have many matches that allegdly share segments 30+ cM. However, they do not match any of my parents.

My explanation is that some recombination of my parents dna happens to match IBDs of other people.

Due to the overrepresentation of Scandinavian and Jewish customers in MyHeritage, I find it normal to have people with these heritages among my "false" matches.

This could be due to no-calls, I have many instances of this with my mother and grandmother, but which I am 100% sure are via my grandmother, so it all comes down to no-calls making the segments appear more fragmented than it actually is. PS: lots of immigrants from southern Europe living in Scandinavia (especially Sweden), so that might also explain the high matches?


Same happened to me, noticed this with a German guy and someone with Scandinavian ancestry too! (haven't looked deeper into that but these 2 cases surprised me as well^^)

We have a few Greek matches, which is due to the presence of Greek/Balkan DNA in the Romani genepool, so that might explain your matches..

Shadogowah
08-16-2020, 09:56 AM
This could be due to no-calls, I have many instances of this with my mother and grandmother, but which I am 100% sure are via my grandmother, so it all comes down to no-calls making the segments appear more fragmented than it actually is. PS: lots of immigrants from southern Europe living in Scandinavia (especially Sweden), so that might also explain the high matches?



Yes and Nope. But thanks a lot for the insight.

I have two profiles in MH because first I transferred the data from FTDNA and later I took a test with them and sent a physical sample.

I noticed that indeed some matches that showed as sharing a relative big segment for one profile split it in two in the other one thus making the relation either more distant or dubious although MH still labels them as "Medium level of confidence".

One is indeed a Swedish gentleman with a 25 cM segment but the paradigmatic case is a German lady with 31 cM.

Neither of them present a profile any close to a southern European. Most of these matches also have part of their genealogy trees visible and there are no names that suggest recent Spanish or Italian ancestry.

The key here is that our shared etnicities percentages seem to be split between a bit of their dominant Scandinavian (I score 7%) and a bit of their minimal MENA (I also score 7%).

My comment therefore referred to the possibility that MH currently is misidentifying as Scandinavian and middle Eastern / North African some contribution that perhaps is common both to berians and northern europeans due to the overrepresentation of Scandinavians and Jews among their customers.

Jessie
08-16-2020, 10:48 AM
You have to bear in mind that Norway (and other parts of northern Europe, such as Gothenburg in Sweden and Gdansk in Poland etc) saw a huge influx of people from Scotland and the northern isles (Shetland and Faeroe Islands to Norway) during the 1500-1600's, so matches might be due to shared British-Irish ancestry and not shared Scandinavian ancestry.



Shouldn't that show as Irish or English though? If it was a shared ancestry I would understand that but a lot of the matches are overlapping ancestries. I have Irish, Scottish and Welsh and they have Scandinavian. That's why I find it odd. Maybe they just use smoothing techniques?

Tolan
08-16-2020, 11:20 AM
The answer is so obvious: IBP!
Why does no one seem to believe it?
The European population is a population which has been endogamous.
If we go back 4,500 years, we don't have many ancestors and they are common to almost all Europeans.

There are still very common segments that can be found in many European countries at the same time.
For that, Jessie, you need to answer these two questions:
How many people match with you in these segments?
If so, are they from several different countries or just one?

evon
08-16-2020, 11:31 AM
Shouldn't that show as Irish or English though? If it was a shared ancestry I would understand that but a lot of the matches are overlapping ancestries. I have Irish, Scottish and Welsh and they have Scandinavian. That's why I find it odd. Maybe they just use smoothing techniques?

It is very difficult to correctly determine admixture in northern Europe using modern labels such as these, as we are so similar to each other up here in north. This is why one has to take estimates with a huge grain of salt. I would say that they get it wrong about 40% of the time.

evon
08-16-2020, 11:39 AM
My comment therefore referred to the possibility that MH currently is misidentifying as Scandinavian and middle Eastern / North African some contribution that perhaps is common both to berians and northern europeans due to the overrepresentation of Scandinavians and Jews among their customers.

Did I understand correctly? you think that the number of tested people and their background have an impact on the estimate?

If so you are mistaken, the database testers have no impact on estimates, at least not during the current model. Overrepresentation only impacts segment sharing.

evon
08-16-2020, 11:47 AM
Any one else not getting notifications from this thread? I only saw it by chance.

Jessie
08-16-2020, 12:02 PM
The answer is so obvious: IBP!
Why does no one seem to believe it?
The European population is a population which has been endogamous.
If we go back 4,500 years, we don't have many ancestors and they are common to almost all Europeans.

There are still very common segments that can be found in many European countries at the same time.
For that, Jessie, you need to answer these two questions:
How many people match with you in these segments?
If so, are they from several different countries or just one?

I don't show matches with other populations other than Northwestern European populations with the highest amount outside of Ireland and Britain being Norwegians. I just find it strange that there is no overlapping ethnicities with a few of these. The other questions you have asked I have looked at but will sometime.

Shadogowah
08-16-2020, 01:48 PM
Did I understand correctly? you think that the number of tested people and their background have an impact on the estimate?

If so you are mistaken, the database testers have no impact on estimates, at least not during the current model. Overrepresentation only impacts segment sharing.


I indeed assume that people tested, their background and genealogy is taken into account to generate an estimate and I cannot understand how it is not the case.

evon
08-16-2020, 06:01 PM
I indeed assume that people tested, their background and genealogy is taken into account to generate an estimate and I cannot understand how it is not the case.

Most DNA companies do not use samples from their tester database for their admixture estimates, and the ones that do only use a few selected samples ,that have been asked to be part of their reference database etc (23andme has a model where they use testers, but as far as I can tell they are only used for the "Timeline" feature, not the actual admixture estimation). Myheritage, LivingDNA and FTDNA etc all use a static set of reference samples for their admixture estimates (as does 23andme for the actual estimates and not the timeline feature). At some of the companies you can see this if you look at the technical data behind the admixture models they use, where they usually list the number of references samples etc..

Saetro
08-18-2020, 01:19 AM
I have way more Scandi matches at MyHeritage than I do at Ancestry and few if any at FTDNA.

To reasons previously given in this thread, I might just add that nobody has mentioned the possible influence of imputed content.
And I am not sure that I can offer anything useful on that score. Just that it could be another source of uncertainty/error.

History should always be considered.
Gustavus Adolphus' armies would have brought some Swedish into Germany during the religious wars.
Maybe not all took their DNA home with them.

C J Wyatt III
08-18-2020, 03:45 AM
History should always be considered.
Gustavus Adolphus' armies would have brought some Swedish into Germany during the religious wars.
Maybe not all took their DNA home with them.

I like to talk about seaman sailing the seven seas, speading their .... Oh nevermind!

boilermeschew827
08-18-2020, 06:23 PM
Some of my closest matches are from my Danish and Swedish lines, many of whom are 100% Scandinavian:
39045


I agree with other users' sentiment, it's just a poor calculator IMO.

Capitalis
08-25-2020, 12:09 PM
To reasons previously given in this thread, I might just add that nobody has mentioned the possible influence of imputed content.
And I am not sure that I can offer anything useful on that score. Just that it could be another source of uncertainty/error.

Good point.

I will add my own experience. My grandmother has a "high confidence" 40.1 cM match (largest segment 18.8 cM) with a Swedish gentleman; both have deep trees with no overlap (his tree is 100% Swedish). There is no third match that produces a triangulated segment.

However, unlike the other examples in this thread, they do have overlapping populations in the ethnicity estimate. My grandmother is on the left:

39153

I like MyHeritage as a company; their layout is pleasant and all of their features are easy to use. Ethnicity estimates are less useful to me than using matching to confirm the accuracy of the family tree I researched 20 years ago.

I can't explain this Swedish match in terms of genealogy and don't expect I ever will.

---

EDIT: My grandmother is 1/4 Southern Welsh. Here is a match with a 100% Southern Welsh woman; notice anything about the "errors"? My grandmother is on the left:

39154

In other words, a portion of their Southern Welsh DNA is quite likely being categorised as N+W European and Middle Eastern.

06lab
09-20-2020, 04:28 PM
I also do get russian matches with no shared locactin

https://i.imgur.com/mcvRL6R.png?1

crazyladybutterfly
01-02-2021, 08:41 AM
I match with French, iberians, north europeans, Slavs, Germans and even some balts even though I don't have ancestry from any of these regions and neither they seem to have common ancestry with me.

Same problem with family tree dna

lana6765
01-02-2021, 09:02 AM
The answer is so obvious: IBP!
Why does no one seem to believe it?
The European population is a population which has been endogamous.
If we go back 4,500 years, we don't have many ancestors and they are common to almost all Europeans.

There are still very common segments that can be found in many European countries at the same time.
For that, Jessie, you need to answer these two questions:
How many people match with you in these segments?
If so, are they from several different countries or just one?

I don’t think that’s what endogamous means. Europe has many different countries and religious groups, and has had a large population for a long time. Groups within Europe may be endogamous, but Europe is too varied to be endogamous as a whole.

Mathematically, everyone everywhere has the same ancestors if you go back far enough. There was a paper published that estimated that all Europeans have among their ancestors the same man or woman who lived around 1400.

Stellaritic
01-02-2021, 09:42 AM
Yes and Nope. But thanks a lot for the insight.

I have two profiles in MH because first I transferred the data from FTDNA and later I took a test with them and sent a physical sample.

I noticed that indeed some matches that showed as sharing a relative big segment for one profile split it in two in the other one thus making the relation either more distant or dubious although MH still labels them as "Medium level of confidence".

One is indeed a Swedish gentleman with a 25 cM segment but the paradigmatic case is a German lady with 31 cM.

Neither of them present a profile any close to a southern European. Most of these matches also have part of their genealogy trees visible and there are no names that suggest recent Spanish or Italian ancestry.

The key here is that our shared etnicities percentages seem to be split between a bit of their dominant Scandinavian (I score 7%) and a bit of their minimal MENA (I also score 7%).

My comment therefore referred to the possibility that MH currently is misidentifying as Scandinavian and middle Eastern / North African some contribution that perhaps is common both to berians and northern europeans due to the overrepresentation of Scandinavians and Jews among their customers.

There is a tendency to think that the European country with the largest population of immigrants is the U.K, but statistically more than 1/4 of the German citizens(including Angela Merkel) have at least one parent with an immigration background.

C J Wyatt III
01-02-2021, 06:49 PM
Mathematically, everyone everywhere has the same ancestors if you go back far enough. There was a paper published that estimated that all Europeans have among their ancestors the same man or woman who lived around 1400.

Perhaps this which makes the common ancestor more like a thousand years back:

https://www.livescience.com/29390-all-europeans-share-recent-ancestors.html

My research points to a much more recent common ancestor.

crazyladybutterfly
01-02-2021, 10:41 PM
I have way more Scandi matches at MyHeritage than I do at Ancestry and few if any at FTDNA.

To reasons previously given in this thread, I might just add that nobody has mentioned the possible influence of imputed content.
And I am not sure that I can offer anything useful on that score. Just that it could be another source of uncertainty/error.

History should always be considered.
Gustavus Adolphus' armies would have brought some Swedish into Germany during the religious wars.
Maybe not all took their DNA home with them.

I feel like they make Scandinavians match with everyone v😂😂😂
I don't score Scandinavian in the estimate , neither do most of the people I asked and most do match with Swedes for some reason , me included but here I seem to also match with many people from England

crazyladybutterfly
01-02-2021, 10:50 PM
Most of my matches without shown south italian ancestry on the estimate (I'm excluding anyone getting south Europe , mena, jewish) are English, german and Slavic I also match a couple of Scandinavians and a Lithuanian . I natch a couple of Iberians too but that's far more likely given my country's history .42252
42253
Both are mine

Among those with whom I share estimates I've seen Italians (south and a north italian) , a couple of Greeks and one Jew

No west asian or arab match tho.

JerryS.
01-03-2021, 04:28 AM
I feel like they make Scandinavians match with everyone v������
I don't score Scandinavian in the estimate , neither do most of the people I asked and most do match with Swedes for some reason , me included but here I seem to also match with many people from England

instead of you matching them from your possible ancient English/Scandinavian, couldn't they be matching you from their possible Italian ancestry?

crazyladybutterfly
01-03-2021, 10:48 AM
instead of you matching them from your possible ancient English/Scandinavian, couldn't they be matching you from their possible Italian ancestry?

They don't seem to have anything southern italian or even southern european, it would need to be quite diluted like 17-18th century but why would a south italian paesant even move there when the region at the time was much poorer.

I doubt we might have common ancestry past medieval age anyway , these apps aren't reliable with less than 100 cM

crazyladybutterfly
01-03-2021, 10:58 AM
instead of you matching them from your possible ancient English/Scandinavian, couldn't they be matching you from their possible Italian ancestry?

A southern Europan admixture would definitely show , if not with mena/south europe theyd score central europeab as minimum.

if it was recent enough, im sure they dont have anything italian in the past 200 years.

while the other way around it could be hidden especially if the person scores mena but I dont think its really the case

The fact we can share as much as 40 cm just shows it's unreliable as a service unless we are talking about close relatives.

Shadogowah
01-03-2021, 02:10 PM
There is a tendency to think that the European country with the largest population of immigrants is the U.K, but statistically more than 1/4 of the German citizens(including Angela Merkel) have at least one parent with an immigration background.

That would bring us again to the way they build their models. I indeed assumed some sort process where clusters are built and labeled using Big Data and they are fed with a big number of samples that can only come from customer profiles. You teach the mathematical thing what "Scandinavian" is by crushing numbers coming from your DB. You could refine by choosing those who report to be "native" to some degree, e.g. like having 4 Swedish grandparents, but even then it could happen (for sure it happens, as you point out) that many have a great-grandmother that was Jewish, Spanish, Italian or Portuguese or whatever. So now the algorithm gets that being "Scandinavian" also implies a bit of southern and eastern flavors. It would also explain why some categories pop up much more than others in people totally unrelated to them (like myself).

But I was told that it is not the case and the samples to build the models do not come from the customers. No Big Data then.

Shadogowah
01-03-2021, 02:15 PM
They don't seem to have anything southern italian or even southern european, it would need to be quite diluted like 17-18th century but why would a south italian paesant even move there when the region at the time was much poorer.

I doubt we might have common ancestry past medieval age anyway , these apps aren't reliable with less than 100 cM

Exactly. Until we go quite into the Industrial era, the Nordics and Central Europe actually were a source of immigrants, not a destination.

JerryS.
01-03-2021, 03:00 PM
That would bring us again to the way they build their models. I indeed assumed some sort process where clusters are built and labeled using Big Data and they are fed with a big number of samples that can only come from customer profiles. You teach the mathematical thing what "Scandinavian" is by crushing numbers coming from your DB. You could refine by choosing those who report to be "native" to some degree, e.g. like having 4 Swedish grandparents, but even then it could happen (for sure it happens, as you point out) that many have a great-grandmother that was Jewish, Spanish, Italian or Portuguese or whatever. So now the algorithm gets that being "Scandinavian" also implies a bit of southern and eastern flavors. It would also explain why some categories pop up much more than others in people totally unrelated to them (like myself).

But I was told that it is not the case and the samples to build the models do not come from the customers. No Big Data then.

why would data used by Ancestry* produce a little different results than Eurogenes G25 using the same raw data from Ancestry*? don't they use the same population sources? which is more accurate, Ancestry* results based on its own pop samples or G25 results using Ancestry raw data but G25 pop samples? I mean what I see being done here more often than not is using raw data from a commercial tester and applying it to G25 coords and then vahaduo resulting it.

crazyladybutterfly
01-03-2021, 05:53 PM
Exactly. Until we go quite into the Industrial era, the Nordics and Central Europe actually were a source of immigrants, not a destination.

Exactly I can only think some weird family ties among european nobility but my family is humble lol

I also have matches with people from eastern Europe like Poland , Russia etc why the hell would an italian go there ?

Shadogowah
01-03-2021, 07:14 PM
Exactly I can only think some weird family ties among european nobility but my family is humble lol

I also have matches with people from eastern Europe like Poland , Russia etc why the hell would an italian go there ?

My bet is Sephardic Jews.

Shadogowah
01-03-2021, 07:24 PM
why would data used by Ancestry* produce a little different results than Eurogenes G25 using the same raw data from Ancestry*? don't they use the same population sources? which is more accurate, Ancestry* results based on its own pop samples or G25 results using Ancestry raw data but G25 pop samples? I mean what I see being done here more often than not is using raw data from a commercial tester and applying it to G25 coords and then vahaduo resulting it.

They key in this case is which is the source coordinates used. That is the model. I don't know which one the comercial companies use and as I explain above, I thought they created their own using statistical methods applied to their customers data.

----
Corrected. I meant the source, not the target

crazyladybutterfly
01-03-2021, 09:11 PM
My bet is Sephardic Jews.

Even worse , that should be more noticeable in a test of a north european lol
For a south european it's harder to tell because they have mena regardless if they are part jews or not, but if someone is let's say British even if he is only 1/32 would show up because more two ethnicities are distant the easier it is to discern the two, you can't do that as successfully with similar ethnicities.
These people I've matched with literally get 100% north european or east European.

I admit I do have some Jewish matches but they're very few somehow. On my heritage I have no one from Israel on lftdna I have one

JerryS.
01-03-2021, 09:25 PM
Even worse , that should be more noticeable in a test of a north european lol
For a south european it's harder to tell because they have mena regardless if they are part jews or not, but if someone is let's say British even if he is only 1/32 would show up because more two ethnicities are distant the easier it is to discern the two, you can't do that as successfully with similar ethnicities.
These people I've matched with literally get 100% north european or east European.

I admit I do have some Jewish matches but they're very few somehow. On my heritage I have no one from Israel on lftdna I have one

1/32 is only 3%. In populations with no known connections to distant populations that is usually viewed as "noise". we see this all the time with "100%" Irish and English people who get ~3% something from the far east et al.

christa
01-04-2021, 12:14 AM
They don't seem to have anything southern italian or even southern european, it would need to be quite diluted like 17-18th century but why would a south italian paesant even move there when the region at the time was much poorer.

In the Middle Ages many Italian merchants lived in England, especially in the 1400s and 1500s

crazyladybutterfly
01-04-2021, 08:35 AM
In the Middle Ages many Italian merchants lived in England, especially in the 1400s and 1500s

They were from north italy and I dont have matches from the region and that s also way too far back in time
These matches are comparable with ancestry which doesn't go back 1800s and thats why I m skeptical.

crazyladybutterfly
01-04-2021, 08:36 AM
1/32 is only 3%. In populations with no known connections to distant populations that is usually viewed as "noise". we see this all the time with "100%" Irish and English people who get ~3% something from the far east et al.

Someone who is 97% swedish and 3% Sephardic jewish would probably get central , south European or Ashkenazi.

Shadogowah
01-04-2021, 09:06 AM
Even worse , that should be more noticeable in a test of a north european lol
For a south european it's harder to tell because they have mena regardless if they are part jews or not, but if someone is let's say British even if he is only 1/32 would show up because more two ethnicities are distant the easier it is to discern the two, you can't do that as successfully with similar ethnicities.
These people I've matched with literally get 100% north european or east European.

I admit I do have some Jewish matches but they're very few somehow. On my heritage I have no one from Israel on lftdna I have one

My point from my very first post on this subject was that whatever those hypothetical small "southern" DNA fragments are and come from, they somehow have been assumed by the algorithm to belong to "nordic" or "eastern" clusters because it has found many more samples from those regions carrying them (due to overrepresentation).

It would explain why some of us, mediterranean people, score scandinavian or eastern european despite not having a single evidence pointing in that direction both in our genealogy or even our land's modern History.

(Yes, perhaps the Normans in the middle ages, but they were not many and seems very far fetched)

christa
01-04-2021, 10:40 AM
They were from north italy and I dont have matches from the region and that s also way too far back in time
These matches are comparable with ancestry which doesn't go back 1800s and thats why I m skeptical.

According to historian Michael Wayatt, there was "a small but influential community" of Italians "that took shape in England in the 15th century initially consisting of ecclesiastics, renaissance humanists, merchants, bankers, and artists."
Mostly of these italians were of Northern or Central Italy.
For history of Italians in England during the centuries:
https://it.m.wikipedia.org/wiki/Italo-britannici

Reading some family trees of my matches in Myheritage or Ftdna I have found several common ancestors in 1700.Few people have family tree more ancient, so I can't verify older links.

lana6765
01-04-2021, 07:20 PM
According to historian Michael Wayatt, there was "a small but influential community" of Italians "that took shape in England in the 15th century initially consisting of ecclesiastics, renaissance humanists, merchants, bankers, and artists."
Mostly of these italians were of Northern or Central Italy.
For history of Italians in England during the centuries:
https://it.m.wikipedia.org/wiki/Italo-britannici

Reading some family trees of my matches in Myheritage or Ftdna I have found several common ancestors in 1700.Few people have family tree more ancient, so I can't verify older links.

Interesting.

I seem to match some Italians. They’re really distant, with the closest sharing a mere 23.4‎ cM with me, but I do score some Southern European on every commercial test except AncestryDNA.

I might have an Iberian ancestor, but the Italian matches are of slightly better quality.

But my weirdest match of all is a 100% Finnish person who I share a single 37.6‎ cM segment with. I score no Finnish, only Scandinavian, Irish and Greek/South Italian.

omgpeeps
01-05-2021, 02:52 AM
I have some matches with no shared ancestry as well on My Heritage, and I think in most cases It's probably the ethnicity algorithm's fault since I don't think they give the best ethnicity results, but I also believe that some of my much lower cM matches (below 9-10 cM I guess) are maybe false positives combined with no shared ancestry or evident genealogical link. For me, this problem is virtually non-existent on 23andme matches.

crazyladybutterfly
01-05-2021, 02:07 PM
According to historian Michael Wayatt, there was "a small but influential community" of Italians "that took shape in England in the 15th century initially consisting of ecclesiastics, renaissance humanists, merchants, bankers, and artists."
Mostly of these italians were of Northern or Central Italy.
For history of Italians in England during the centuries:
https://it.m.wikipedia.org/wiki/Italo-britannici

Reading some family trees of my matches in Myheritage or Ftdna I have found several common ancestors in 1700.Few people have family tree more ancient, so I can't verify older links.

The weird thing is that I have no north and central italian matches and I assure you I'm of humble origins on both sides. I find this scenario just as unlikely as the opposite route.

Another weird thing is that the Italian peaks in england instead of southern italy when you go for distant time periods , not only that also the "Montenegrin" peaks in france

crazyladybutterfly
01-05-2021, 02:15 PM
My point from my very first post on this subject was that whatever those hypothetical small "southern" DNA fragments are and come from, they somehow have been assumed by the algorithm to belong to "nordic" or "eastern" clusters because it has found many more samples from those regions carrying them (due to overrepresentation).

It would explain why some of us, mediterranean people, score scandinavian or eastern european despite not having a single evidence pointing in that direction both in our genealogy or even our land's modern History.

(Yes, perhaps the Normans in the middle ages, but they were not many and seems very far fetched)

The Scandinavian , etc in the estimate is just trying to balance , I think italian on "my heritage" means actually laziale/Tuscan which is why north italians get Scandinavian, and other more northern ethnicities while south italians get some mena and some italians get a mix of both. Their "Italian" is way too much "limited".

As for medieval etc admixture , it shouldn't show off simply because now that foreign Admixture is part of the modern population admixture. My heritage compares you to modern people not ancient or medieval . the Norman admixture is low but by that logic Balkan people should score 50% Ukrainian or something similar which doesn't usually happen .
Regardless medieval admixture shouldnt match you with distant foreigners with whom you share a medieval ancestors , its too far back in time and now the inherited dna isn't only diluted but quite fragmented.

crazyladybutterfly
01-05-2021, 02:23 PM
I have some matches with no shared ancestry as well on My Heritage, and I think in most cases It's probably the ethnicity algorithm's fault since I don't think they give the best ethnicity results, but I also believe that some of my much lower cM matches (below 9-10 cM I guess) are maybe false positives combined with no shared ancestry or evident genealogical link. For me, this problem is virtually non-existent on 23andme matches.

I actually share an high number of SNPs with many I dont get it.

christa
01-05-2021, 06:08 PM
The weird thing is that I have no north and central italian matches and I assure you I'm of humble origins on both sides. I find this scenario just as unlikely as the opposite route.

Another weird thing is that the Italian peaks in england instead of southern italy when you go for distant time periods , not only that also the "Montenegrin" peaks in france

Many kits used for genetic groups are of Americans. For example if You see the genetic group of Bari (Italy) You have to see the map 1950-2000, it's coloured America and South Italy, so in this case they have used to form this genetic group Italians of Apulia and Americans of apulians and others origins (England and other nations).
The map of 1600-1650 indicated the areas of people on the genealogies of the kits that have a family trees. For Italy (Bari), 788 kits are used to form this group, but only 315 have family trees. These trees are used for the maps of different periods.
Maybe someone know better their enghish ancestors in genealogies and they know only few generations of Italians ancestors.For this reason the map of 1600-1650 is more colorful in England.

crazyladybutterfly
01-06-2021, 07:46 AM
Many kits used for genetic groups are of Americans. For example if You see the genetic group of Bari (Italy) You have to see the map 1950-2000, it's coloured America and South Italy, so in this case they have used to form this genetic group Italians of Apulia and Americans of apulians and others origins (England and other nations).
The map of 1600-1650 indicated the areas of people on the genealogies of the kits that have a family trees. For Italy (Bari), 788 kits are used to form this group, but only 315 have family trees. These trees are used for the maps of different periods.
Maybe someone know better their enghish ancestors in genealogies and they know only few generations of Italians ancestors.For this reason the map of 1600-1650 is more colorful in England.

And why can't they fix this simple problem ?
Montenegrin ethnicity colors france but no Balkans.
So basically they're giving faux results .42353
4235442353

42351
42352

crazyladybutterfly
01-06-2021, 07:49 AM
Many kits used for genetic groups are of Americans. For example if You see the genetic group of Bari (Italy) You have to see the map 1950-2000, it's coloured America and South Italy, so in this case they have used to form this genetic group Italians of Apulia and Americans of apulians and others origins (England and other nations).
The map of 1600-1650 indicated the areas of people on the genealogies of the kits that have a family trees. For Italy (Bari), 788 kits are used to form this group, but only 315 have family trees. These trees are used for the maps of different periods.
Maybe someone know better their enghish ancestors in genealogies and they know only few generations of Italians ancestors.For this reason the map of 1600-1650 is more colorful in England.

They shouldn't use Americans and mixed people to make ethnicities lol

christa
01-06-2021, 01:23 PM
They should use only Europeans to form genetic groups of Europe.

Suri
01-10-2021, 04:25 PM
I match a lot of Finns and Swedes distantly. I do think they are real matches though, because there seems to be a match "cluster". They all match both me and my mother and then also match the same people that also match with us. Like a circle. Many of them don't share any ethnicity with us while others seem to have some Southern European ancestors even if they are from a different group. From that I am assuming there was some movement from Southern Europe to Scandinavia. I noticed that almost all of them have South Asian ethnicity as well which could mean that they are Roma.