Is the Central Italian component in Living DNA a good proxy for Roman ancestry?

Maciamo

Veteran member
Admin
Messages
9,948
Reaction score
3,228
Points
113
Location
Lothier
Ethnic group
Italo-celto-germanic
I now have the results of several Belgian people who tested with Living DNA. All got a bit of Scandinavian, and some got a from 1 to 3% of French, Basque, West Balkans, North Anatolian or Kurdish. But what really stood out is that all of them had considerable amount of Italian DNA, and in all but one samples that was exclusively Central Italian/Tuscan DNA. It ranged from 11% (in West Flanders) to 19% (in central Wallonia), with an average of about 15%. This is what I originally estimated to be the genetic contribution of the Romans to the Belgian gene pool, before I revised that estimate down to 5-10% based on the more conservative 'Italian' and 'Broadly South European' components in 23andMe.

Estimating Roman ancestry based on Y-DNA alone is very difficult because the Romans shared most of their haplogroups (such as R1b-U152, G2a-L497, G2a-U1, J2b2a-L283 and maybe even E-V13) with the Hallstatt and La Tène Celts (including the Gauls and Belgae). It's also not clear how haplogroups such as E-M34, E-M81, E-V22, J1, J2a, L1 and T1a got to the Benelux, although the ancient Romans/Italians are prime candidates. Altogether the former group (Italo-Celtic) makes up about 17% of the population. The latter group, that is to say the more recent Middle Eastern influx to the East Med, probably derived for the most part from the Kura-Araxes culture via Greece, amounts to 7.5%.

Based on the modern Y-DNA frequencies in the Latium, each group represent about half of the non-Germanic haplogroups. The ancient and modern Latium might be very different in terms of Y-DNA, but as it is the only good proxy we have at the moment, it is reasonable to say that 15% (± 5%) of Belgian Y-DNA might be of Roman origin. And that is exactly what the autosomal results from Living DNA say! At least now we know that, among all the potential Mediterranean populations that could have contributed to the above haplogroups, the Belgians correlate almost exclusively with the Central Italians autosomally. And indeed there is no other reasonable historical explanation for their presence. 5 or 6 years ago we might still have imagined that E-M34, E-V22, J1, J2a and T1a might have come to the Low Countries with Neolithic farmers. But that can be pretty much ruled out now. Even the G2a in Belgium is mostly of post-Indo-European origin, not Neolithic remnants from the LBK, Rössen or Michelsberg cultures.

Living DNA's autosomal composition isn't perfect yet, as all Belgians are reported to be mostly a mix of Southeast England, East Anglia, South England, Devon and Yorkshire. Overall Belgians get between 60% and 85% of British, depending on the individual, due to the close proximity and historical connections (Belgae, Anglo-Saxons) between the two populations. That should be resolved soon when more samples from the Benelux get tested. But when it comes to clearly differentiated populations (i.e. not immediate neighbours), like Central Italians, the results are pretty amazing.

I am really interested in seeing how other Europeans from former parts of the Roman Empire, especially outside Italy (too high Roman ancestry) and Britain (too low Roman ancestry) will turn out. Please share your results.
 
I now have the results of several Belgian people who tested with Living DNA. All got a bit of Scandinavian, and some got a from 1 to 3% of French, Basque, West Balkans, North Anatolian or Kurdish. But what really stood out is that all of them had considerable amount of Italian DNA, and in all but one samples that was exclusively Central Italian/Tuscan DNA. It ranged from 11% (in West Flanders) to 19% (in central Wallonia), with an average of about 15%. This is what I originally estimated to be the genetic contribution of the Romans to the Belgian gene pool, before I revised that estimate down to 5-10% based on the more conservative 'Italian' and 'Broadly South European' components in 23andMe.

Estimating Roman ancestry based on Y-DNA alone is very difficult because the Romans shared most of their haplogroups (such as R1b-U152, G2a-L497, G2a-U1, J2b2a-L283 and maybe even E-V13) with the Hallstatt and La Tène Celts (including the Gauls and Belgae). It's also not clear how haplogroups such as E-M34, E-M81, E-V22, J1, J2a, L1 and T1a got to the Benelux, although the ancient Romans/Italians are prime candidates. Altogether the former group (Italo-Celtic) makes up about 17% of the population. The latter group, that is to say the more recent Middle Eastern influx to the East Med, probably derived for the most part from the Kura-Araxes culture via Greece, amounts to 7.5%.

Based on the modern Y-DNA frequencies in the Latium, each group represent about half of the non-Germanic haplogroups. The ancient and modern Latium might be very different in terms of Y-DNA, but as it is the only good proxy we have at the moment, it is reasonable to say that 15% (± 5%) of Belgian Y-DNA might be of Roman origin. And that is exactly what the autosomal results from Living DNA say! At least now we know that, among all the potential Mediterranean populations that could have contributed to the above haplogroups, the Belgians correlate almost exclusively with the Central Italians autosomally. And indeed there is no other reasonable historical explanation for their presence. 5 or 6 years ago we might still have imagined that E-M34, E-V22, J1, J2a and T1a might have come to the Low Countries with Neolithic farmers. But that can be pretty much ruled out now. Even the G2a in Belgium is mostly of post-Indo-European origin, not Neolithic remnants from the LBK, Rössen or Michelsberg cultures.

Living DNA's autosomal composition isn't perfect yet, as all Belgians are reported to be mostly a mix of Southeast England, East Anglia, South England, Devon and Yorkshire. Overall Belgians get between 60% and 85% of British, depending on the individual, due to the close proximity and historical connections (Belgae, Anglo-Saxons) between the two populations. That should be resolved soon when more samples from the Benelux get tested. But when it comes to clearly differentiated populations (i.e. not immediate neighbours), like Central Italians, the results are pretty amazing.

I am really interested in seeing how other Europeans from former parts of the Roman Empire, especially outside Italy (too high Roman ancestry) and Britain (too low Roman ancestry) will turn out. Please share your results.

Interesting, I got 3.5% Italian (1.8% Tuscany and 1.6% Sardinian). I just figured my Italian results were a proxy of my French Canadian Ancestry and my 3% East European ancestry as a nod to my polish 3rd Great Grandmother. If it's true that I am as French as I am Roman, that would be quite impressive. Are there any events in Belgian history that would cause your Roman ancestry to be so high? :)
 
Interesting; I remember seeing elsewhere a few Brits who had Tuscan as well, though most did not have it.
They give me 68,8% GB (most SE England, Lincolnshire, East Anglia), and then interestingly 19% Scandinavian and 1,5% Germanic (I would expect it the other way around; I am mostly central/south Dutch; calculators show me closest to South Dutch and West German second; the Frankish heartland). I do have 1,5% Iberian, 1,5% West-Balkan and 2,5% Kurdish, which I tend to connect to my 1/16 Jewish. The remainder 4,1% is SE Asian, stemming from colonial times.
I don't have the Italian component; I do maybe have a Roman Y-lineage, since my closests Y37-matches seem to be a Z145+-cluster; I am waiting for SNP results below U152 though; could be totally different.

It's now waiting for the German project (and hopefully the Benelux; I can't take part since I am an urban mixture).
 
I now have the results of several Belgian people who tested with Living DNA. All got a bit of Scandinavian, and some got a from 1 to 3% of French, Basque, West Balkans, North Anatolian or Kurdish. But what really stood out is that all of them had considerable amount of Italian DNA, and in all but one samples that was exclusively Central Italian/Tuscan DNA. It ranged from 11% (in West Flanders) to 19% (in central Wallonia), with an average of about 15%. This is what I originally estimated to be the genetic contribution of the Romans to the Belgian gene pool, before I revised that estimate down to 5-10% based on the more conservative 'Italian' and 'Broadly South European' components in 23andMe.

Estimating Roman ancestry based on Y-DNA alone is very difficult because the Romans shared most of their haplogroups (such as R1b-U152, G2a-L497, G2a-U1, J2b2a-L283 and maybe even E-V13) with the Hallstatt and La Tène Celts (including the Gauls and Belgae). It's also not clear how haplogroups such as E-M34, E-M81, E-V22, J1, J2a, L1 and T1a got to the Benelux, although the ancient Romans/Italians are prime candidates. Altogether the former group (Italo-Celtic) makes up about 17% of the population. The latter group, that is to say the more recent Middle Eastern influx to the East Med, probably derived for the most part from the Kura-Araxes culture via Greece, amounts to 7.5%.

Based on the modern Y-DNA frequencies in the Latium, each group represent about half of the non-Germanic haplogroups. The ancient and modern Latium might be very different in terms of Y-DNA, but as it is the only good proxy we have at the moment, it is reasonable to say that 15% (± 5%) of Belgian Y-DNA might be of Roman origin. And that is exactly what the autosomal results from Living DNA say! At least now we know that, among all the potential Mediterranean populations that could have contributed to the above haplogroups, the Belgians correlate almost exclusively with the Central Italians autosomally. And indeed there is no other reasonable historical explanation for their presence. 5 or 6 years ago we might still have imagined that E-M34, E-V22, J1, J2a and T1a might have come to the Low Countries with Neolithic farmers. But that can be pretty much ruled out now. Even the G2a in Belgium is mostly of post-Indo-European origin, not Neolithic remnants from the LBK, Rössen or Michelsberg cultures.

Living DNA's autosomal composition isn't perfect yet, as all Belgians are reported to be mostly a mix of Southeast England, East Anglia, South England, Devon and Yorkshire. Overall Belgians get between 60% and 85% of British, depending on the individual, due to the close proximity and historical connections (Belgae, Anglo-Saxons) between the two populations. That should be resolved soon when more samples from the Benelux get tested. But when it comes to clearly differentiated populations (i.e. not immediate neighbours), like Central Italians, the results are pretty amazing.

I am really interested in seeing how other Europeans from former parts of the Roman Empire, especially outside Italy (too high Roman ancestry) and Britain (too low Roman ancestry) will turn out. Please share your results.

If this is the case in that central italians are close to the belgae , then one must investigate the central Italian Umbri people and where they originate from, they are far older than the etruscans and IF one was to follow italian scholars , the Umbri where from upper rhine area....( but I have not checked this out ).

The Germanics of the Belgae should be a later intruder
 
File:Romancoloniae.jpg

There was no massive population influx like in some other zones (Balkans). So I would say yes.
 
Last edited:
My grandfather was born in Frosinone, Latzio, but his father came from Abruzzo. He never gets any central (or much of any) Italian.

This is Living DNA
Henry Living DNA.JPG

Here are the others for reference
Henry rest.JPG
 
I'm not an expert of genetic but maybe is correlated with Italic ancestry...Central Italy was inhabited by a lot of tribes, not just Romans. Other Latins, Sabines, Picenes, Etruscans, Umbrians, Faliscans etc..
 
Interesting; I remember seeing elsewhere a few Brits who had Tuscan as well, though most did not have it.
They give me 68,8% GB (most SE England, Lincolnshire, East Anglia), and then interestingly 19% Scandinavian and 1,5% Germanic (I would expect it the other way around; I am mostly central/south Dutch; calculators show me closest to South Dutch and West German second; the Frankish heartland). I do have 1,5% Iberian, 1,5% West-Balkan and 2,5% Kurdish, which I tend to connect to my 1/16 Jewish. The remainder 4,1% is SE Asian, stemming from colonial times.
I don't have the Italian component; I do maybe have a Roman Y-lineage, since my closests Y37-matches seem to be a Z145+-cluster; I am waiting for SNP results below U152 though; could be totally different.

It's now waiting for the German project (and hopefully the Benelux; I can't take part since I am an urban mixture).

I am not surprised by your lack of Italian admixture as the Romans only settled in a few specific parts of the Netherlands, cities along the Meuse and Rhine like Maastricht (historically a Belgian city), Nijmegen, Utrecht and Leiden. These were essentially military garrisons to protect the borders and may have been populated by legionaries from any part of the empire, including Iberia, the Balkans and Anatolia (hence your various admixtures from these regions).

In contrast, most of Belgium, except the coastal areas and Kempen, were heavily settled by the Romans, who had hundreds of villas and (re-)founded cities/towns like Tournai, Mons, Kortrijk, Tongeren, Namur, Gembloux, Ciney, Huy and Arlon. Tongeren was one of the first Roman colony outside of Italy and Spain, founded c. 50 BCE like Paris.

One of my ancestral villages had three Roman villas, the same concentration per km² as you'd expected in the Italian countryside. Parts of Roman Belgium might have been ethnically close to the Italo-Celtic (Gallo-Roman) mix found in northern Italy at the time. Interestingly, castle-farms were built next these Roman villas in the Middle Ages and covered the same territories, as if continuing the Roman estates in a nearly uninterrupted manner. That may be because Belgium was the only part of the Western Roman Empire that never suffered invasions from Germanic tribes. The Franks were the only Germanic tribe who was allowed to settled peacefully in Roman times and became thoroughly Romanised, providing many Roman generals and senators, protecting the borders against other Germanics, and eventually rebuilding the Roman Empire under Charlemagne. That is why the Roman population did not suffer as much as elsewhere. In Britain, most of the Roman citizens fled back to Italy when the Anglo-Saxons arrived, explaining the very low Italian admixture among modern Britons. In the Netherlands or the Rhine, legions also moved away from the border to defend Italy, leaving Roman cities depopulated. That didn't happen in Belgium.

Once you remove the more recent Germanic haplogroups from the Belgian population, the closest match is indeed northern to central Italy. Here are the inferred Y-DNA frequencies of Gallo-Roman Belgium, once Germanic haplogroups are excluded. I kept all the E1b1b and J2 as it's not clear yet if the Franks brought any of it.

E-M34 : 2%
E-M78 : 9%
G2a : 10%
J1 : 1%
J2a : 5%
J2b : 1%
L1 : 0.5%
R1b-L23/Z2103 : 6%
R1b-U152 : 30.5%
R1b-P312(xU152) : 30.5%
T1a : 4%

If we also exclude Germanic haplogroups in Italy, these frequencies are very similar to Tuscany and Lombardy. The main differences are:

1) Belgium has less J2, but more T1a. This could be due either to a founder effect among the Roman colonists or to a too small sample size for minor haplogroups in Belgium. Minor haplogroups vary also widely between studies in Italian provinces.

2) Belgium has proportionally a bit more R1b-P312 (L21, DF27, etc.) than U152 compared to Italy, which is expected considering the higher Gaulish ancestry.

Everything else is nearly a perfect match.
 
Last edited:
I'm not an expert of genetic but maybe is correlated with Italic ancestry...Central Italy was inhabited by a lot of tribes, not just Romans. Other Latins, Sabines, Picenes, Etruscans, Umbrians, Faliscans etc..

Of course. It depends what you consider 'Roman', or rather 'when'. My definition of Roman here is inhabitants of the Roman Republic before the expansion outside Italy (i.e. before the Punic Wars). Therefore they include the Latins, Sabines, Faliscans, Umbrians, Etruscans, etc.

At present Living DNA only divides Italy in four parts: North, Centre, South and Sardinia. It will be interesting to see the correlation between Italian provinces and other Europeans once the scale is further refined. But it's quite telling that the current match with Belgians is almost exclusively Central Italians (and they use Tuscany as a population reference, so it's not really all Central Italy). I had already noticed that on the PCA plot in first version of 23andMe a few Belgians were clustered both with French-Germans and Tuscans (but no other part of Italy). One of my HLA types is also found almost exclusively in central Italy. I think that the Roman ancestry is quite incontrovertible.
 
If you don't mind me to ask for how many generations have your accenstors been in Abruzzo?

I don't know. We always just knew about Frosinone and some of the aunts were in contact with the family there. When I did DNA testing I found that most of the Italian connections (that I could reasonably trace) were from western Abruzzo. I dug deeper and found that both my grandfather and his brother listed Arischia as their birthplace. This is in L'Aguilla, Abruzzo and due west of Rieta in the map below.

central_italy.JPG

I used the surname website that Angela posted and saw that his surname Capanna, is common in western Abruzzo, but not Frosinone. His mother's surname, Quatrini is common in Frosinone but not L'Aquilla.
 
Last edited:
I don't know. We always just knew about Frosinone and some of the aunts were in contact with the family there. When I did DNA testing I found that most of the Italian connections (that I could reasonably trace) were from western Abruzzo. I dug deeper and found that both my grandfather and his brother listed Arischia as their birthplace. This is in L'Aguilla, Abruzzo and due west of Rieta in the map below.

View attachment 9072

I used the surname website that Angela posted and saw that his surname Capanna, is common in western Abruzzo, but not Frosinone. His mother's surname, Quatrini is common in Frosinone but not L'Aquilla.

Interesting. The Abruzzese provided the Rome countryside (Agro Romano) with much of its agricultural labour force for centuries.

Be careful with your spelling.
It's L'Aquila and Rieti.

BTW L'Aquila is east of Rieti.
 
My grandfather was born in Frosinone, Latzio, but his father came from Abruzzo. He never gets any central (or much of any) Italian.

This is Living DNA
View attachment 9068

Here are the others for reference
View attachment 9069

Ecohaidh, what I find confusing is these results. They are of a 100% Italian, from the Abruzzi and Lazio? It's fine and expected he would get little to no Northern Italian and Tuscan, but where is his southern Italian, and/or Greek?

If I'm not misreading them, then Living DNA has some serious problems in detecting Italian ancestry.
 
Maybe it is coincidence that your Italian genetic cousins did not test themselves thus the result end up like this.
 
Ecohaidh, what I find confusing is these results. They are of a 100% Italian, from the Abruzzi and Lazio? It's fine and expected he would get little to no Northern Italian and Tuscan, but where is his southern Italian, and/or Greek?
If I'm not misreading them, then Living DNA has some serious problems in detecting Italian ancestry.
Yeah, and I only see 20 something percent of his genome in each of those images; Ecohaidh, could you show us what the remaining 75-80 percent is like?
I'll admit, I think he's posting his own results and what he got from his grandfather (since he's only showing roughly 25 percent of his (assuming those are his scores and not his grandfather's) total inheritance in each of those images).
 
Ecohaidh, what I find confusing is these results. They are of a 100% Italian, from the Abruzzi and Lazio? It's fine and expected he would get little to no Northern Italian and Tuscan, but where is his southern Italian, and/or Greek?

If I'm not misreading them, then Living DNA has some serious problems in detecting Italian ancestry.

I cannot say with certainty whether he had any unusual ancestry. My aunts who visited his sisters in Italy did not mention anything interesting like that.

If you note the other companies (MyHeritage is excluded, but about the same), they are all over the place as well. The second version of FTDNA is also very different that the earlier version. The first version had a distinct west Europe (Iberian) tilt. The new one is now tilting to east Europe.

It may be that all of the companies has issues with Italy.
 
Last edited:
Yeah, and I only see 20 something percent of his genome in each of those images; Ecohaidh, could you show us what the remaining 75-80 percent is like?
I'll admit, I think he's posting his own results and what he got from his grandfather (since he's only showing roughly 25 percent of his (assuming those are his scores and not his grandfather's) total inheritance in each of those images).

The rest is generally Ireland. Here is a picture. Both MyHeritage and Living DNA have me at 84% Irish or Great Britain and Ireland. That is why I included some south England numbers in the original graphic to make up the about 25% that he needs.

Henry all.JPG
 
Thanks. You mention that your grandfather scores little to no Italian, are you guessing this based on what you scored?
 

This thread has been viewed 32012 times.

Back
Top