I will be posting in english since it's easier for me to do so without butchering grammar too much, and also because it opens the topic to whoever might be interested in contributing and discussing. I was interested in making a simple study about the variation of different sources of ancestry within non-Basque Iberian populations, so I made a simple model in nMonte3 using regular/unscaled G25 coordinates with the standard penalty, as advised by Huijbregts. Some people prefer scaled coordinates, but I'm sticking to my guns.
I'm not striving for exact ancestry proportions, but rather on how these vary across the peninsula. There are multiple ways of making these models, I decided to go by populations from around the Iron and Bronze Ages, however since there are no IA samples form Iberia avaliable, I used a few others as proxies. I'm not dogmatic about the model choice, if you feel these populations should be changed, and a few others added feel free to give your input or post your own models (G25 unscaled, pen=0.001). I'll make a quick explanation of why I used these references before moving on to the results.
Italy_Medieval_Collegno_o1 - Since Ryu's report about the Italian study placed the bulk of Imperial Roman samples around modern day South Italians/Sicilians/Greeks, I'm using these individuals as a proxy for Imperial Roman ancestry. It might not be perfect, but it should be close enough for its purpose;
Hallstatt_Bylany - Proxy for Celtic-like ancestry in Iberia, I'm keeping this reference for the bulk of Steppe-related ancestry in the model. I'm not expecting pre-Roman Iberians to be exactly like Hallstatt Celts of the Czech Republic, so I'll add something else to compensate;
Iberia_Central_CA - So I chose this one. I suppose any Bronze Age Iberian samples without Steppe ancestry would work too, but for some reason I'm getting slightly better fits with this one, hence my choice. I'd wager Hallstatt+Iberia_CA to represent pre-Roman ancestry in Iberia without extra Mediterranean ancestry;
Levant_BA_North - Proxy for Levantine ancestry, it's more recent and contains higher Armenia_EBA (which I won't be using in the model) ancestry than Levant_BA_South;
Iberomaurusian - I was actually really torn between using this or the Guanche11 sample. The latter is identical to modern Moroccans, but would probably make a lot of sense given it's dated from the 8th century. However it contains Steppe and EEF ancestry, and as a consequence Guanche11 score would vary depending on the other references. I decided to use Iberomaurusian simply because it's a North African-specific source of ancestry and and not related to other groups at all.
As I said, I'm not dogmatic about this choice of references, it's just a way to start.
Below are the results for different regions of the peninsula. Keep in mind that some/most references were built with only 3 or 4 samples, so it's best not to take the results too literally. Besides there's personal variation to account for, and so few references doesn't do it enough justice.
Guess that some people are scoring Levant_BA_N by getting a lower amount of Collegno (or the other way around), so I'm not too happy about this choice in particular, it might not be very useful to begin with.
Hallstatt seems to be highest in the north and dropping as you move away - although I suspect some "Basque effect" might be in play, as they would score high Hallstatt aswell, but because of higher HG ancestry, not necessarily Steppe-related one.
On the other hand the Roman proxy is clearly higher in the south, and around coastal areas, the Baleares in particular seem to score a very high amount of it.
And finally, as expected, North African-specific ancestry is higher in the West, and gradually lowers as you move East.
A simplistic analysis indicates the Portuguese reference as perhaps the least autochthonous of the bunch, increasing as you move close to the north/northeast, although as you'll see below personal variation rules a simple scenario out of the window. It may be a trend, but hardly a rule.
Edit: Made a list with Colegno+Levant_BA_North for all the above regions
Portugal 24,8
Galicia 23,6
Extremadura 24,2
Castilla y León 17,0
Castilla La Mancha 20,6
Andalucia 24,8
Cantabria 14,6
Murcia 25,2
Valencia 18,2
Catalonia 17,0
Aragón 14,0
Baleares 28,4
Canarias 22,2
Same, but with the supposed pre-Roman references (Hallstatt+Iberia)
Portugal 70,8
Galicia 72,4
Extremadura 71,4
Castilla y León 78,8
Castilla La Mancha 77,6
Andalucia 72,6
Cantabria 83,6
Murcia 72,4
Valencia 79,8
Catalonia 80,6
Aragón 80,4
Baleares 70,6
Canarias 70,0
Below I'll post a few results for Portuguese forum members and relatives, just for fun. If you are Iberian and want to be added, just post your G25 unscaled coordinates.
For reference:
Ruderico - North-centre
MrsRuderico - East (next to Extremadura)
rxavierflima - Northwest
elisabetenjoaquim - West
Lusitano - Centre
Last edited by Ruderico; 02-08-2019 at 07:33 PM.
Reason: Added Castilla La Mancha, sorry Manchegos
YDNA E-Y31991>PF4428>Y134097>Y134104>Y168273>FT17866 (TMRCA ~1100AD) - Domingos Rodrigues, b. circa 1690 Hidden Content, Viana do Castelo, Portugal - Stonemason, miller.
mtDNA H20 - Monica Vieira, b. circa 1700 Hidden Content, Porto, Portugal
Can you refer me to a source (or give a quick summary) of regional population genetic differences within Portugal? Familiar with the situation across the border, but knowledge of inter-regional differences in Portugal might clear up some issues I’m having modeling crypto-Jews from Trás-is-Montes.
Can you refer me to a source (or give a quick summary) of regional population genetic differences within Portugal? Familiar with the situation across the border, but knowledge of inter-regional differences in Portugal might clear up some issues I’m having modeling crypto-Jews from Trás-is-Montes.
Unfortunetly we don't have such information, eyeballing though various results from Gedmatch calculators is very difficult to establish any form of regional variation because you find people with ancestry from roughly the same region with "very different" results. It may be that personal variation is more important than regional variation (although to be fair I haven't seem results explicitly from the south of the country).
This seems to be in line with PCA plots in which Portugal and Spain have a very large overlap, it appears Portugal despite its small size holds a fair degree of internal variation. For example, as you can see my result is virtually the same as the Castilla y León reference. My ancestry centroid is relatively close to that of rxavierflima's (we even have ancestors from the same parish), except more eastern. Similarly my missus place of origin is not too far from Lusitano's, yet our results are nothing alike.
Xavier and I have been trying to persuade more people into getting their G25 coordinates, so we could try to spot regional differences, but so far we've been unsuccesful
Last edited by Ruderico; 02-08-2019 at 04:53 PM.
YDNA E-Y31991>PF4428>Y134097>Y134104>Y168273>FT17866 (TMRCA ~1100AD) - Domingos Rodrigues, b. circa 1690 Hidden Content, Viana do Castelo, Portugal - Stonemason, miller.
mtDNA H20 - Monica Vieira, b. circa 1700 Hidden Content, Porto, Portugal
To put things in prespective I decided to make a few G25 scatter plots with different datasets, just to see how each group (and individuals) change their relative position as the dataset has additional references.
Each group will have its unique colour, which I'll list below. Individuals will be easier to spot because they'll have their name standing next to their symbol.
Basques - darkblue
French_South - darkblue
Pais Vasco - green
Portuguese - mediumblue
Andalucia - darkgreen
Aragon - orangered
Baleares - yellow
Canarias - lime
Cantabria - saddlebrown
Castilla La Mancha - gold
Castilla y León - purple
Catalonia - crimson
Extremadura - black
Galicia - skyblue
Murcia - olive
Valencia - orange
I'll start off with a scatterplot exclusively with Iberian populations, just for future reference.
Next we'll add North African populations, to check who gets pulled the most. Since the scatter plot will make the Iberian cluster smaller and thus harder to read, you can see a zoomed image here.
As expected from the previous model, Canarians and West Iberians get to plot closer to North African populations than the rest. One Portuguese sample seems to be pulled as much or more than Canarians, for some reason.
The following scatter plot includes (north)West Europeans and all Italian populations (including Sardinians).
The mediterranean cline is easily spotted, with South and West Iberians being more affected by it. Here's the zoomed scatter plot
And finally the standard scatter plot with all European populations, which you're probably already used to see.
Once again you can see the zoomed version.
As I said in the previous post, the Portuguese references seem to harbour quite a bit of variation between them, which is kind of unexpected for a small country. The similar thing might happen with the Spanish regions, but unfortunetly G25 doesn't have enough samples to confirm it
Last edited by Ruderico; 02-08-2019 at 07:39 PM.
YDNA E-Y31991>PF4428>Y134097>Y134104>Y168273>FT17866 (TMRCA ~1100AD) - Domingos Rodrigues, b. circa 1690 Hidden Content, Viana do Castelo, Portugal - Stonemason, miller.
mtDNA H20 - Monica Vieira, b. circa 1700 Hidden Content, Porto, Portugal
Há uma coisa que não percebo segundo o teu modelo: o que me faz tender para a Itália é ter mais Italy_Medieval_Collegno_o1 mas o que é faz elisabete tender para a França? Não é ter Hallstatt_Bylany mais alta por si só porque tu tens ainda mais. Pelo que percebo é ter também bastante Levant_BA_North em comparação com os outros. Mas isso dá a entender que o que separa os franceses dos portugueses é o facto dos primeiros terem mais sangue do médio oriente, o que nao me parece correcto. Isto é o tal problema de passar 25 eixos para 2?
É possível, mas também acho que o modelo inicial é demasiado simples para fazer essa distinção, ja que tanto a Elisabete como eu temos valores altos de Hallstatt, mas ela tende para França enquanto que eu para o Norte de Espanha. Já a minha namorada plota relativamente perto de mim e tem menos 20%..
Penso que o mais útil deste modelo é ver como as referências usadas como proxy variam ao longo da península, especialmente as do Norte de África e Romana, mas ainda assim se calhar o modelo precisava de ser aperfeiçoado. Com tempo e novas samples lá chegaremos
Por falar em Romanos, há isto. As Baleares foram durante muito tempo parte do império romano, talvez explique terem tanto de Collegno
Last edited by Ruderico; 02-09-2019 at 07:24 AM.
YDNA E-Y31991>PF4428>Y134097>Y134104>Y168273>FT17866 (TMRCA ~1100AD) - Domingos Rodrigues, b. circa 1690 Hidden Content, Viana do Castelo, Portugal - Stonemason, miller.
mtDNA H20 - Monica Vieira, b. circa 1700 Hidden Content, Porto, Portugal
Sorry to post in English but at least since a couple of week I decided to learn Spanish ( by myself). I found this interesting article and I want to share with you.
Há uma coisa que não percebo segundo o teu modelo: o que me faz tender para a Itália é ter mais Italy_Medieval_Collegno_o1 mas o que é faz elisabete tender para a França? Não é ter Hallstatt_Bylany mais alta por si só porque tu tens ainda mais. Pelo que percebo é ter também bastante Levant_BA_North em comparação com os outros. Mas isso dá a entender que o que separa os franceses dos portugueses é o facto dos primeiros terem mais sangue do médio oriente, o que nao me parece correcto. Isto é o tal problema de passar 25 eixos para 2?
I decided to pick up from here, because the previous model was indeed unable to distinguish between different sources of "northern" admixture.
Certain individuals or groups scored similar amounts of Hallstatt but plot in very different areas of the Iberian cluster with other European populations present - Elisabete plots on the border between Iberia and France, while I plot near Cantabria, closer to Basque individuals, just as an example. This is symptomatic of a higher degree of Northwestern/Central European ancestry in her, so I decided to get rid of the Iberia_Central_CA and replace it with a reference from a similar time period and region which already harbours steppe ancestry. So I went with the most obvious, Iberia_BA, which should have around half the amount of steppe ancestry as the Hallstatt reference, while having higher WHG and more affinity with Basques. I also got rid of the Levant_BA_North, as it was only occasionally appearing but always in small quantities. Just to test it, I preemptively ran Basque_Spanish, just to check it they indeed get a large chunk of Iberia_BA
[1] "distance%=1.2471"
Basque_Spanish
Iberia_BA,76.8
Hallstatt_Bylany,23.2
The model's fits improve a bit too, which is good. I'll remind everyone that the regional references were built few samples, so the results shouldn't be taken too literally.
And below, on a separate group, the same individual Portuguese samples as before. You can check the G25 PCA plot again here
Again, for reference:
Ruderico - North-centre
MrsRuderico - East (next to Extremadura)
rxavierflima - Northwest
elisabetenjoaquim - West
Lusitano - Centre
Looking at the regional frequencies again, Hallstatt peaks in Castilla y León - although this component can be picking up more recent CEU ancestry than simply Hallstatt groups that moved into Iberia since the late Bronze/early Iron Ages. This makes sense, given it's where most of the Meseta is located, and it has been historically important in the process of IE influence in Iberia, notably Cogotas I and II, earlier Beaker influence, although mostly on the Castilla area, closer to Aragón, and even later Celtiberian tribes, also on the easternmost areas.
Galicia scores relatively low, not totally unexpected as its process of Celtisation was late, and continued even by the time Romans were fighting Celtiberians further East. Although this could simply be the consequence of the very few samples avaliable. Also predictable that populations close to Basques scored high Iberia_BA, notably Cantabria.
Last edited by Ruderico; 02-13-2019 at 10:17 AM.
YDNA E-Y31991>PF4428>Y134097>Y134104>Y168273>FT17866 (TMRCA ~1100AD) - Domingos Rodrigues, b. circa 1690 Hidden Content, Viana do Castelo, Portugal - Stonemason, miller.
mtDNA H20 - Monica Vieira, b. circa 1700 Hidden Content, Porto, Portugal
Galicia scores relatively low, not totally unexpected as its process of Celtisation was late, and continued even by the time Romans were fighting Celtiberians further East. Although this could simply be the consequence of the very few samples avaliable. Also predictable that populations close to Basques scored high Iberia_BA, notably Cantabria.
Reply With Quote
Originally Posted by hartaisarlag
.png)
