PDA

View Full Version : Modeling of Iberian populations on Global25. Could these be improved?



Sikeliot
11-17-2019, 10:42 PM
I also get a Levantine component not captured by North African for some of these populations.

Target: Spanish_Canarias
Distance: 1.4874% / 0.01487431
Aggregated
43.8 Bell_Beaker_Iberia
27.0 Moroccan_North
15.2 DEU_MA_ACD
14.0 ITA_Ardea_Latini_IA_o

Target: Spanish_Castilla_Y_Leon
Distance: 1.2980% / 0.01297956
Aggregated
41.2 Bell_Beaker_Iberia
26.8 DEU_MA_ACD
14.0 Moroccan_North
12.0 ITA_Ardea_Latini_IA
5.2 ITA_Ardea_Latini_IA_o
0.8 Levant_Canaanite_MBA

Target: Spanish_Extremadura
Distance: 1.0861% / 0.01086072
Aggregated
34.8 Bell_Beaker_Iberia
32.0 DEU_MA_ACD
10.8 Moroccan_North
9.8 Levant_Canaanite_MBA
8.2 ITA_Ardea_Latini_IA
4.4 ITA_Ardea_Latini_IA_o

Target: Spanish_Galicia
Distance: 0.9243% / 0.00924275
Aggregated
37.8 Bell_Beaker_Iberia
35.4 DEU_MA_ACD
10.8 Moroccan_North
7.6 Levant_Canaanite_MBA
7.0 ITA_Ardea_Latini_IA
1.4 ITA_Ardea_Latini_IA_o

Target: Spanish_Asturias
Distance: 2.3181% / 0.02318115
Aggregated
54.6 Bell_Beaker_Iberia
25.2 DEU_MA_ACD
10.2 Moroccan_North
10.0 ITA_Ardea_Latini_IA_o

Target: Portuguese
Distance: 0.9591% / 0.00959077
Aggregated
43.4 Bell_Beaker_Iberia
30.4 DEU_MA_ACD
10.2 Moroccan_North
8.0 ITA_Ardea_Latini_IA_o
6.0 Levant_Canaanite_MBA
2.0 ITA_Ardea_Latini_IA

Target: Spanish_Valencia
Distance: 0.7645% / 0.00764522
Aggregated
48.2 Bell_Beaker_Iberia
30.8 DEU_MA_ACD
14.4 ITA_Ardea_Latini_IA_o
3.6 Moroccan_North
3.0 Levant_Canaanite_MBA

Target: Spanish_Terres_de_l'Ebre
Distance: 0.7183% / 0.00718252
Aggregated
46.6 Bell_Beaker_Iberia
23.2 DEU_MA_ACD
16.2 ITA_Ardea_Latini_IA_o
9.4 ITA_Ardea_Latini_IA
4.0 Levant_Canaanite_MBA
0.6 Moroccan_North

Target: Spanish_Soria
Distance: 1.5913% / 0.01591298
Aggregated
48.0 Bell_Beaker_Iberia
35.4 DEU_MA_ACD
13.0 ITA_Ardea_Latini_IA
3.0 Levant_Canaanite_MBA
0.6 Moroccan_North

Target: Spanish_Pirineu
Distance: 0.7806% / 0.00780627
Aggregated
54.6 Bell_Beaker_Iberia
26.6 DEU_MA_ACD
6.8 Levant_Canaanite_MBA
6.6 ITA_Ardea_Latini_IA
3.2 Moroccan_North
2.2 ITA_Ardea_Latini_IA_o

Target: Spanish_Peri-Barcelona
Distance: 0.7490% / 0.00748968
Aggregated
39.8 Bell_Beaker_Iberia
35.8 DEU_MA_ACD
11.6 ITA_Ardea_Latini_IA
7.8 Levant_Canaanite_MBA
2.6 ITA_Ardea_Latini_IA_o
2.4 Moroccan_North

Target: Spanish_Penedes
Distance: 0.9750% / 0.00975008
Aggregated
52.0 Bell_Beaker_Iberia
30.6 DEU_MA_ACD
8.8 Levant_Canaanite_MBA
5.4 ITA_Ardea_Latini_IA_o
2.6 ITA_Ardea_Latini_IA
0.6 Moroccan_North

Target: Spanish_Pais_Vasco
Distance: 1.9205% / 0.01920518
Aggregated
70.0 Bell_Beaker_Iberia
21.8 DEU_MA_ACD
8.0 ITA_Ardea_Latini_IA_o
0.2 Moroccan_North

Target: Spanish_Navarra
Distance: 1.8570% / 0.01857012
Aggregated
48.6 Bell_Beaker_Iberia
33.0 DEU_MA_ACD
14.8 Levant_Canaanite_MBA
2.4 ITA_Ardea_Latini_IA_o
1.2 Moroccan_North

Target: Spanish_Murcia
Distance: 0.8392% / 0.00839246
Aggregated
44.2 Bell_Beaker_Iberia
39.6 DEU_MA_ACD
8.4 Levant_Canaanite_MBA
7.8 Moroccan_North

Target: Spanish_Menorca
Distance: 0.7837% / 0.00783669
Aggregated
29.4 DEU_MA_ACD
29.0 Bell_Beaker_Iberia
23.2 ITA_Ardea_Latini_IA
7.6 ITA_Ardea_Latini_IA_o
7.2 Levant_Canaanite_MBA
3.6 Moroccan_North

Target: Spanish_Mallorca
Distance: 0.7611% / 0.00761099
Aggregated
38.4 DEU_MA_ACD
25.4 Bell_Beaker_Iberia
20.2 ITA_Ardea_Latini_IA
9.0 Levant_Canaanite_MBA
3.8 ITA_Ardea_Latini_IA_o
3.2 Moroccan_North

Target: Spanish_Lleida
Distance: 0.8334% / 0.00833408
Aggregated
49.8 Bell_Beaker_Iberia
34.2 DEU_MA_ACD
11.0 ITA_Ardea_Latini_IA_o
4.2 Levant_Canaanite_MBA
0.8 Moroccan_North

Target: Spanish_La_Rioja
Distance: 1.1698% / 0.01169846
Aggregated
55.0 Bell_Beaker_Iberia
31.6 DEU_MA_ACD
13.4 Levant_Canaanite_MBA

Target: Spanish_Girona
Distance: 0.8613% / 0.00861253
Aggregated
36.4 Bell_Beaker_Iberia
36.2 DEU_MA_ACD
16.8 ITA_Ardea_Latini_IA
8.0 Levant_Canaanite_MBA
1.8 Moroccan_North
0.8 ITA_Ardea_Latini_IA_o


Target: Spanish_Eivissa
Distance: 0.9903% / 0.00990277
Aggregated
32.6 DEU_MA_ACD
30.6 Bell_Beaker_Iberia
13.6 ITA_Ardea_Latini_IA
12.2 ITA_Ardea_Latini_IA_o
8.0 Levant_Canaanite_MBA
3.0 Moroccan_North

Target: Spanish_Catalunya_Central
Distance: 0.8011% / 0.00801061
Aggregated
40.6 Bell_Beaker_Iberia
34.6 DEU_MA_ACD
12.6 ITA_Ardea_Latini_IA
6.8 Levant_Canaanite_MBA
5.4 ITA_Ardea_Latini_IA_o

Target: Spanish_Cataluna
Distance: 0.8480% / 0.00848016
Aggregated
41.4 DEU_MA_ACD
41.2 Bell_Beaker_Iberia
12.8 ITA_Ardea_Latini_IA_o
4.6 Moroccan_North


Target: Spanish_Castilla_La_Mancha
Distance: 0.7785% / 0.00778516
Aggregated
52.8 Bell_Beaker_Iberia
22.4 DEU_MA_ACD
14.0 ITA_Ardea_Latini_IA_o
5.6 Levant_Canaanite_MBA
4.4 Moroccan_North
0.8 ITA_Ardea_Latini_IA

Target: Spanish_Castello
Distance: 0.7802% / 0.00780165
Aggregated
52.0 Bell_Beaker_Iberia
26.2 DEU_MA_ACD
11.4 Levant_Canaanite_MBA
10.2 ITA_Ardea_Latini_IA
0.2 Moroccan_North

Target: Spanish_Cantabria
Distance: 1.3941% / 0.01394085
Aggregated
49.4 Bell_Beaker_Iberia
20.8 DEU_MA_ACD
15.8 ITA_Ardea_Latini_IA
5.8 ITA_Ardea_Latini_IA_o
4.8 Levant_Canaanite_MBA
3.4 Moroccan_North


Target: Spanish_Camp_de_Tarragona
Distance: 0.6801% / 0.00680067
Aggregated
43.8 Bell_Beaker_Iberia
30.0 DEU_MA_ACD
9.4 ITA_Ardea_Latini_IA
9.0 ITA_Ardea_Latini_IA_o
5.4 Levant_Canaanite_MBA
2.4 Moroccan_North

Target: Spanish_Barcelones
Distance: 0.9644% / 0.00964399
Aggregated
38.0 DEU_MA_ACD
37.8 Bell_Beaker_Iberia
10.8 ITA_Ardea_Latini_IA
9.8 ITA_Ardea_Latini_IA_o
3.6 Moroccan_North

Target: Spanish_Baleares
Distance: 1.0120% / 0.01012019
Aggregated
32.4 ITA_Ardea_Latini_IA
29.8 DEU_MA_ACD
23.4 Bell_Beaker_Iberia
6.8 Levant_Canaanite_MBA
4.8 Moroccan_North
2.8 ITA_Ardea_Latini_IA_o


Target: Spanish_Aragon
Distance: 1.1325% / 0.01132540
Aggregated
55.6 Bell_Beaker_Iberia
27.2 DEU_MA_ACD
6.0 ITA_Ardea_Latini_IA_o
5.0 Levant_Canaanite_MBA
4.4 Moroccan_North
1.8 ITA_Ardea_Latini_IA

Target: Spanish_Andalucia
Distance: 1.1079% / 0.01107932
Aggregated
49.2 Bell_Beaker_Iberia
18.6 DEU_MA_ACD
13.8 ITA_Ardea_Latini_IA_o
7.2 Moroccan_North
6.0 ITA_Ardea_Latini_IA
5.2 Levant_Canaanite_MBA

Target: Spanish_Alacant
Distance: 1.2881% / 0.01288128
Aggregated
50.2 Bell_Beaker_Iberia
27.2 DEU_MA_ACD
10.0 ITA_Ardea_Latini_IA_o
7.6 Levant_Canaanite_MBA
5.0 Moroccan_North

LTG
11-17-2019, 11:35 PM
Your models are lacking the appropriate Neolithic Iberian component. Here they are with the web runner; results may be more accurate when using the manual nMonte with default penalty but I do not have the necessary datasheets at the moment.

Spanish_Aragon
0.8106% / 0.00810599
54.4 Bell_Beaker_NLD
40.4 Iberia_N
5.2 Tunisian

Spanish_Baleares
1.3834% / 0.01383392
53.6 Bell_Beaker_NLD
38.8 Iberia_N
7.6 Tunisian

Spanish_Barcelones
1.0909% / 0.01090920
58.4 Bell_Beaker_NLD
37.0 Iberia_N
4.6 Tunisian

Spanish_Cataluna
0.8533% / 0.00853347
57.2 Bell_Beaker_NLD
35.2 Iberia_N
7.6 Tunisian

Spanish_Valencia
0.9861% / 0.00986103
55.0 Bell_Beaker_NLD
36.0 Iberia_N
7.0 Tunisian

Spanish_Murcia
Distance: 1.0732% / 0.01073178
49.0 Bell_Beaker_NLD
38.6 Iberia_N
12.4 Tunisian

Spanish_Canarias
1.1718% / 0.01171820
42.6 Bell_Beaker_NLD
32.0 Iberia_N
25.4 Tunisian

Seabass
11-18-2019, 12:30 AM
I think you should use gaunche11 for anything berber. I also think you should strongly consider using the samples from olalde 2019 rather then the bell beaker iberian. In G25 there are now samples that are likely anything from moriscos to celtiberians. The iberia empuries 2 (greeks from sw turkey) should be included too. Amazingly iberian today show no real impact from Greek settlements since AD as opposed to Italians. I'm still surprised some or someone on this forum suggest some iberian could descend up to 25 percent from Romans.

Also DEU mac ACD are the samples that had their skulls purposely deformed. They are said to on average be of medieval balkan origin. Do you think they are relevant here? DEU MA on the other hand could be, but in G25 there are apparently a few celtiberians samples available.

Sikeliot
11-18-2019, 12:39 AM
I think you should use gaunche11 for anything berber. I also think you should strongly consider using the samples from olalde 2019 rather then the bell beaker iberian. In G25 there are now samples that are likely anything from moriscos to celtiberians. The iberia empuries 2 (greeks from sw turkey) should be included too. Amazingly iberian today show no real impact from Greek settlements since AD as opposed to Italians. I'm still surprised some or someone on this forum suggest some iberian could descend up to 25 percent from Romans.

Also DEU mac ACD are the samples that had their skulls purposely deformed. They are said to on average be of medieval balkan origin. Do you think they are relevant here? DEU MA on the other hand could be, but in G25 there are apparently a few celtiberians samples available.

I didn’t use Guanches because they wouldn’t have absorbed some of the admixture that modern Berbers have that may have made it to Iberia by the Moors. Guanches may represent “pure” Berbers.

Ajeje Brazorf
11-18-2019, 07:56 AM
I didn’t use Guanches because they wouldn’t have absorbed some of the admixture that modern Berbers have that may have made it to Iberia by the Moors. Guanches may represent “pure” Berbers.

For a purely Berber component, you should use "Iberia_Central_CA_Afr" which is a fully North African genome that lived during the Bronze Age.

Iberia_Central_CA_Afr
Distance: 3.6471%
43.8 Anatolia_Barcin_N
29.0 MAR_Iberomaurusian
14.6 Levant_PPNB
8.8 Dinka
2.4 WHG
1.4 Yoruba

sweuro
11-18-2019, 08:03 AM
Your models are lacking the appropriate Neolithic Iberian component. Here they are with the web runner; results may be more accurate when using the manual nMonte with default penalty but I do not have the necessary datasheets at the moment.

Spanish_Aragon
0.8106% / 0.00810599
54.4 Bell_Beaker_NLD
40.4 Iberia_N
5.2 Tunisian

Spanish_Baleares
1.3834% / 0.01383392
53.6 Bell_Beaker_NLD
38.8 Iberia_N
7.6 Tunisian

Spanish_Barcelones
1.0909% / 0.01090920
58.4 Bell_Beaker_NLD
37.0 Iberia_N
4.6 Tunisian

Spanish_Cataluna
0.8533% / 0.00853347
57.2 Bell_Beaker_NLD
35.2 Iberia_N
7.6 Tunisian

Spanish_Valencia
0.9861% / 0.00986103
55.0 Bell_Beaker_NLD
36.0 Iberia_N
7.0 Tunisian

Spanish_Murcia
Distance: 1.0732% / 0.01073178
49.0 Bell_Beaker_NLD
38.6 Iberia_N
12.4 Tunisian

Spanish_Canarias
1.1718% / 0.01171820
42.6 Bell_Beaker_NLD
32.0 Iberia_N
25.4 Tunisian
if you dont add a east-med or near-east source, the Tunisian skyrockets..

Ruderico
11-18-2019, 10:56 AM
What's the point of using early Latins when we have actual Iron Age Iberians? North Moroccans are also not advisable, we have earlier samples (Guanches) eventhough they are trending towards Europe.
There's a whole thread about this in the Spanish sections, I suggest you go there.

Sikeliot
11-18-2019, 12:37 PM
What's the point of using early Latins when we have actual Iron Age Iberians? North Moroccans are also not advisable, we have earlier samples (Guanches) eventhough they are trending towards Europe.
There's a whole thread about this in the Spanish sections, I suggest you go there.

Guanches probably shift slightly toward Europe because they have slightly less SSA and Arabian than do continental Berbers... although they do not differ much at all from modern day northern Moroccans.

Really they might be more purely Berber than the continentals.

Ruderico
11-18-2019, 12:41 PM
Guanches probably shift slightly toward Europe because they have slightly less SSA and Arabian than do continental Berbers... although they do not differ much at all from modern day northern Moroccans.

Really they might be more purely Berber than the continentals.

No SSA references on the dataset, but:

https://i.postimg.cc/9cT10fpJ/berber-guanche.png

LTG
11-18-2019, 01:05 PM
if you dont add a east-med or near-east source, the Tunisian skyrockets..

There isn't any significant ancestry from those regions in Spain and Portugal. It's nothing more than low level admixture from Northern Africa as a result of historical movements in both directions leading to shared ancestry. If they had ancestry from the Eastern Mediterranean, which is Greece, Crete, Cyprus and coastal Anatolia, then they would harbor at least some additional components that we see in that region; CHG, for example, but they don't. Greek colonization may have a genetic legacy in some areas of Iberia but it isn't really needed to model these populations.

Portuguese
2.1056% / 0.02105609
51.6 Anatolia_Barcin_N
32.0 Yamnaya_RUS_Samara
10.4 WHG
6.0 MAR_Iberomaurusian

Spanish_Aragon
2.8188% / 0.02818790
53.0 Anatolia_Barcin_N
31.2 Yamnaya_RUS_Samara
12.8 WHG
3.0 MAR_Iberomaurusian

Spanish_Valencia
2.2847% / 0.02284703
52.8 Anatolia_Barcin_N
31.0 Yamnaya_RUS_Samara
13.0 WHG
3.2 MAR_Iberomaurusian

Spanish_Canarias
2.3153% / 0.02315297
51.2 Anatolia_Barcin_N
27.4 Yamnaya_RUS_Samara
10.2 MAR_Iberomaurusian
10.0 WHG
0.6 ETH_4500BP
0.6 Yoruba

Sikeliot
11-18-2019, 01:12 PM
No SSA references on the dataset, but:

https://i.postimg.cc/9cT10fpJ/berber-guanche.png

Guanches look more European-shifted than Moroccans, but are definitely not the most European-shifted Berbers. Guanches might be what Moroccans were like before extra SSA.

Ruderico
11-18-2019, 01:13 PM
There isn't any significant ancestry from those regions in Spain and Portugal. It's nothing more than low level admixture from Northern Africa as a result of historical movements in both directions leading to shared ancestry. If they had ancestry from the Eastern Mediterranean, which is Greece, Crete, Cyprus and coastal Anatolia, then they would harbor at least some additional components that we see in that region; CHG, for example, but they don't. Greek colonization may have a genetic legacy in some areas of Iberia but it isn't really needed to model these populations.


It may be an issue with using older components, or just G25, because you can clearly see the Roman_Imperial influence in modern Iberia if you do a model with IronAge/RomanEra samples, the Olalde paper clearly stated it too. This sort of 'East Med' influence wasn't a thing before the Romans arrived

Ruderico
11-18-2019, 01:15 PM
Guanches look more European-shifted than Moroccans, but are definitely not the most European-shifted Berbers. Guanches might be what Moroccans were like before extra SSA.

Those Berbers to the left are the ones with the least amount of SSA and Levantine admixture, according to the models ran in another thread, and likely the best candidates for the pre-Islamic population of Morocco.


Edit: Sorry double post

LTG
11-18-2019, 01:52 PM
It may be an issue with using older components, or just G25, because you can clearly see the Roman_Imperial influence in modern Iberia if you do a model with IronAge/RomanEra samples, the Olalde paper clearly stated it too. This sort of 'East Med' influence wasn't a thing before the Romans arrived

I'll give it a go later on the manual version; the online tool without default penalty creates very sharp models without minor admixtures. I wish whoever made it kept both functions rather than omitting the default one.

Bealfire
11-18-2019, 01:58 PM
Guanches probably shift slightly toward Europe because they have slightly less SSA and Arabian than do continental Berbers... although they do not differ much at all from modern day northern Moroccans.

Really they might be more purely Berber than the continentals.

Indeed that's true.

Don't mind Ruderico he has an agenda and thinks there is alot of Islamic Arab admixture in Morocco lol.

Whatever, G25 modeling shows Guanches don't differ anymore from North Moroccans (including Middle Atlas) at all.

Also most Moors in Iberia were North Moroccans identical to moderns.

Ruderico
11-18-2019, 02:00 PM
I'll give it a go later on the manual version; the online tool without default penalty creates very sharp models without minor admixtures. I wish whoever made it kept both functions rather than omitting the default one.

I never bother with very ancient models myself, I'm not sure G25 is an adequate tool for those to be honest. But when sticking to more recent ones the outputs appear at least plausible and relatively in line with published papers. However I still take them with a handful of salt, particularly when some of the references in the input are relatively close/similar (besides Germanic ancestry).
Here's a sample

[1] "distance%=0.7743"

Spanish_Murcia

Iberia_IA,67.6
ITA_Rome_Imperial,19.8
Germanic_IA,6.6
North_African,6



Indeed that's true.

Don't mind Ruderico he has an agenda and thinks there is alot of Islamic Arab admixture in Morocco lol.

Whatever, G25 modeling shows Guanches don't differ anymore from North Moroccans (including Middle Atlas) at all.

Also most Moors in Iberia were North Moroccans identical to moderns.

I don't really care, actually, much less have an agenda, I just look at the data. Truth is we don't have ancient continental samples, so not much can be said.

https://anthrogenica.com/showthread.php?18761-Who-are-North-Africans-descended-from&p=614801#post614801

Bealfire
11-18-2019, 03:43 PM
I never bother with very ancient models myself, I'm not sure G25 is an adequate tool for those to be honest. But when sticking to more recent ones the outputs appear at least plausible and relatively in line with published papers. However I still take them with a handful of salt, particularly when some of the references in the input are relatively close/similar (besides Germanic ancestry).
Here's a sample

[1] "distance%=0.7743"

Spanish_Murcia

Iberia_IA,67.6
ITA_Rome_Imperial,19.8
Germanic_IA,6.6
North_African,6




I don't really care, actually, much less have an agenda, I just look at the data. Truth is we don't have ancient continental samples, so not much can be said.

https://anthrogenica.com/showthread.php?18761-Who-are-North-Africans-descended-from&p=614801#post614801

Models using only Levant N are obviously bad.

Granted, models using BA Sidon/Lebanon instead are even more bad and don't make sense , so there is always worse next door.

In the absence of Mesolithic/Neolithic Egypt (and Libya) samples the best thing is to use neutral components like Natufian Levant, CHG, Iran_N and interpret consequently as one please.

Iran-N is rich in Semitic and Arab populations. Iran_N is too small in North Moroccans (averaging 3-4%) to represent any kind of obvious Post-Islamic admixture.

In good G25 models, North Moroccans score little Iran_N, and they do score significant Natufian-like ancestry just like the Guanche sample. So the two are clearly not the same and don't represent the same stream of admixture.

Ruderico
11-18-2019, 03:51 PM
My point is that there's a cline going on, and those Berber individuals on the end of the cline are close, but not really the same as, other populations such as North Morocco or whatnot. We'll see what samples from BA/IA/Antiquity Morocco look like when those are published. For now we can only speculate, and mine was that they could be analogous to what modern Basques are to Iberians (ie. similar, but closer to an older less admixed population, in our case IA_Iberia) since they too are at the edge of our cline. Mistaking this for "having an agenda" would be funny if it were not offensive. It's just speculation.

Sikeliot
11-18-2019, 06:56 PM
Indeed that's true.

Don't mind Ruderico he has an agenda and thinks there is alot of Islamic Arab admixture in Morocco lol.

Whatever, G25 modeling shows Guanches don't differ anymore from North Moroccans (including Middle Atlas) at all.

Also most Moors in Iberia were North Moroccans identical to moderns.

Guanches seem not necessarily close to isolated berbers but to typical Moroccans from places like Casablanca or Tangier. This might mean northern Morocco hasn’t changed much and that all Moroccans were not once like Mozabites.

sweuro
11-18-2019, 08:18 PM
There isn't any significant ancestry from those regions in Spain and Portugal. It's nothing more than low level admixture from Northern Africa as a result of historical movements in both directions leading to shared ancestry. If they had ancestry from the Eastern Mediterranean, which is Greece, Crete, Cyprus and coastal Anatolia, then they would harbor at least some additional components that we see in that region; CHG, for example, but they don't. Greek colonization may have a genetic legacy in some areas of Iberia but it isn't really needed to model these populations.

The east-med input in Iberia is not directly from the East-Med source (besides some minor phoenician or carthagian) it came indirectly from the Romans , which carry it in high levels