Genetic history of Calabrian Greeks reveals ancient events and long term isolation in

@All
How do you interpret those p-values? They say "we considered a P-value threshold of 0.01 to assess the significance of tested models". We should expect that the p-values were lower than 0.01, shouldn't we? Yet, they're all higher than 0.01. I'm certainly missing something. Anyway, most of them are lower than 0.05. (1)(2)
North Italy and Sardinia get higher p-values; N. Italy's would be supposedly off(?): > 0.05. So I wonder if Tepecik-like ancestry was the one used instead of Barcin-like ancestry, given the focus on Calabria, which could perhaps explain the p-value obtained for N. Italy.
 
It's not very much more. The big difference between them, which changes the percentages, is the amount of steppe in Northern Italy, which is even a bit higher in Northeastern Italy, as Regio pointed out. You can see it in the ydna in Sardinia. They do have R1b-U152 in Sardinia, but a lot of it is in the north where a good part of the area speaks a Corsican dialect, not the parts of the island where Sardinian is spoken.

As I mentioned in another post, the Romans very quickly Romanized Northern Italy, and not just in draining swamps, building roads and settlements etc., or introducing Latin, or educating people, or adding new gods and imposing their political structure, but also genetically through veterans colonies. There was also the natural movement of people from Southern Italy.

I don't want to give the wrong impression here. There was already Iran Neo in central Italy in the Neolithic according to Antonio et al, a little more in the Copper Age, and it was there in the Republican Era Roman samples, and not just in the one sample with high Aegean like ancestry. Jovialis pointed it out by using material from Antonio et al on this page of that dedicated thread, as he recently reminded me.

See:
https://www.eupedia.com/forum/threads/39501-Moots-Ancient-Rome-Paper/page35?highlight=Antonio

The Empire would just have increased it.

I think all the confusion arises from the fact that Antonio et al, and a lot of its readers did not take into consideration that some of the "Imperial" and "Post Imperial" samples could have been travelers, traders etc.

To really approach understanding what went on we need North Italian, i.e. North of Rome samples from the Iron Age, and samples from Bronze Age to Iron Age samples from southern Italy. We also need Iron Age samples from, say, the Peloponnese and the Aegean islands.
 
In S7 of the Supplement, all of the Anatolia Neolithic farming samples are included so far as I can tell. I don't know if the following means they used them all.

"To test temporal patterns of genetic relationships, we finally merged the “modern extended” dataset with genomic data for 1059 ancient samples (Suppl. Table S7) extracted from the literature62,63,64,65,66,67,68,69,70,71,72 and genotyped on the 1240 K panel (V37.2.1240K, https://reich.hms.harvard.edu/), finally obtaining a common “modern-plus-ancient” dataset of 326,832 SNPs. For the genotype-based analyses involving also ancient samples we applied a LD-pruning procedure by excluding one SNP for each pair of loci showing r2 values higher than 0.4 within a 200‐SNPs window, sliding 25 loci at the time (PLINK option --indep-pairwise 200 25 0.4), for a total of 286,656 SNPs left after pruning."
 
Wow, very interesting paper ! Thanks for posting kingjohn.

I am very surprised, Northern Italy has more Iran Neo than Sardinia. I am not an expert on Italy but can someone explain the difference. I think for North Italy there is a little too much Iran Neo.
The HGDP Sardinians are from the interior, maybe the ones from the coast are a bit more Iran Neo and Steppe and less WHG.

Inviato dal mio POT-LX1T utilizzando Tapatalk
 
In S7 of the Supplement, all of the Anatolia Neolithic farming samples are included so far as I can tell. I don't know if the following means they used them all.

"To test temporal patterns of genetic relationships, we finally merged the “modern extended” dataset with genomic data for 1059 ancient samples (Suppl. Table S7) extracted from the literature62,63,64,65,66,67,68,69,70,71,72 and genotyped on the 1240 K panel (V37.2.1240K, https://reich.hms.harvard.edu/), finally obtaining a common “modern-plus-ancient” dataset of 326,832 SNPs. For the genotype-based analyses involving also ancient samples we applied a LD-pruning procedure by excluding one SNP for each pair of loci showing r2 values higher than 0.4 within a 200‐SNPs window, sliding 25 loci at the time (PLINK option --indep-pairwise 200 25 0.4), for a total of 286,656 SNPs left after pruning."
Thanks, Angela.

It seems they didn't use them all for the 4-way mixtures. Notice that S7 includes all kinds of samples: Ust_Ishim, Kostenki, MA1, SHG, Unetice, Bell Beaker, Levant BA and on and on. Apparently they used them all for merging modern and ancient, i.e., for selecting the dataset of SNPs used in their work. And for PCA*. It doesn't mean that all those ancient samples were used in the model.

The p-values for North Italy and Sardinia contrast with the Calabrian, and they seem to point to bad fits. Not sure, but one of the possible explanations is that they used the Anatolian source that makes more sense to Calabria (Tepecik?), and had to use them for North Italy and Sardinia as well, which resulted in those extremely high p-values.

*ED: Based on the next post (thanks, Duarte).
 
Last edited:
Supllementary figures:

https://static-content.springer.com.../MediaObjects/41598_2021_82591_MOESM1_ESM.pdf (good definition).

As example, S8 and S7 (screenshots Low definition):

bf48sEG.jpg

u5x86xl.jpg
 
Who brought Iran Neo in North Italy and Sardinia? Imperial Romans?

Inviato dal mio POT-LX1T utilizzando Tapatalk


I would not swear to the accuracy of Iran_N's percentages in this paper. One has to see what reference ancient samples she exactly used.

I've checked in the supp info. For example as reference samples to represent Steppe she did not use the usual Yamanya Samara only, but threw in as reference samples also individuals from Potapovka, Afanasievo etc...

Unfortunately, in the analyzes there are none of the southern European populations genetically close to the Italians, so we cannot have any comparison.

It is strange :unsure:
That acording to this paper
Sardinians have more whg component 12.2% than north italians 7.9%
Is there explanation to it ?

More WHG in Sardinians is not strange, it has always been so. What is strange is Iran_N being inflated everywhere.
 
I would not swear to the accuracy of Iran_N's percentages in this paper. One has to see what reference ancient samples she exactly used.





More WHG in Sardinians is not strange, it has always been so. What is strange is Iran_N being inflated everywhere.

Ok...
I didn't knew the averge sardinian share more allells with the whg reference than averge north italian do :cool-v:
 
Ok...
I didn't knew the averge sardinian share more allells with the whg reference than averge north italian do :cool-v:

The Sardinians, particularly the HGDP sample, has long been modelled as 85% EEF and 15% WHG. When no one in northern Italy, as I remember, has ever reached 15+% WHG. Now how Sardinians can reach 15% of Iran_N is quite strange.
 
Last edited:
The Sardinians, particularly the HGDP sample, has long been modelled as 85% EEF and 15% WHG. When no one in northern Italy has ever reached 15% WHG. Now how Sardinians can reach 15% of Iran_N is quite strange.

w3KrzYN.jpg


Fernandes et al 2020 seems to model Sardinians similarly to the paper, however. In fact, Sardinians seem to have more Iran_N in the Fernandes paper.:

ONagPuX.jpg
 
The Sardinians, particularly the HGDP sample, has long been modelled as 85% EEF and 15% WHG. When no one in northern Italy has ever reached 15% WHG. Now how Sardinians can reach 15% of Iran_N is quite strange.

Interesting:unsure:
How much whg % do spaniards( non basque) have ?
Also significant like sardinian ?
 
Interesting:unsure:
How much whg % do spaniards( non basque) have ?
Also significant like sardinian ?

I am not certain of the percentage, but Spaniards have a relatively large amount of WHG. Portuguese and many Spaniards were also modeled to have respectable amounts of Anatolian_BA, as well, in Raveane et al. 2018. So I would imagine, they would have respectable amounts of Iran_N.

X5FQhf9.jpg
 
w3KrzYN.jpg


Fernandes et al 2020 seems to model Sardinians similarly to the paper, however. In fact, Sardinians seem to have more Iran_N in the Fernandes paper.


In fact also the Fernandes 2020 models on modern samples are not very credible. Fernandes is the same who models Sicilians as 46.9% North African using Moroccan Late Neolithic (he literally says that modern Sicilians show a predominant component of North African ancestry). Really? A meaningless model, since Moroccan Late Neolithic are samples that show a strong gene flow from Iberia to Morocco, they cannot be representative samples of the modern populations of North Africa.


Not to mention yet another confusion made by geneticists. In the table at the top in the Sarno paper it says IRAN_N but in the caption it says CHG/IRAN_N. So, in 2021 the geneticists are still not having a clear idea of what is CHG and what is IRAN_N. Can we really take them seriously?


p15TK6k.png
 
Interesting:unsure:
How much whg % do spaniards( non basque) have ?
Also significant like sardinian ?

In the old models, percentages similar to those of Sardinians up to 20% and more (in the Basques). Now who knows, given the latest trend among geneticists, I imagine they will be at 20% Iran_N (I'm joking).
 
The mixture in G25.

(Scaled)
TargetDistanceIRN_Ganj_Dareh_NTUR_Tepecik_Ciftlik_NWHGYamnaya_RUS_Samara
Sardinian:HGDP006650.04041752085.214.80
Sardinian:HGDP006710.0469156082.817.20
Sardinian:HGDP006720.04612349086.413.60
Sardinian:HGDP006740.0660315083170
Sardinian:HGDP010630.04658002084.314.61.1
Sardinian:HGDP010660.06251437084.615.40
Sardinian:HGDP010670.04357858084.715.30
Sardinian:HGDP010730.05871846085.214.80
Sardinian:HGDP010750.05328043085.614.40
Sardinian:HGDP010780.04876746081.915.82.3
Sardinian:S_Sardinian-20.04848495082.216.31.5
Italian_Bergamo:HGDP011470.03288339058.69.232.2
Italian_Bergamo:HGDP011510.04753794062.712.924.4
Italian_Bergamo:HGDP011520.0323124061.68.130.3
Italian_Bergamo:HGDP011530.02962768062.9928.1
Italian_Bergamo:HGDP011550.02969563064.17.628.3
Italian_Calabria:ALP5820.018283439.670.91.318.2
Italian_Calabria:ALP5960.016647657.270.71.121
Italian_Calabria:BEL570.02364372470.5025.5
Average0.041686541.176.211.511.2

If we use Barcin:
TargetDistanceIRN_Ganj_Dareh_NTUR_Barcin_NWHGYamnaya_RUS_Samara
Sardinian:HGDP006650.024013091.782.412.13.8
Sardinian:HGDP006710.02931803080.414.15.5
Sardinian:HGDP006720.025331390.181.6117.3
Sardinian:HGDP006740.03675824079.714.75.6
Sardinian:HGDP010630.02482242081.810.87.4
Sardinian:HGDP010660.0378662083.913.62.5
Sardinian:HGDP010670.02806582.879.814.43
Sardinian:HGDP010730.038212150.583.413.42.7
Sardinian:HGDP010750.02786165083.611.35.1
Sardinian:HGDP010780.03350293079.312.28.5
Sardinian:S_Sardinian-20.03436949080.112.87.1
Italian_Bergamo:HGDP011470.01798027055.96.737.4
Italian_Bergamo:HGDP011510.02826248059.19.831.1
Italian_Bergamo:HGDP011520.02391725059.1535.9
Italian_Bergamo:HGDP011530.01511783060.16.333.6
Italian_Bergamo:HGDP011550.016929480605.334.7
Italian_Calabria:ALP5820.0190416615.964.11.518.5
Italian_Calabria:ALP5960.0209064913.864.1022.1
Italian_Calabria:BEL570.0269222212.962.1025
Average0.026799952.572.79.215.6

Both:
Target Distance IRN_Ganj_Dareh_N TUR_Barcin_N TUR_Tepecik_Ciftlik_N WHG Yamnaya_RUS_Samara
Sardinian:HGDP00665 0.02401309 1.7 82.4 0 12.1 3.8
Sardinian:HGDP00671 0.02931803 0 80.4 0 14.1 5.5
Sardinian:HGDP00672 0.02533139 0.1 81.6 0 11 7.3
Sardinian:HGDP00674 0.03675824 0 79.7 0 14.7 5.6
Sardinian:HGDP01063 0.02482193 0 81.8 0 10.9 7.3
Sardinian:HGDP01066 0.0378662 0 83.9 0 13.6 2.5
Sardinian:HGDP01067 0.02804244 2.6 77 3.3 14.5 2.6
Sardinian:HGDP01073 0.03821215 0.5 83.4 0 13.4 2.7
Sardinian:HGDP01075 0.02786165 0 83.6 0 11.3 5.1
Sardinian:HGDP01078 0.03350268 0 79.3 0 12.1 8.6
Sardinian:S_Sardinian-2 0.03436949 0 80.1 0 12.8 7.1
Italian_Bergamo:HGDP01147 0.01798027 0 55.9 0 6.7 37.4
Italian_Bergamo:HGDP01151 0.02826248 0 59.1 0 9.8 31.1
Italian_Bergamo:HGDP01152 0.02380149 0 53.8 5.9 5.4 34.9
Italian_Bergamo:HGDP01153 0.01506149 0 57.1 3.2 6.5 33.2
Italian_Bergamo:HGDP01155 0.01692948 0 60 0 5.3 34.7
Italian_Calabria:ALP582 0.01402107 12 29.1 40.1 0.8 18
Italian_Calabria:ALP596 0.01453407 10 27.4 40.9 0.6 21.1
Italian_Calabria:BEL57 0.02247109 7.9 23 43.9 0 25.2
Average 0.02595572 1.8 66.2 7.2 9.2 15.5
 
By all means let's take the findings of a program created by a skinhead instead of results from academics. Makes perfect sense to me.

I'm sure the amount of Levantine is huge in southern Italians according to that.

What else would you expect from a virulent anti-Semite who posted Southern Italians should be kicked out of Europe.
 
Does anyone have to hand a file showing the breakdown based on ancient samples for Northern Italy from Fernandez et al, Raveane et al, and Antonio et al (the Moots paper), for comparison?
 
Even using program made by others, Sardinians never reach 15% of Iran_N. This of inflating Iran_N is only a recent trend in academic studies. As we have seen so many times in the past, geneticists are anything but infallible, and they are often wrong especially with these very speculative models.
 
By all means let's take the findings of a program created by a skinhead instead of results from academics. Makes perfect sense to me.

I'm sure the amount of Levantine is huge in southern Italians according to that.

What else would you expect from a virulent anti-Semite who posted Southern Italians should be kicked out of Europe.
I've posted it just for comparison. I also wanted to compare Barcin vs. Tepecik, but I have no idea on how to use other tools such qpAdm. :)

These would be the amounts of Levantine suggested for Southern Italians by the tool (naturally they may vary depending on the references chosen):
Target Distance IRN_Ganj_Dareh_N Levant_PPNB MAR_EN TUR_Barcin_N TUR_Tepecik_Ciftlik_N WHG Yamnaya_RUS_Samara
Sardinian 0.0268524 0 0 0 81.8 0 12.2 6
Italian_Abruzzo 0.00768384 7.3 6.7 0 42.9 13.5 0.6 29
Italian_Molise 0.00761903 6.5 4.3 0 44.8 13.5 0.5 30.4
Italian_Campania 0.00787971 9.4 8.4 0 35.8 21.1 0 25.3
Italian_Apulia 0.00818471 7.2 3.2 0.2 33.5 29.2 0.2 26.5
Italian_Basilicata 0.0073524 7.9 4.8 0.1 36.3 24.2 0 26.7
Italian_Calabria 0.00959004 9.8 2 1.5 33.8 30.2 0 22.7
Sicilian_East 0.00960535 7.2 7 1.3 27.6 31.1 1.8 24
Sicilian_West 0.01100361 8.4 6.5 2.6 35.6 19.5 4.4 23
Average 0.01064123 7.1 4.8 0.6 41.3 20.3 2.2 23.7
 
I've posted it just for comparison. I also wanted to compare Barcin vs. Tepecik, but I have no idea on how to use other tools such qpAdm. :)

These would be the amounts of Levantine suggested for Southern Italians by the tool (naturally they may vary depending on the references chosen):
TargetDistanceIRN_Ganj_Dareh_NLevant_PPNBMAR_ENTUR_Barcin_NTUR_Tepecik_Ciftlik_NWHGYamnaya_RUS_Samara
Italian_Abruzzo0.007683757.36.8042.913.40.629
Italian_Molise0.007619056.54.3044.713.60.530.4
Italian_Campania0.007875839.48.6035.920.7025.4
Italian_Apulia0.00818217.22.90.333.629.30.226.5
Italian_Basilicata0.00734971850.136.723.5026.7
Italian_Calabria0.009590979.82.31.433.730.1022.7
Sicilian_East0.009605357.271.327.631.11.824
Sicilian_West0.011002988.36.52.635.2204.423
Average0.0086137285.40.736.322.70.926

No offense, but G25, and all of the various armature tools used by enthusiasts, really should be taken more seriously than professionals adept in using more sophisticated tools like qpAdm.
 

This thread has been viewed 44775 times.

Back
Top