Ancient DNA, admx. history and endogamy in the prehistoricAegean Skourtaniotietal2022

[eupator]

Maybe you don't want to produce the model with Greece_N or Minoan because it will prove my point. That's not fair.

Yes, it's not fair, all opinions should be examined, using that poster's model.

Outgroup list:

right = c('Russia_Ust_Ishim.DG', 'Russia_Kostenki14.SG', 'Russia_MA1_HG.SG', 'Belgium_UP_GoyetQ116_1', 'Spain_ElMiron', 'Czech_Vestonice', 'Turkey_Epipaleolithic', 'Ethiopia_4500BP.SG','Iberomaurusian', 'EHG')

Sicilian results:

target   left                      weight     se     z
  <chr>    <chr>                      <dbl>  <dbl> <dbl>
1 Sicilian Greece_Minoan             0.566  0.0320 17.7 
2 Sicilian Russia_Samara_EBA_Yamnaya 0.213  0.0285  7.48
3 Sicilian Iran_N                    0.127  0.0387  3.27
4 Sicilian Natufian                  0.0945 0.0284  3.32
> results$popdrop
# A tibble: 15 × 15
   pat      wt   dof  chisq         p f4rank Greece_Minoan
   <chr> <dbl> <dbl>  <dbl>     <dbl>  <dbl>         <dbl>
 1 0000      0     6   10.4 1.10e-  1      3         0.566

56.6% Minoan + 21.3% Yamnaya + 12.7% Iran_N + 9.45% Natufian, all std. errors below 5%, p-value=0.110.


Italian_South:

target        left                      weight     se     z
  <chr>         <chr>                      <dbl>  <dbl> <dbl>
1 Italian_South Greece_Minoan             0.603  0.0425 14.2 
2 Italian_South Russia_Samara_EBA_Yamnaya 0.199  0.0353  5.64
3 Italian_South Iran_N                    0.140  0.0468  2.99
4 Italian_South Natufian                  0.0578 0.0365  1.59
> results$popdrop
# A tibble: 15 × 15
   pat      wt   dof   chisq         p f4rank Greece_Minoan
   <chr> <dbl> <dbl>   <dbl>     <dbl>  <dbl>         <dbl>
 1 0000      0     6    9.95 1.27e-  1      3         0.603

60.3% Minoan + 19.9% Yamnaya + 14% Iran_N + 5.78% Natufian, all std. errors below 5%, p-value 0.127.


PS. WHG on the left wrecks the model, so it is omitted.

 

[Hawk]

nice maybe the source of E ( e1b1b1) among them

Yes, definitely. Natufians as Dzudzuana can be even spotted by cranial studies alone. Populations with higher ANA were more robust and taller than Natufians and Dzudzuana-like. The gracilization in North Africa happened due to Dzudzuana-like admixture.

From all these chronologies, one i wonder is, during Bronze Age how did Y-DNA J1 got to be Afro-Asiatic speaking. Because odds are extremely low to zero for them to have been the originators of the language, they can be labelled up to Proto-Semitic but not further than that.

For me personally, it looks like initially they were under influence from E-M35 Natufians/PPNB, and somewhere after that, after accepting the tongue they took the upper hand in Levant during EBA and further south in Saudi Arabia (but i doubt E-M35 was ever present in Saudi Arabia).

 

[Francesco]

There is clear 100% confirmed excess Levant_N admix in South Italy. G25 models are consistent with qpAdm models. But you can force Levant N out and just use Iran N as a proxy on qpAdm, that is a forced model. Once you add Levant N it will pick Levant N. Also Iran N was not even close to those numbers in the Bronze Age South Italy.

Also you can use academic tools such as qpAdm to make stupid models such as this below. A lot of papers such the one modeling Italians with just Iran N instead of Levant N made mistakes such as this obviously unrealistic model below

Using that Sicilian bronze age sample, wich is an individual from the bell beaker horizon from the early bronze age, as a proxy for early iron age southern Italy is pretty approximate, since we will need southern italic tribes and Greek colonists to have a good idea of the genetic landscape of the region in the Iron Age. On the other hand, you seem to have chosen one of the least levant admixed mycenean samples, so, again, that could be a problem if we want to have a clear picture of the early iron age.
Anyway, an excess of anatolian ancestry after the imperial age is likely, not only in the south, but in the rest of Italy and in all the wester part of the Empire, we are just saying that the Greek colonization likely had a much bigger demographic impact in southern Italy, as many studies have shown.
Lastly, that G25 model for Calabria and Puglia seems to conflate a bit too much both the Yamnaya and the levant Neolithic component: you would need not only a pretty big arrival from the east, but also a substantial impulse from the north, both of which are not in line with the historical record (the invading barbarians were just a few thousand men), nor are suggested by the aforementioned studies. Maybe the samples used are not ideal.

 

[Leopoldo Leone]

Like i've said qpAdm is not perfect. In this case G25 models might be better than qpAdm models due to the limitation of too many proxies. Also Dodecad is an outdated calculator. It makes no sense to use just Minoans as a proxy for Italians when they clearly much more excess Iran N and Levant N stuff relatively to Bronze-Iron age Sicilians-Italians.

Your mistake is to believe that G25 samples are representative, but it is clear there is something fishy about them, and there is no real historical process that explains how all that Levantine-Anatolian ancestry ended up in south Italy (muh slaves and hellenistic "Greeks" do not make sense either), conveniently offset by another unexplained northern gene flow. Furthermore your bunch must stop pretending as if professional genetists were clueless and did not care to use relevant outgroups in their tests, and in the literature you find papers that do use middle easterners as a reference: https://www.biorxiv.org/content/10.1101/2022.02.26.482072v1

However, when the affinity of Italian groups with African and Middle Eastern populations was tested, Southern Italians resulted not significantly closer to any of the two

Furthermore there is no Levant_N excess in south Italians compared to ancient SE european samples, it is just the fantasies of delusional hobbyists who keep finding it DESPITE the fact that professional genetists somehow can't find it (and ridiculous excuses as to why they keep overlooking it ensue). I might as well claim that modern north Europeans carry a sizeable Siberian admixture offset by another southern European admixture, find northern Europeans admixed with Sami and southern Europeans, label them as "northern European" and use the data I myself have collected to prove my theory, pace to the eventual fact that the scientific literature would not support my view and actually debunk it. It is how ludicrous these theories about muh east meds have become.

 

[Leopoldo Leone]

Well, I have cited two studies (Raveane et al. 2022 and Sarno et al. 2021) that suggest that it does make sense. Raveane et al. 2022 particularly with modeling Apulians with Minoans. You can play around with the out groups all you'd like. But I'm more inclined to believe that the authors are competent in coming to their conclusions.


Also, it is an appeal to novelty fallacy to disregard Dodecad as being worse than G25, simply because it is older.


Also, this is a red herring, you are trying to divert attention away from my initial remark that because South Italians and Maniots have more an affinity to Greece_N/Minoans, they are the closest to the ancient Greeks. So now you have to come up with sophistry to try to disprove two academic papers to try to pretend there's all this Levantine in these populations. Come on man, I wasn't born yesterday. Maybe you don't want to produce the model with Greece_N or Minoan because it will prove my point. That's not fair.

You see, actually amateur hobbyists who believed such BS that south Italians were literally a mix of a north-Italian-like population with a straight Lebanese-like population and kept changing their claims when the evidence debunked their previous claims- always concocting though models in which there is a conspicuous Levantine admixture- are more trustworthy than professional genetists, who keep making naive blunders, as if they had no precise idea of what they are doing. Especially the former are more trustworthy when they insist on using their data (either collected or processed, and the second part is as opaque as the first) and their calculator,

You are discussing with a group of people that has forsaken any contact with reality and has decided to believe in their headcanon, so there is no data getting in the ways of their fantasies; I've learnt to draw the lines when the other party decides not to engage with the relevant data and uses instead their own data (for example, how can I forget the anthrogenica users' self-collected Y haplo list from self-identified south Italians from the internet which very conveniently showed "a sizeable overlap between south Italy's Y haplo gene pool and the Levant's"? It is not as if the internet is teeming with genetic kits of fake Italians and as if there are plenty of official papers with samples from Italy).

 

[Idontknowwhatimdoing]

Your mistake is to believe that G25 samples are representative, but it is clear there is something fishy about them, and there is no real historical process that explains how all that Levantine-Anatolian ancestry ended up in south Italy (muh slaves and hellenistic "Greeks" do not make sense either), conveniently offset by another unexplained northern gene flow. Furthermore your bunch must stop pretending as if professional genetists were clueless and did not care to use relevant outgroups in their tests, and in the literature you find papers that do use middle easterners as a reference: https://www.biorxiv.org/content/10.1101/2022.02.26.482072v1


Furthermore there is no Levant_N excess in south Italians compared to ancient SE european samples, it is just the fantasies of delusional hobbyists who keep finding it DESPITE the fact that professional genetists somehow can't find it (and ridiculous excuses as to why they keep overlooking it ensue). I might as well claim that modern north Europeans carry a sizeable Siberian admixture offset by another southern European admixture, find northern Europeans admixed with Sami and southern Europeans, label them as "northern European" and use the data I myself have collected to prove my theory, pace to the eventual fact that the scientific literature would not support my view and actually debunk it. It is how ludicrous these theories about muh east meds have become.

If you scroll a bit up you will see my qpAdm models proving that South Italians have excess Natufian admix.

 

[Leopoldo Leone]

If you scroll a bit up you will see my qpAdm models proving that South Italians have excess Natufian admix.

1) What samples did you use?
2) Explain why professional genetists did not find such a conspicuous admixture.

 

[Idontknowwhatimdoing]

Well, I have cited two studies (Raveane et al. 2022 and Sarno et al. 2021) that suggest that it does make sense. Raveane et al. 2022 particularly with modeling Apulians with Minoans. You can play around with the out groups all you'd like. But I'm more inclined to believe that the authors are competent in coming to their conclusions.


Also, it is an appeal to novelty fallacy to disregard Dodecad as being worse than G25, simply because it is older.


Also, this is a red herring, you are trying to divert attention away from my initial remark that because South Italians and Maniots have more an affinity to Greece_N/Minoans, they are the closest to the ancient Greeks. So now you have to come up with sophistry to try to disprove two academic papers to try to pretend there's all this Levantine in these populations. Come on man, I wasn't born yesterday. Maybe you don't want to produce the model with Greece_N or Minoan because it will prove my point. That's not fair.

Bruh, i dont disagree with you about South Italians being the closest to ancient Greeks. The excess Iran N admix in South Italians might even help them plot closer to ancient Greeks too, + the ancient Italian and Greek admix

 

[Idontknowwhatimdoing]

1) What samples did you use?
2) Explain why professional genetists did not find such a conspicuous admixture.

I used the default dataset for qpAdm from Harvard.

2) I explained why their models were wrong, did you people not read my explanation?

3) "Professional" geneticists does not hold much value here. being paid to produce research for universities doesn't make you an expert. They make mistakes and are often very outdated. Even Lazaridis from Harvard the creator of qpAdm made mistakes which people on anthrogenica were pointing out before Lazaridis made a new paper that fixed the mistakes.

 

[mount123]

great work
also this dude
GLI002- Glika Nera Attica - ( according to the ytree site he is E-S9621 an upstream branch to e-m84)
https://www.theytree.com/usersample/1e27887ee0abb47a9d5bcdaa9f987013.html

Looking over this again, I see no upstream pre-M84 (<M34) branch on FTDNA & YFull but yes, he doesn't seem to be M84 so I'll just add the designation S9621 for the sample (pre M84).

 

[Leopoldo Leone]

I used the default dataset for qpAdm from Harvard.

2) I explained why their models were wrong, did you people not read my explanation?

3) "Professional" geneticists does not hold much value here. being paid to produce research for universities doesn't make you an expert. They make mistakes and are often very outdated. Even Lazaridis from Harvard the creator of qpAdm made mistakes which people on anthrogenica were pointing out before Lazaridis made a new paper that fixed the mistakes.

1) Your explanation to why they are wrong is that they didn't use Levant_N alongside Iran_N, but they did use as reference Levant_N in the Moot paper about Rome to model the Latin outliers and they came up with no excess Levant_N, and they used modern MENA samples in a f4 tests in the paper I linked and it gave negative results, and you are wrong in assuming it is literally Iran_N when the actual literature can't tell it apart from CHG, which is a more plausible input which came from Anatolia first to SE Europe and then south Italy.
2) Here it is where we depart, since you are accusing not one but genetists in general of being less adept than you are at doing their job; one thing it is to criticise one paper, another to believe that all are wrong.
3) If Lazaridis, the creator of qpAdm, made mistakes, using it I presume, why do you think you are not making big mistakes too? 10-15% natufian in excess in modern south Italians translates to at least 20-30% Levant_BA admixture, which is utterly implausible, and isn't supported by any other piece of evidence.

 

[Idontknowwhatimdoing]

Also, this is a red herring, you are trying to divert attention away from my initial remark that because South Italians and Maniots have more an affinity to Greece_N/Minoans, they are the closest to the ancient Greeks. So now you have to come up with sophistry to try to disprove two academic papers to try to pretend there's all this Levantine in these populations. Come on man, I wasn't born yesterday. Maybe you don't want to produce the model with Greece_N or Minoan because it will prove my point. That's not fair.

1) I've never denied that South Italians are closer genetically to ancient Greeks than most mainland Greeks. It is common sense. The closest modern Greeks to ancient Greeks also have excess Levantine Natufian admix.

2) Academic papers are often wrong, a lot of the people working on them are a bit inexperienced with autosomal archaeaogenetics and make mistakes. I explained very well why they couldnt use Natufian as a proxy with their outgroups. qpAdm needs Anatolian_HG and Taforalt in order to help qpAdm tell apart Natufian from Anatolia_N, they did not use them for some reason, it seems that they did not pay attention to Lazaridis's paper outgroups, one of the creators of qpAdm.

3) South Italians still score excess Natufian even when using Minoans or Peloponnese_N as a proxy. As you can see below. This model still proves my point and not yours. There is obvious excess Levantine Natufian admix in South Italians compared to Bronze age and Iron Age Italians and Aegeans.

You see, actually amateur hobbyists who believed such BS that south Italians were literally a mix of a north-Italian-like population with a straight Lebanese-like population and kept changing their claims when the evidence debunked their previous claims- always concocting though models in which there is a conspicuous Levantine admixture- are more trustworthy than professional genetists, who keep making naive blunders, as if they had no precise idea of what they are doing. Especially the former are more trustworthy when they insist on using their data (either collected or processed, and the second part is as opaque as the first) and their calculator

Leopoldo, i made qpAdm models using improved outgroups from the paper that improve STD errors using the default Harvard qpAdm dataset, i did not just use "calculators". Maybe next time pay attention to the qpAdm models i posted using the official dataset from the creators of qpAdm instead of assuming that i just used "calculators".

These below are qpAdm models using similar outgroups to the paper.

 

[Idontknowwhatimdoing]

1) Your explanation to why they are wrong is that they didn't use Levant_N alongside Iran_N, but they did use as reference Levant_N in the Moot paper about Rome to model the Latin outliers and they came up with no excess Levant_N, and they used modern MENA samples in a f4 tests in the paper I linked and it gave negative results, and you are wrong in assuming it is literally Iran_N when the actual literature can't tell it apart from CHG, which is a more plausible input which came from Anatolia first to SE Europe and then south Italy.

3) If Lazaridis, the creator of qpAdm, made mistakes, using it I presume, why do you think you are not making big mistakes too? 10-15% natufian in excess in modern south Italians translates to at least 20-30% Levant_BA admixture, which is utterly implausible, and isn't supported by any other piece of evidence.

1) I've said that they should use Anatolian_HG and Taforalt as outgroups/right pops in order for qpAdm to tell apart the left pops/proxies Anatolia_N from Levant_N or Natufian. They did not use those 2 important outgroups. How is qpAdm going to distinguish between Levant_N and Anatolia_N if Anatolian_HG and Taforalt or Natufian are not used as right pops/outgroups? Do you even know the difference between left pops/proxies and right pops/outgroups?

3) Do you even pay attention to what i say? I've already explained why Natufian is inflated unrealistically if you scroll up. On qpAdm when using too many related proxies it gives high STD errors. South Italians have slight North African admix but i cant add it as a proxy on qpAdm for that reason. When not adding the small North African % in South Italians then qpAdm is forced to compensate for it with Natufian.

On g25 it works fine and it shows normal levels of Natufian in South Italians because i can use Taforalt as a proxy too.

G25

vs qpAdm models

​

 

[Leopoldo Leone]

3) Do you even pay attention to what i say? I've already explained why Natufian is inflated unrealistically if you scroll up. On qpAdm when using too many related proxies it gives high STD errors. South Italians have slight North African admix but i cant add it as a proxy on qpAdm for that reason. When not adding the small North African % in South Italians then qpAdm is forced to compensate for it with Natufian.

On g25 it works fine and it shows normal levels of Natufian in South Italians because i can use Taforalt as a proxy too.

G25

vs qpAdm models

​

1)G25 is off the chart, it shows as well 20-30% Levant_BA/15% Levant_N in south Italy which is BS, and it is not my word but the consensus of the actual scientific literature, which you dismiss by saying that "official papers are often wrong", and I can retort that often hobbyists like yourself are wrong, and I have cited papers that used other tests to check whether modern south Italians show any trace of admixture with MENA people (besides the low north Africans, mostly concentrated in Sicily), and their results were negative, though I take it that you dismiss them too.
2) Instead of assuming that south Italians have Levantine admixture, which is not only not supported but debunked by the actual data, since you say that the qpAdm is inflating natufian because it is compensating for the lack of north African proxies, which you can't use together with natufian, why don't you use just a north african proxy and see the results? As far as I am seeing it, you are claiming that qpAdm needs something extra to model south Italians compared to BA and IA ones (at least Sicilians and "south Italians" who are Calabrians), which you also say it is the low amount north african in them (and no one disputes that) which inflates greatly the natufian in qpAdm. Fine, nothing wrong up to here; what is a logical gap is to say that G25 does not present this problem and hence is "fine" when there are plenty of reasons to think it is not.
3) You claim that also the modern Greeks closest to ancient ones present an excess of Natufian/Levantine? Ancient Greeks did not have such extra natufian and you'd need also an extra northern gene flow to compensate the southern one, which is an extremely dubious model.

 

[Idontknowwhatimdoing]

1)G25 is off the chart, it shows as well 20-30% Levant_BA/15% Levant_N in south Italy which is BS, and it is not my word but the consensus of the actual scientific literature, which you dismiss by saying that "official papers are often wrong", and I can retort that often hobbyists like yourself are wrong, and I have cited papers that used other tests to check whether modern south Italians show any trace of admixture with MENA people (besides the low north Africans, mostly concentrated in Sicily), and their results were negative, though I take it that you dismiss them too.
2) Instead of assuming that south Italians have Levantine admixture, which is not only not supported but debunked by the actual data, since you say that the qpAdm is inflating natufian because it is compensating for the lack of north African proxies, which you can't use together with natufian, why don't you use just a north african proxy and see the results? As far as I am seeing it, you are claiming that qpAdm needs something extra to model south Italians compared to BA and IA ones (at least Sicilians and "south Italians" who are Calabrians), which you also say it is the low amount north african in them (and no one disputes that) which inflates greatly the natufian in qpAdm. Fine, nothing wrong up to here; what is a logical gap is to say that G25 does not present this problem and hence is "fine" when there are plenty of reasons to think it is not.
3) You claim that also the modern Greeks closest to ancient ones present an excess of Natufian/Levantine? Ancient Greeks did not have such extra natufian and you'd need also an extra northern gene flow to compensate the southern one, which is an extremely dubious model.

1) 15% Levant_N does imply around 20% Levant_BA which is absolutely not bullshit. Did you see the Rome_Imperial samples in Italy? Its full of straight up Levantines.

2) When using too many related proxies on qpAdm it gives high STD errors, i explained that many times, i cant use North africans on top of Natufian and on top of ANF as proxies.

3) I've never claimed that ancient Greeks had such extra Natufian, i've said it to show that despite the excess Natufian they are still one of the closest people genetically to them.

 

[kingjohn]

Looking over this again, I see no upstream pre-M84 (<M34) branch on FTDNA & YFull but yes, he doesn't seem to be M84 so I'll just add the designation S9621 for the sample (pre M84).

Yes they are not in yfull and ftdna ytree as
pre-e-m84
But he carried in yfull 6 negetive calls + 5 positive calls
In the e-m84 level ( according to the reading of his bam file by theytree site)
Thats why this sample is huge and extremely important because it created a splitt inside
e-m84
Honestly i wish pribislav would take a look at it as well
Yes you can put e-s9621 at the moment

P.s
It is likely glica nera site remains are from late bronze age not the iron age as i thought
he could be some rare branch of
e-m34 who made it to attica greece in late bronze age but died out
As most modern day greek e-m34 are mainly
e-m34> e-z841 branches and likely arrived to greece
Much later
In roman period and late antiquity

 

[Leopoldo Leone]

1) 15% Levant_N does imply around 20% Levant_BA which is absolutely not bullshit. Did you see the Rome_Imperial samples in Italy? Its full of straight up Levantines.

2) When using too many related proxies on qpAdm it gives high STD errors, i explained that many times, i cant use North africans on top of Natufian and on top of ANF as proxies.

3) I've never claimed that ancient Greeks had such extra Natufian, i've said it to show that despite the excess Natufian they are still one of the closest people genetically to them.

1) It is BS for the evidence points to a relatively stable population structure in Europe, https://www.biorxiv.org/content/10.1101/2022.05.15.491973v1.full, furthermore "muh Rome imperial samples" actually are mostly "east med" who were half Anatolian-half Balkanites ( https://www.biorxiv.org/content/10.1101/2021.08.30.458211v1.full see how the near eastern cluster is modelled in the supplementaries, and we know that those near easterners were Anatolians), only 28% were "near easterners" or Levantine, and if you remember the moot paper it said that the tail to the Levant disappeared after the capital of the empire was moved to Constantinople, and southern Italians overlaps with the "mediterranean" or Greek-like cluster, which existed in Italy before the imperial era, and the genetic literature supports high level of similarity between ancient Greeks and Italians, so the most plausible theory is that southern Italians mostly descend from a Greek-like pop (actual ancient Greeks and Sicily_IA), not your implausible "east med shift" during the empire followed by a "west Europe shift" during the middle age, which somehow cancelled each other out.
2) Just use talforat instead of Natufian.
3) Let's say that modern Greeks have indeed extra Natufian, for the argument's sake, have you tried using Anatolia_BA as a source? The problem with the models I see on anthrogenica is that they require straight Levantine on top of any Anatolian, which is by the way often still blown out of proportion, again to be offset by a generous Slavic admixture even in the extreme south of Greece.

 

[Francesco]

1) I've said that they should use Anatolian_HG and Taforalt as outgroups/right pops in order for qpAdm to tell apart the left pops/proxies Anatolia_N from Levant_N or Natufian. They did not use those 2 important outgroups. How is qpAdm going to distinguish between Levant_N and Anatolia_N if Anatolian_HG and Taforalt or Natufian are not used as right pops/outgroups? Do you even know the difference between left pops/proxies and right pops/outgroups?
3) Do you even pay attention to what i say? I've already explained why Natufian is inflated unrealistically if you scroll up. On qpAdm when using too many related proxies it gives high STD errors. South Italians have slight North African admix but i cant add it as a proxy on qpAdm for that reason. When not adding the small North African % in South Italians then qpAdm is forced to compensate for it with Natufian.
On g25 it works fine and it shows normal levels of Natufian in South Italians because i can use Taforalt as a proxy too.
G25

vs qpAdm models

​

Assuming an impulse from the levant, southern Italy should show an excess of Iran_N as well (since both levantine and Anatolian populations of Bronze and Iron Age were roughly an equal mix of the two), while your model shows a much greater component of Levant_N, wich doesn't add up very well. For it to be correct, we should maybe conclude tha all the Iran_N ancestry in Italy was bought by imperial Era immigrants, but we know that it precedes them.

 

[Idontknowwhatimdoing]

1)G25 is off the chart, it shows as well 20-30% Levant_BA/15% Levant_N in south Italy which is BS, and it is not my word but the consensus of the actual scientific literature, which you dismiss by saying that "official papers are often wrong", and I can retort that often hobbyists like yourself are wrong, and I have cited papers that used other tests to check whether modern south Italians show any trace of admixture with MENA people (besides the low north Africans, mostly concentrated in Sicily), and their results were negative, though I take it that you dismiss them too.
2) Instead of assuming that south Italians have Levantine admixture, which is not only not supported but debunked by the actual data, since you say that the qpAdm is inflating natufian because it is compensating for the lack of north African proxies, which you can't use together with natufian, why don't you use just a north african proxy and see the results? As far as I am seeing it, you are claiming that qpAdm needs something extra to model south Italians compared to BA and IA ones (at least Sicilians and "south Italians" who are Calabrians), which you also say it is the low amount north african in them (and no one disputes that) which inflates greatly the natufian in qpAdm. Fine, nothing wrong up to here; what is a logical gap is to say that G25 does not present this problem and hence is "fine" when there are plenty of reasons to think it is not.
3) You claim that also the modern Greeks closest to ancient ones present an excess of Natufian/Levantine? Ancient Greeks did not have such extra natufian and you'd need also an extra northern gene flow to compensate the southern one, which is an extremely dubious model.

I used the same outgroups, methods to model Cypriots as the paper and according to their shitty paper outgroups the Greek Cypriots have no Levantine admix therefore zero Phoenician. Obviously their outgroups are bad, i explained why many times. Even Cypriots who have 20-30% Phoenician admix score zero Natufian/Levant_N with their outgroups.

I want you to explain me why when using the same outgroups as the paper the Cypriots also can be modeled like Italians with 0% Levantine. Are you actually claiming their models are so good that they proved Cypriots have no Phoenician admix?

Cypriot results:
60.2% Anatolia_N
28.3% Iran_N
11.6% Yamnaya

South Italian results:
58.1% Anatolia_N
17.3% Iran_N
24.2% Yamnaya
0,465% WHG

 

[Leopoldo Leone]

I used the same outgroups, methods to model Cypriots as the paper and according to their shitty paper outgroups the Greek Cypriots have no Levantine admix therefore zero Phoenician. Obviously their outgroups are bad, i explained why many times. Even Cypriots who have 20-30% Phoenician admix score zero Natufian/Levant_N with their outgroups.

Cypriot results:
60.2% Anatolia_N
28.3% Iran_N
11.6% Yamnaya

South Italian results:
58.1% Anatolia_N
17.3% Iran_N
24.2% Yamnaya
0,465% WHG

Your mistake is to assume that what applies to Cyprus must apply to south Italy as well: for Cyprus zero Levantine ancestry is a priori implausible, but for south Italy 20% Levantine ancestry is a priori implausible since it is separated by 1000+ km from the Levant and there is no historical event which might explain such an admixture. On the other hand, low levels of north african input in south Italy are plausible and confirmed, which are what causes the ridiculously inflated levels of natufian in qpAdm, but you refuse to use just talforat to see what happens and you are appealing to ignorance to what the true levels of north african are according to qpAdm to claim that there are both, which is a logical gap.

P.S. Furthermore, if there is 20% Levant_BA in south Italy, on top of the low levels of north Africans, the only way for them to be still able to be similar to ancient Greek samples (and also ancient Thracian samples) is for them to have also some extra European-like admixture, and that is "accounted" on your models by the 30% Steppe in south Italians, which translates to at least a double digit contribution from central/north European-like population, and this is exactly the "anthrogenica model" and which is not supported either by the majority of genetic literature or history.