Genetic and Cultural Differences between Jews and Greeks

Palermo Trapani said:
Leopoldo: That Palermo Trapani is the one typing and posting to you right now. Glad you cited my comment from over there. I jumped the gun and did not do my due diligence and homework. But after examining Morocco_LN, who is at best 10% Local Berber, if that, 10% of 46% Morroco_LN will get you to the estimate of 4.6% reported in Sazzini et al 2016 and the similar admixture rate reported in Raveane et al 2019.

Thanks for finding that comment. I think it makes my point without saying anymore and again I appreciate you citing that here.(y)
Click to expand...

Good work, Palermo.

More than due diligence.
 
Leopoldo Leone said:
In reality the only fault of the paper was that it didn't clarify that the "Moroccan_LN" is not what you might think at a first glance, as the comment of Palermo trapani under a comment of a "guest" clarifies, https://www.biorxiv.org/content/10.1101/584714v1.
Click to expand...

Err... not. See KEB (that's Morocco_LN): https://www.pnas.org/content/pnas/115/26/6774/F2.large.jpg?width=800&height=600&carousel=1

Nothing Mycenaean/Minoan-like about it, unless you think Mycenaean/Minoan-like was undisputably North African.

The paper (https://www.pnas.org/content/115/26/6774) that released the data on that sample unmistakably asserts:

"At K = 5, TOR is composed of the component associated with the European Early Neolithic and IAM is composed of the North African component observed in Mozabites. KEB is placed in an intermediate position, with ∼50% each of European Early Neolithic and North African ancestries."

Moreover, KEB samples are placed halfway between the IAM and Anatolian/European farmer clusters, in close proximity to Levant aDNA samples and also to Guanche samples (16) (from the indigenous population of the Canary Islands known to have a Berber origin; ref. 23).

The same pattern is observed for Taforalt samples, which shows its lower FST distance with IAM, followed again by KEB (0.129), the Guanches (0.150), and modern North Africans (0.130 to 0.149).

[FONT=&quot]All comparisons produced negative values of the f3-statistic, which suggests that the [/FONT][FONT=&quot]KEB population can be modeled as a mixture of IAM and Anatolian/European Neolithic."[/FONT][FONT=&quot][/FONT]

CLOSEST MODERN INDIVIDUALS COMPARED TO MOROCCO_LN (KEB):

Distance to:MAR_LN:KEB.4
0.07736599Moroccan_North:MNA3
0.08080144Moroccan_North:MNA6
0.08179377Berber_Tunisia_Chen:T39
0.08412465Algerian:ALG200
0.08431946Berber_Tunisia_Chen:T35
0.08434668Algerian:ALG900
0.08694789Mozabite:HGDP01263
0.08713594Moroccan_North:MCH9
0.08899931Berber_Tunisia_Chen:T52
0.08941020Berber_Tunisia_Chen:T44
0.09028865Algerian:ALG600
0.09094180Moroccan_North:MNA1
0.09154340Moroccan_North:MCH13
0.09198705Moroccan_North:MOJ2
0.09220874Berber_Tunisia_Chen:T42
0.09229322Moroccan_North:MCH11
0.09295668Berber_Tunisia_Chen:T43
0.09315742Moroccan_North:MCH8
0.09375413Algerian:Algerian43A34
0.09397438Berber_Tunisia_Chen:T11
0.09417848Berber_Tunisia_Chen:T1
0.09488236Moroccan_North:MCH18
0.09513748Moroccan_North:MNA4
0.09580645Berber_Tunisia_Sen:BerSH2
0.09631119Tunisian:Tunisian20C4


CLOSEST ANCIENT INDIVIDUALS TO MOROCCO_LN (KEB)

Distance to:MAR_LN:KEB.4
0.06336924ITA_Sardinia_Punic:VIL011
0.07837974ITA_Rome_Imperial:RMPR132
0.08394185Canary_Islands_Guanche:gun005
0.08541817ITA_Sardinia_Punic:VIL010
0.08676415ITA_Sardinia_Punic:VIL006
0.09003592Iberia_Central_CA_Afr:I4246
0.09388315Canary_Islands_Guanche:gun008
0.10630851ITA_Sardinia_Punic:VIL007
0.10642779Canary_Islands_Guanche:gun012
0.10773508Iberia_Southeast_c.10-16CE:I8146
0.10814317Canary_Islands_Guanche:gun002
0.10939985ITA_Rome_Imperial:RMPR80
0.11118689Canary_Islands_Guanche:gun011
0.11165249ITA_Etruscan_o:RMPR475b
0.11399745Iberia_Southeast_c.5-8CE:I3980
0.11559692Levant_PPNB:I0867
0.11717896Levant_PPNC:I1699
0.11796247Levant_PPNB:I1704
0.11802590Iberia_Southeast_c.3-4CE:I3983
0.12120813Iberia_Southeast_c.10-16CE:I7457
0.12379721Levant_PPNB:I1707
0.12390148Levant_PPNB:BAJ001
0.12399837Iberia_Southeast_c.10-16CE:I3808
0.12507911Iberia_Southeast_c.3-4CE:I4054
0.12616588Iberia_Southeast_c.5-8CE:I3581
0.12646303Iberia_Southeast_c.5-8CE:I3981
0.12697912Iberia_Ibiza_Punic:MS10614
0.12701377EGY_Late_Period:JK2134
0.12722007Iberia_Southeast_c.5-8CE:I3576
0.12758295Iberia_Southeast_c.10-16CE:I7499

 
Palermo Trapani said:
Leopoldo: That Palermo Trapani is the one typing and posting to you right now. Glad you cited my comment from over there. I jumped the gun and did not do my due diligence and homework. But after examining Morocco_LN, who is at best 10% Local Berber, if that, 10% of 46% Morroco_LN will get you to the estimate of 4.6% reported in Sazzini et al 2016 and the similar admixture rate reported in Raveane et al 2019.

Thanks for finding that comment. I think it makes my point without saying anymore and again I appreciate you citing that here.(y)
Click to expand...

I think this is much more plausible.

5bkGrqm.png


Sicilians would not plot where they do, if Moroccan_LN was not actually very Greeks-like, plus that amount of Iran_N, Anatolian_N, and bit of steppe.
 
Palermo Trapani said:
Ygorcs: I am not going to get into the squabble here but I would be willing to bet that you are aware that Morroco_LN has little native Berber ancestry, in fact looks 80% Anatolian_EEF, some WHG, and some Iran_NEO, so to get the appropriate North African_Berber admixture in modern sicily, Morroco_LN was used. You can see the Fernandes et al 2020 Table 2 which has the ancestral admixture of Morroco_LN. Marcus et al 2020 I think also used a 3-Way model which included Morroco_LN, or some similar source and the result of the Sicilian admixture chart is presented in Figure 4 right next to Modern Tuscany. My personal opinion, I would expect better for a Moderator here rather than this type of childish behavior. Of course, that last statement is just my opinion


bQQdpnc.jpg



FvnEvaH.jpg
Click to expand...

Not childish at all. It's the latest model made by professional geneticists that we have, and of course it's also a heavily mixed and somewhat drifted source population, so they must be able to distinguish it from other ANF-admixed source populations. Otherwise, we would be totally unable to distinguish, for example, Levant_N from Anatolia_N, or North African ancestry from European ancestry. They share ancestry, but are different. Also, I'm not sure why you're pointing that out. I had explained before that the higher Levant_N and Morocco_EN in modern Sicilians and South Italians must arrived together with other admixtures, so already diluted, but you and others didn't seem to understand that point. Now all of a sudden you can perfectly understand that those ancestry models must not be interpreted literally and that some source population may be only representing a certain combination of admixtures that may have arrived together all at once or accumulated over time? Why the difference in understanding in one case and in the other?
 
Jovialis said:
I think this is much more plausible.

5bkGrqm.png


Sicilians would not plot where they do, if Moroccan_LN was not actually very Greeks-like, plus that amount of Iran_N, Anatolian_N, and bit of steppe.
Click to expand...

Didn't you point out from the start, regarding Greeks vs. Jews, that plotting on a similar position does not mean necessarily having the same admixtures, let alone in similar proportions? And I agree with you, indeed.

Secondly, I'm glad to know that now you guys do think you can sometimes question and even know better than the authors of the study that presents the aDNA samples themselves. It's alright if that happens. When problems are not in the models themselves due to the choice of not really proximate sampls (e.g. the use of Levant_N in the recent Fatyanovo paper), problems with the literal interpretation of models that are actually just suggestive of the real things in genetic history are also usual.

https://www.pnas.org/content/pnas/115/26/6774.full.pdf

"At K = 5, TOR is composed of thecomponent associated with the European Early Neolithic and IAMis composed of the North African component observed in Mozabites. KEB is placed in an intermediate position, with ∼50% each ofEuropean Early Neolithic and North African ancestries.

Moreover, KEB samples are placed halfway between the IAM and Anatolian/European farmer clusters, in closeproximity to Levant aDNA samples and also to Guanche sample.

Lastly, KEB can be explained ashaving both IAM-like and Iberian Early Neolithic components (Fig.2). The same admixture profile is observed in the Guanche samples,but the amount of IAM ancestry is consistently higher in all of thesamples."
 
But Ygorcs, the same study shows that Morrocan_LN indeed has very little actual berber in it, to Palemo's point. Which perfectly explains why Raveane et al 2019 shows so little Moroccan-like ancestry in Sicily.

If they had that much North African, they would be in a radically different postion on the PCA.

I have to say, even the people at AG, and Davidski didn't buy this result when it first came out, and were suprised it passed peer-review.
 
Jovialis said:
But Ygorcs, the same study shows that Morrocan_LN indeed has very little actual berber in it, to Palemo's point. Which perfectly explains why Raveane et al 2019 shows so little Moroccan-like ancestry in Sicily.
Click to expand...

Berber Moroccan-like ancestry necessarily and automatically includes a lot of ANF (and some other more minor admixtures). Considering Moroccan-like only what is Taforalt/Morocco_EN-like is a mistake. It'd be like saying that only the Yamnaya admixture is really indicative of European genetic contribution. That's why I said before: Morocco_EN, Levant_N and other ancient admixtures are not to be interpreted literally, they may have arrived in completely diluted forms within other generally much more distant populations that simply had acquired some of that admixture before (even very long before).
 
Jovialis said:
I have to say, even the people at AG, and Davidski didn't buy this result when it first came out, and were suprised it passed peer-review.
Click to expand...

I don't either. I think it's bad modelling. I think using Levant_N to give a lot of Levant_N and negligible WHG in Fatyanovo samples is also the result of bad modelling. And I also think other peer-reviewed and released studies made bad modellings and should be interpreted cautiously because they are not really picking everything that is there in an accurate way. But I'm just an amateur with some good sense. I think you may not have really got the main point of my former comment.
 
@Ygorcs, I don't take what is published as gospel, which is evident by my position to the conclusions made by Antonio m. et al 2019. I think it makes a lot of naïve and broad assumpstions, with a very limited sample set. Moreover, in regards to Fernandes et al 2020, I think the fact that Morocco_LN looks very Mycenaean-like, makes their conclusion extremely tenuous. The main issue I had was with G25, and the components constructed from averages, which I don't agree is a proper way to analyze the publicly available samples. I am all for hobbyists like us examining samples, which is why me and Duarte made so many coordinates for Dodecad on Vahaduo. Though I recognize that it too has some flaws, due to the fact it uses modern components, which is not adequate for some ancient samples, particularly those that fall outside of modern genetic continuums.
 
Incidentally, I have found that, comparing the best qpAdm models obtained by the Sicilian/Sardinian aDNA paper with G25 results using the same pool of samples, there is not a lot of difference at all, quite on the contrary. Maybe the real problem is when you don't have the best samples and don't choose the most parsimonious and realistic ones. E.g.:

In the supplement the authors write: We were able to model the Bronze Age Sardinians in a similar way as the two previous main groups, here as 82.9 ± 0.9% Anatolia_Neolithic and 17.1 ± 0.9% WHG (p=0.380) (Supplementary Table 14). qpAdm provided no insights as to why I10365 was an outlier in qpWave analysis, as Sardinia_BA10365 also produced a valid P-value for this 2-way model (p=0.120), with 85.6 ± 2.2% Anatolia_Neolithic and 14.4 ± 2.2% WHG.

What I got using G25 samples with similar basic source populations: 82.5% Anatolia_Neolithic + 17.4% WHG + 0.1% others (very similar results). For the 10365 outlier (in qpWave, but not in other tools) individual, I got 85.4% Anatolia_Neolithic + 14.4% + 0.2% others (again almost identical to the authors' best fit model).

TargetDistanceGEO_CHGMAR_ENTUR_Barcin_NTUR_Boncuklu_NWHG
ITA_Sardinia_EBA:I02090.056582040.00.083.00.017.0
ITA_Sardinia_EBA:ISB0010.041777810.00.082.60.017.4
ITA_Sardinia_EBA:LON0010.044903250.00.083.20.016.8
ITA_Sardinia_EBA:MA850.056354140.02.479.40.018.2
ITA_Sardinia_EBA:MA880.032667440.00.080.80.019.2
ITA_Sardinia_EBA:SUA0020.039643900.00.081.40.018.6
ITA_Sardinia_EBA:SUA0070.043594370.00.081.80.018.2
ITA_Sardinia_EBA:SUC0010.035961420.00.081.20.018.8
ITA_Sardinia_EBA:SUC0020.046573010.00.284.80.015.0
ITA_Sardinia_EBA:SUC0030.040780760.00.282.40.017.4
ITA_Sardinia_EBA:SUC0040.046174870.00.079.62.617.8
ITA_Sardinia_EBA:SUC0050.038694370.00.079.80.020.2
ITA_Sardinia_EBA:SUC0060.046358890.00.085.40.014.6
ITA_Sardinia_EBA:SUC0070.035705670.00.474.20.025.4
ITA_Sardinia_EBA:SUC0090.047662830.00.083.20.016.8
ITA_Sardinia_Nuragic:ISC0010.041729980.00.080.41.817.8
ITA_Sardinia_Nuragic:I37410.046325180.00.081.81.217.0
ITA_Sardinia_Nuragic:I37430.040529980.00.083.00.017.0
ITA_Sardinia_Nuragic:I103640.041780530.00.084.20.015.8
ITA_Sardinia_Nuragic:I103650.026268990.20.066.818.614.4
ITA_Sardinia_Nuragic:I105520.058284740.00.082.20.017.8
ITA_Sardinia_Nuragic:I105530.038299170.00.081.40.018.6
ITA_Sardinia_Nuragic:I105540.043056440.00.084.40.015.6
ITA_Sardinia_Nuragic:I161610.045935220.00.081.00.019.0
ITA_Sardinia_Nuragic:I161680.051591240.00.075.410.813.8
ITA_Sardinia_Nuragic:I161690.055897960.00.080.40.019.6
ITA_Sardinia_Nuragic:I161700.047321770.00.082.60.017.4
ITA_Sardinia_Nuragic:I161830.043616420.00.476.47.815.4
ITA_Sardinia_Nuragic:MA820.043708060.00.081.20.018.8
ITA_Sardinia_Nuragic:MA1100.054298870.00.080.40.019.6
ITA_Sardinia_Nuragic:ORC0010.038743970.00.083.01.815.2
ITA_Sardinia_Nuragic:ORC0030.043971960.00.068.414.617.0
ITA_Sardinia_Nuragic:ORC0040.041027130.00.078.85.615.6
ITA_Sardinia_Nuragic:ORC0050.041451300.00.076.47.815.8
ITA_Sardinia_Nuragic:ORC0060.046399240.00.085.00.015.0
ITA_Sardinia_Nuragic:ORC0070.044588860.00.083.80.016.2
ITA_Sardinia_Nuragic:ORC0080.048001200.00.082.80.017.2
ITA_Sardinia_Nuragic:ORC0090.037334650.00.082.00.018.0
ITA_Sardinia_Nuragic:SUA0010.036236950.00.060.622.017.4
ITA_Sardinia_Nuragic:SUA0030.051270210.00.051.430.218.4
ITA_Sardinia_Nuragic:SUA0060.054459230.00.082.40.017.6
Average0.044282050.00.179.53.017.4


The authors write: In Sicily, the Middle Neolithic individuals can be modeled as 2-way mixtures of Anatolia_Neolithic (89.5 ± 1.2%) and WHG (10.5 ± 1.2%) (p=0.338).
What I got using G25 models: 89.9% Anatolia_Neolithic + 10.1% WHG (virtually identical)


Target
DistanceTUR_Barcin_NTUR_Boncuklu_NWHG
ITA_Sicily_MN:I40620.0225600784.68.47.0
ITA_Sicily_MN:I40630.0236867370.419.010.6
ITA_Sicily_MN:I40640.0211613591.40.28.4
ITA_Sicily_MN:I40650.0171841885.60.014.4
Average0.0211480883.06.910.1

The authors write: The outliers Sicily_EBA11443, Sicily_EBA8561, and Sicily_EBA3124 possessed large proportions of Yamnaya_Samara-related ancestry (39.0 ± 3.5, 22.1 ± 3.6%, and 13.5 ± 3.4% respectively), showing this ancestry was already in Sicily at least in outlier individuals by this time.
What I get using the proper G25 model: 43.6%, 19.8% and 15.2%, respectively, well within the range the authors estimated.


TargetDistanceIRN_Ganj_Dareh_NMAR_ENTUR_Barcin_NTUR_Boncuklu_NTUR_Kumtepe_NTUR_Tepecik_Ciftlik_NWHGYamnaya_RUS_Samara
ITA_Sicily_EBA:I31220.026499200.00.071.018.60.00.010.40.0
ITA_Sicily_EBA:I31230.024676241.60.477.40.00.01.29.010.4
ITA_Sicily_EBA:I31240.021125850.00.474.40.00.00.010.015.2
ITA_Sicily_EBA:I78070.031811340.00.054.024.014.21.85.20.8
ITA_Sicily_EBA:I85610.029042520.00.045.022.40.00.012.819.8
ITA_Sicily_EBA:I114420.022149850.01.675.40.00.04.08.410.6
ITA_Sicily_EBA:I114430.022727140.00.245.03.00.00.08.243.6
Average0.025433160.20.463.29.72.01.09.114.3

This exercise seems to confirm what RegioX had argued a few days ago in his own analysis of the matter: G25 doesn't seem to be getting very incorrect results at all when you happen to use the correct combination of samples and interpret them correctly (that is, not literally and with some basic common sense and deduction skill).

By the way, using these exercises I could definitely confirm: the authors are really including all Neolithic Anatolian samples, including those that already had CHG/Iran and Levant_N, in their analyses, so any minor propotion of CHG/Iran and Levant_N will either disappear or become nearly negligible when you model them, especially in more parsimonious 3-way or 4-way models, without several possible source populations. Most of the discrepancy between G25 and the models in published genetic papers virtually disappears when you use Anatolia_N as a sum of ALL the individuals of Neolithic Anatolia.

That finally explains perfectly well why the authors do not identify much or even minimally significant Levant_N in places like Sicily and Greek islands: some Levant_N-shifted individual samples of the Anatolia_N pool of samples is already taking whatever little proportions of it there may be. Models using the entire pool of Anatolia_N samples under the label Anatolia_N will always hide any minor Levant_N and CHG/Iran_N and will show them only if they're in excess of what already existed in some part of Neolithic Anatolia.
 
Jovialis said:
@Ygorcs, I don't take what is published as gospel, which is evident by my position to the conclusions made by Antonio m. et al 2019. I think it makes a lot of naïve and broad assumpstions, with a very limited sample set. Moreover, in regards to Fernandes et al 2020, I think the fact that Morocco_LN looks very Mycenaean-like, makes their conclusion extremely tenuous. The main issue I had was with G25, and the components constructed from averages, which I don't agree is a proper way to analyze the publicly available samples. I am all for hobbyists like us examining samples, which is why me and Duarte made so many coordinates for Dodecad on Vahaduo. Though I recognize that it too has some flaws, due to the fact it uses modern components, which is not adequate for some ancient samples, particularly those that fall outside of modern genetic continuums.
Click to expand...

Then we both in fact agree, because I don't fancy the use of those artificially created population averages at all and prefer using individual samples and, if necessary, combining them all to form the proportion of ancestry coming from a given population. I repeatedly told you in this thread that's what I do, instead of using average components, so I never understood why you were disagreeing so strongly with what I was doing.

As for your description of some models and conclusions made by professional geneticists in their studies being too "naive and broad", I completely concur. I would also describe some of them, especially when involving more or less recently related populations (e.g. Morocco_LN vs. EEF-derived samples samples), as too simplistic. By saying that they can model modern Sicilians as simply Anatolia_N + Morocco_LN + Iran_Ganj_Dareh + Yamnaya_Samara, they are simplifying things so much that the software will look for the closest related source sample that has those other admixtures that are lacking in the others, among the 4 samples included. That enormously inflates the proportions of that sample, which wouldn't happen at all in a more complex model based on the also much more complex demographic history of populations since the Neolithic.

I'm honestly at a loss why in that model Punic_Ibiza looks Mycenaea-like even if they. Maybe another too broad modelling, because it's hard to believe that a population sample with significant (and very divergent in the West Eurasian context) Taforalt ancestry (according to the very analysis of the authors that published it, as well as my own amateurish analysis) is reeeeally very similar to Mycenaeans and Minoans. I for one don't buy it. There must be some unexplained problem there.
 
Ygorcs in regards to Post 591, my comment about childish is that you most likely new well Morocco_LN is very, very, little indigenous berber. The other ancestries in Morrocco_LN are way, way, East, Anatolian Neolithic and Iran_Neo and if there is some residual Levant_Neolithic as part of some Anatolian_Neolithic EEF, then that to would be East. Jovialis made that point as well. It is a a rhetorical question but if in fact Morrocco_LN was 100% Berber or ancient North African from the Maghreb, Sicily would plot way, way, way, further West than it does, don't you at least agree with that.

And to repeat myself, I have never come here and not fully disclosed what Ancestry and NAT GENO results said, 3% Levant/Middle East in Ancestry and 8% Asia Minor in NAT GENO (which is Northern Levant to Armenia. So based on these 2 DNA tests, 3 to 8% is Levant, if all of the 8% in Asia-Minor in NAT GENO is Levant related. As an FYI, I haven't gotten any Berber ancestry that has shown up so far, but doesn't mean some ancestry from North Africa via Phoenicians/Carthaginians isn't part of my ancestry, could be, could not be? I have no evidence of it and it could be that some of the Levant admixture is via Phoenician/Carthage in ancient Tunisia, with no local Berber element? I don't know and to be honest, I really don't care how the whatever the amount of Levant admixture in me is and how it got there. As Doris Day sang in Hitchcock's The Man who new To Much (2nd version of it) "Que Sera, Sera (Whatever Will Be, Will Be)"

As for your G25 analysis, I have stated to you that that if the G25 is a valid model, and I conceded after your work and Regio_X's, it does look like a valid model. However, when results get crazy it is sample selection and models chosen/variable selection that is the problem, which is the essence of what your first paragraph is about.
 
Palermo Trapani said:
Ygorcs in regards to Post 591, my comment about childish is that you most likely new well Morocco_LN is very, very, little indigenous berber. The other ancestries in Morrocco_LN are way, way, East, Anatolian Neolithic and Iran_Neo and if there is some residual Levant_Neolithic as part of some Anatolian_Neolithic EEF, then that to would be East. Jovialis made that point as well. It is a a rhetorical question but if in fact Morrocco_LN was 100% Berber or ancient North African from the Maghreb, Sicily would plot way, way, way, further West than it does, don't you at least agree with that.
Click to expand...

No, I really didn't. Actually, I still have strong doubts about that after reading several statements about KEB samples in the study itself and analyzed the pictures in it. I think it does look very much like a slightly more European-shifted Berber (Northern/Mediterranean Berber, of course), with too much Taforalt ancestry to be really Mycenaean-like (EEF+Steppe+CHG/Iran) as it only appears to be in some broad admixture model. Morocco_LN was basically pure ANF + Taforalt. There is a lot of shared ancestry on the ANF part, but already quite drifted, I believe, and the rest involves extremely divergent ancestral admixtures.

I absolutely agree that if would plot very differently, so that's precisely the reason I was ironic in that I actually think this study is another evidence that models are sometimes too vague, too simplistic, too broad (as Jovialis pointed out just now) and because of that they end up getting wrong or at least on the surface very misleading results.
 

Great answer, I can tell you as a European American trying to claim you are a Native American would be cause for a great uproar - its not true. This should not be happening in Europe. No doubt these people have an agenda. No cultural or racial identity for Europeans - from Sicily to Ireland, Sweden to Russia. This includes also people of European decent all over the world. Would it be safe to say if Europe can’t be European then Japan can’t be Japanese, Korea can’t be Korean and Israel can’t be Jewish? I’m sure those people wouldn’t take that lying down.
 
Palermo Trapani said:
As for your G25 analysis, I have stated to you that that if the G25 is a valid model, and I conceded after your work and Regio_X's, it does look like a valid model. However, when results get crazy it is sample selection and models chosen/variable selection that is the problem, which is the essence of what your first paragraph is about.
Click to expand...

Agreed. We can only examine the samples we have, preferably using the best samples (for instance, avoiding using samples that are not representative or are just too mixed between themselves that the results will be a lot less reliable). But I maintain that I see nothing "crazy" in my G25 models of South Italian samples. They may not be representative of the whole population, particularly because they're so few, but still they are there and they exist and are part of the population.

Now that I'm finally proving that Anatolia_N is indeed a pool of many Neolithic Anatolian samples (actually the Fernandes et al. supplement explicitly says that, though they don't specify which individual samples they're using and from where they are), I'm getting results that do not look that different from what previous studies have asserted, with the caveat that doing so will hide a bit of the CHG/Iran and Levant_N that may be there not because it arrived with ANF farmers, but due to later admixtures blended into a more unmixed Barcin-N population.

Levant_N is still there in my models (in lower proportions and fewer individuals, of course), but in all honesty I don't really think most geneticists were interested in making more complicated ancestry models with less profoundly divergent samples like Anatolia_N and Levant_N, particularly if they can clearly see that the proportions of the latter are so minor that they could be basically explained away by Anatolia_N and focus the model more on very divergent population movements, like Iran_N, Taforalt-related and Pontic-Caspian Seppe ones. There's also the issue of the aim of the study: they won't care about modelling a more convoluted demographic history involving a few percents of Levantine or North African ancestry when and if the aim of their study is to track Indo-European migrations or understand when Iran/CHG arrived in a certain part of the world (just some examples). Ultimately what I want to say is: the models aren't talking for themselves, they are also being modelled according to what the authors want to detect and to explain.

Let me show you one example just with Abruzzo and Sicilian Italians:

TargetDistanceGEO_CHGIRN_Ganj_Dareh_NITA_Grotta_Continenza_MesoLevant_PPNBMAR_ENRUS_Karelia_HGRUS_Khvalynsk_EnTUR_Barcin_NTUR_Boncuklu_NTUR_Kumtepe_NTUR_Tepecik_Ciftlik_NWHG
Italian_Abruzzo:Alp0900.025743033.610.20.09.20.00.018.650.60.00.05.82.0
Italian_Abruzzo:Alp1400.027821624.25.80.05.60.00.021.641.40.01.617.82.0
Italian_Abruzzo:ALP1610.023109296.07.40.01.20.00.021.254.60.00.08.61.0
Italian_Abruzzo:Alp1620.018467100.010.20.00.80.00.021.232.20.00.032.82.8
Italian_Abruzzo:ALP2050.016340551.68.00.04.60.00.017.620.40.00.047.40.4
Italian_Abruzzo:Alp3800.016720060.06.81.22.60.00.021.022.00.05.041.40.0
Italian_Abruzzo:Alp5030.027987655.46.40.05.60.00.021.653.40.00.07.60.0
Italian_Abruzzo:Alp6160.024069865.08.60.02.20.40.019.053.80.00.011.00.0
Italian_Abruzzo:ItalyAbruzzo130.028150550.010.60.07.20.00.020.253.20.00.05.83.0
Italian_Abruzzo:ItalyAbruzzo140.020848261.07.00.02.00.00.022.236.80.00.031.00.0
Italian_Abruzzo:ItalyAbruzzo150.018904000.88.60.00.00.00.022.641.60.00.024.81.6
Italian_Abruzzo:ItalyAbruzzo160.018951650.45.00.00.00.00.021.419.80.01.049.82.6
Italian_Abruzzo:ItalyAbruzzo170.017328163.84.40.05.00.00.025.651.40.00.09.40.4
Italian_Abruzzo:ItalyAbruzzo190.019910653.26.60.00.00.00.021.426.00.00.041.61.2
Italian_Abruzzo:ItalyAbruzzo200.019741456.28.00.00.00.00.021.851.00.00.012.20.8
Italian_Abruzzo:ItalyAbruzzo210.027322783.63.60.01.00.80.020.430.40.00.036.63.6
Italian_Abruzzo:ItalyAbruzzo220.017321844.03.20.00.00.00.022.637.40.00.031.61.2
Italian_Abruzzo:ItalyAbruzzo230.026364232.65.40.00.00.05.617.630.20.00.038.60.0
Italian_Abruzzo:ItalyAbruzzo90.028222793.67.00.010.60.00.024.040.00.00.014.60.2
Sicilian_East:EastSicilian2H0.019607215.83.80.04.01.00.017.433.20.08.425.21.2
Sicilian_East:EastSicilian5H0.025409310.08.40.011.20.40.021.636.80.00.020.01.6
Sicilian_East:EastSicilian8H0.018392505.28.80.08.20.80.012.614.40.011.234.44.4
Sicilian_West:WestSicilian10H0.021701910.010.60.02.80.00.017.223.40.011.828.65.6
Sicilian_West:WestSicilian4H0.018928330.610.00.00.84.00.013.813.42.64.247.43.2
Sicilian_West:WestSicilian7H0.022914380.08.60.03.22.80.017.632.40.00.030.25.2
Average0.022011172.77.30.03.50.40.220.136.00.11.726.21.8

In other words, clustering the samples together as studies often do: 64.0% ANATOLIA_N (Tepecik-Ciftlik + Kumtepe included) + 10.0% CHG/IRAN + 3.5% LEVANT_N + 20.3% STEPPE + 1.8% WHG + 0.4% MOROCCO_EN (TAFORALT-LIKE)
Does that look really crazy, extremely implausible considering the entire history of the region? I don't think so.
 
Palermo Trapani said:
Ygorcs in regards to Post 591, my comment about childish is that you most likely new well Morocco_LN is very, very, little indigenous berber. The other ancestries in Morrocco_LN are way, way, East, Anatolian Neolithic and Iran_Neo and if there is some residual Levant_Neolithic as part of some Anatolian_Neolithic EEF, then that to would be East. Jovialis made that point as well. It is a a rhetorical question but if in fact Morrocco_LN was 100% Berber or ancient North African from the Maghreb, Sicily would plot way, way, way, further West than it does, don't you at least agree with that.

And to repeat myself, I have never come here and not fully disclosed what Ancestry and NAT GENO results said, 3% Levant/Middle East in Ancestry and 8% Asia Minor in NAT GENO (which is Northern Levant to Armenia. So based on these 2 DNA tests, 3 to 8% is Levant, if all of the 8% in Asia-Minor in NAT GENO is Levant related. As an FYI, I haven't gotten any Berber ancestry that has shown up so far, but doesn't mean some ancestry from North Africa via Phoenicians/Carthaginians isn't part of my ancestry, could be, could not be? I have no evidence of it and it could be that some of the Levant admixture is via Phoenician/Carthage in ancient Tunisia, with no local Berber element? I don't know and to be honest, I really don't care how the whatever the amount of Levant admixture in me is and how it got there. As Doris Day sang in Hitchcock's The Man who new To Much (2nd version of it) "Que Sera, Sera (Whatever Will Be, Will Be)"

As for your G25 analysis, I have stated to you that that if the G25 is a valid model, and I conceded after your work and Regio_X's, it does look like a valid model. However, when results get crazy it is sample selection and models chosen/variable selection that is the problem, which is the essence of what your first paragraph is about.
Click to expand...

(CONTINUING) And, since Levant_N is already itself ~50% ANF, most geneticists, unless they are specifically studying the history of the expansion of Levantine farmers, won't care about those minor proportions that might even be explained by a Tepecik-Ciftlik population with a bit more Levant_N than the few samples extracted so far had. Since we're not talking about a very divergent population contributing ancestry to Sicilians and South Italians, the models don't even change much if we simply ignore this distinction between Anatolia_N and Levant_N:

TargetDistanceGEO_CHGIRN_Ganj_Dareh_NITA_Grotta_Continenza_MesoMAR_ENMAR_LNRUS_Karelia_HGRUS_Khvalynsk_EnTUR_Barcin_NTUR_Boncuklu_NTUR_Kumtepe_NTUR_Tepecik_Ciftlik_NWHG
Italian_Abruzzo:Alp0900.027067870.011.60.00.80.00.018.640.80.00.026.22.0
Italian_Abruzzo:Alp1400.028229012.46.20.01.00.00.021.436.60.00.030.22.2
Italian_Abruzzo:ALP1610.023128445.67.60.00.00.00.021.254.40.00.010.21.0
Italian_Abruzzo:Alp1620.018478580.010.20.00.00.00.021.031.40.00.034.82.6
Italian_Abruzzo:ALP2050.016511130.67.60.00.40.00.017.416.80.00.057.00.2
Italian_Abruzzo:Alp3800.016834960.07.01.40.60.00.020.220.00.04.646.20.0
Italian_Abruzzo:Alp5030.028332012.87.00.00.00.00.022.043.80.00.024.40.0
Italian_Abruzzo:Alp6160.024117944.68.40.00.80.00.018.852.40.00.015.00.0
Italian_Abruzzo:ItalyAbruzzo130.028737960.011.20.00.00.00.019.052.80.00.014.42.6
Italian_Abruzzo:ItalyAbruzzo140.020907170.07.20.00.00.00.022.433.00.00.037.40.0
Italian_Abruzzo:ItalyAbruzzo150.018899980.68.60.00.00.00.022.841.60.00.025.01.4
Italian_Abruzzo:ItalyAbruzzo160.018955080.44.80.00.20.00.021.620.40.00.249.82.6
Italian_Abruzzo:ItalyAbruzzo170.017717782.84.80.00.02.40.025.047.40.00.017.20.4
Italian_Abruzzo:ItalyAbruzzo190.019910653.26.60.00.00.00.021.426.00.00.041.61.2
Italian_Abruzzo:ItalyAbruzzo200.019742086.08.20.00.00.00.021.851.00.00.012.20.8
Italian_Abruzzo:ItalyAbruzzo210.027322703.23.60.01.00.00.020.429.00.00.039.23.6
Italian_Abruzzo:ItalyAbruzzo220.017321844.03.20.00.00.00.022.637.40.00.031.61.2
Italian_Abruzzo:ItalyAbruzzo230.026364232.65.40.00.00.05.617.630.20.00.038.60.0
Italian_Abruzzo:ItalyAbruzzo90.029574230.29.20.01.60.00.023.635.40.00.029.60.4
Sicilian_East:EastSicilian2H0.019648844.84.00.01.60.20.017.632.80.03.834.01.2
Sicilian_East:EastSicilian5H0.026959110.07.40.02.40.01.818.426.00.00.042.21.8
Sicilian_East:EastSicilian8H0.018816383.28.60.02.20.00.013.213.00.05.050.64.2
Sicilian_West:WestSicilian10H0.021839020.09.80.00.60.00.816.421.20.011.434.45.4
Sicilian_West:WestSicilian4H0.018927010.210.20.04.20.00.014.013.23.22.849.23.0
Sicilian_West:WestSicilian7H0.023017740.08.00.03.60.00.017.229.40.00.036.45.4
Average0.022294471.97.50.10.80.10.319.833.40.11.133.11.7

The fits get extremely little worse than with Levant_PPNB. Why? 1) very low proportions won't make a lot of difference; 2) very low proportions of a population group ALREADY quite related to another population group included in the pool of source samples will make even less difference. But I still find that extra minor Levant_PPNB quite realistic, and see that it its exactly within the range of Levantine affinity that you yourself found using other tools. ;)
 
Ygorcs said:
No, I really didn't. Actually, I still have strong doubts about that after reading several statements about KEB samples in the study itself and analyzed the pictures in it. I think it does look very much like a slightly more European-shifted Berber (Northern/Mediterranean Berber, of course), with too much Taforalt ancestry to be really Mycenaean-like (EEF+Steppe+CHG/Iran) as it only appears to be in some broad admixture model. Morocco_LN was basically pure ANF + Taforalt. There is a lot of shared ancestry on the ANF part, but already quite drifted, I believe, and the rest involves extremely divergent ancestral admixtures.

I absolutely agree that if would plot very differently, so that's precisely the reason I was ironic in that I actually think this study is another evidence that models are sometimes too vague, too simplistic, too broad (as Jovialis pointed out just now) and because of that they end up getting wrong or at least on the surface very misleading results.
Click to expand...

Ok, I just assumed you new that Morroco_LN was a European_shifted Berber as you put it. Thus, I will take you at your word and retract my childish statement.

At this point, I am not going to get into how he or she got to where they are admixture wise, and what ANF in terms of source population was. Maybe rather than Mycenaean like, maybe more Minoan like, EEF+CHG+Iran, no Steppe. But the Ibiza_Pheonician who was way over in Spain still seemed to plot back further East as well, or at least, not in the Maghreb, isn't that correct?
 
Palermo Trapani said:
bQQdpnc.jpg
Click to expand...

Palermo, now I understood what's wrong with your (not just yours) interpretation of this picture. It's clear what the issue is: they didn't use enough K= to be able to identify and distinguish Morocco_EN on itself (Morocco_EN was 100% Taforalt-like, though of course very drifted after millennia, with no signal of extra admixture). So, the model accoridngly won't pick the entire amount of Taforalt-like admixture in Morocco_LN and any other sample with Taforalt-like or perhaps even Natufian-like ancestry.

Notice what happens to the Morocco_EN sample: since it is VERY divergent from all the others, the K= doesn't work for it - it has a unique blue component (which is probably an indication of the Hadza-related ANA ancestry), but I'd say half or more of it LOOKS LIKE (but in fact IS NOT) derived from the main populations ancestral to Anatolia_N and Iran_N. That makes no sense, and I don't think the authors of the study were particularly interested in that sampple and others not directly related to their paper, either.

Therefore, in the Morocco_LN what you're actually seeing is a MUCH higher percentage of Morocco_EN admixture than the model is showing, because a whole lot of that admixture will appear as darker blue + some pink + a bit of yellow. It'll appear as Iran_N-like and ANF_like. That's just an artifact of the model, which does not distinguish Morocco_EN/Taforalt ancestry properly. If half or less than half of Morocco_EN is light blue, then you should double or slightly more than double the proportion of that light blue component in Morocco_LN to get close to the real genetic input from Morocco_EN-like individuals.

I'd suggest reading the study on those Moroccan samples again, they make it clear that Morocco_LN samples are INTERMEDIATE between EEF and Morocco_EN, therefore pretty much North African Berber-like (not including those modern groups more admixed with Arabs and Subsaharan Africans).

Finally, I wanted to say that I think it's clear from my thousands of messages and my attitudes here that I'm not that childish nor that fake smartass as I've been accused of, even though I'm sure no infractions will ever be given to the users who use such impolite ad hominem attacks when they think they should react to a perceived "passive-aggressive" offense (but never name-calling).
 
Last edited:
Palermo Trapani said:
Ok, I just assumed you new that Morroco_LN was a European_shifted Berber as you put it. Thus, I will take you at your word and retract my childish statement.

At this point, I am not going to get into how he or she got to where they are admixture wise, and what ANF in terms of source population was. Maybe rather than Mycenaean like, maybe more Minoan like, EEF+CHG+Iran, no Steppe. But the Ibiza_Pheonician who was way over in Spain still seemed to plot back further East as well, or at least, not in the Maghreb, isn't that correct?
Click to expand...

Okay, no problem, Palermo. Thanks.

Indeed, and some even said that Ibiza_Phoenician was actually not Phoenician, Punic or whatever at all, but Greek or at least Aegean admixed with Phoenicians (or more broadly Levantines) and Punics (Berber-Phoenician). I'm not sure, I'd say only European + Levantine + North African. But I honestly don't understand how they could model it as mostly Morocco_LN. It's very genetically distant from Morocco_LN (>12.0), while it's much closer to Mycenaean (~5.0-6.0). Considering the genetic description of Morocco_LN samples as basically Taforalt-shifted ANF in the study they were published, I find it hard to believe that would somehow create a population very similar to a people with ANF+CHG/Iran+Steppe like the Mycenaeans. Totally different additional components added to ANF.

With some Steppe + WHG ancestry, I doubt very much that the Ibiza_Phoenician is really mostly derived from mostly North African people, even if more ANF-shifted like Morocco_LN. I can only guess the model simply saw the more divergent of all those admixtures (Taforalt/Morocco_EN) and since it had to "explain" it it took all the ANF and Levant_N and attributed to Morocco_LN, giving a totally misleading result. Compare:
TargetDistanceGEO_CHGIRN_Ganj_Dareh_NLevant_PPNBMAR_ENRUS_Khvalynsk_EnTUR_Barcin_NTUR_Boncuklu_NTUR_Kumtepe_NTUR_Tepecik_Ciftlik_NWHG
MAR_LN:KEB.40.032275760.00.08.429.40.059.60.00.00.02.6
Iberia_Northeast_Empuries2:I82080.024323588.04.80.00.08.255.80.00.023.20.0
Iberia_Northeast_Empuries2:I82150.022360945.24.00.00.013.662.40.00.014.80.0
GRC_Mycenaean:I90060.028561668.67.00.00.05.465.40.013.60.00.0
GRC_Mycenaean:I90100.025829712.29.86.60.03.245.622.60.09.20.8
GRC_Mycenaean:I90330.0316255412.65.80.03.410.261.42.82.80.01.0
GRC_Mycenaean:I90410.022724488.05.43.20.010.671.80.00.01.00.0
Iberia_Ibiza_Punic:MS106140.024209530.010.20.48.89.446.40.016.20.08.6


EDIT: From the original paper that published the Ibiza_Phoenician sample - in the figures it's shown that it plots closest to Levant_N (which does not mean it is Levant_N, of course, plotting close may be because of a specific combination of many admixtures in certain proportions that in the aggregate end up positioning the sample right on top of another in fact quite different sample).

"When considering more fine-grained analysis including only the Levantine and Iberian modern data with MS10614, the Ibizan positioned between the Iberian and Levantine samples, with more affinity to the Levantine populations (Fig. 6a). We conducted a further test, adding North African populations48 to the analyses (Fig. 6b) and see that MS10614 is situated more closely to the Levantine and Iberian samples and not the North Africans."

"The model k = 3 shows a new component (dark blue) that impacts all of the Eastern Mediterranean populations and the modern Iberians. Sample MS10614 has an intermediate amount of this component, with more than modern Spanish, but less than modern Lebanese and ancient Levantines."

"As can be seen in Fig. 6a,b, the Ibizan Phoenician sample, MS10614, plots in between modern Levantine and Iberian populations but closer to both of these than to modern North African populations. In Fig. 6c, it is positioned most closely to a Levantine Neolithic sample, but in between a Sidon Bronze Age and European/Iberian Bronze Age samples. The ADMIXTURE result (k = 2) in Fig. 7 also indicates that the genetic ancestry of this individual was intermediate between an Iberian Bronze Age and Levantine Bronze Age individuals."

I find it hard to believe Ibiza_Phoenician was indeed mostly Morocco_LN-like and Morocco_LN was on its turn Mycenaean-like. That would ultimately mean that Mycenaeans were like Levant-shifted Europe_BA. Those apparent results must be just a misinterpretation of insufficiently comprehensive models that are missing some key ancestral components.
 
You must log in or register to reply here.

This thread has been viewed 190594 times.

Back
Top