Genetic and Cultural Differences between Jews and Greeks

Regio_X: Nice analysis in post #275. I respect the diplomacy and caution with your statement "possibly overstated by G25" but I think a 2.3/2.4 to 3.3 time magnitude in difference in admixture is quite large with respect to Levant_PNNB. Especially given the views in some of the post above, or that have been made in the past, that there was some massive migration from the Levant post "fill in your period" that supposedly shifted Italian populations from modern Central Italy as far as Lazio and Abruzzo to Sicily significantly from what Rome was during the Iron Age and Republican Rome period. The Antonio et al 2019 paper and the 11 Iron Age_Republican Rome samples already had 3 of the 11 that were Southern Italian shifted. There is no way people here at Eupedia whose ancestors are all 100% from regions in Modern Italy south of Lazio on the Southern peninsula (Campania, Puglia, Calabria, Basilicata) and yes, sorry to disappoint some, Sicily as well, are that close to Iron Age Romans unless there were Iron Age Romans that were genetically no different that modern Southern Italians.

My distances in post 137 and 274 are not statistical anomalies. Here are my distances vs. Modern Italian populations. I mean if you look at my distances, and just go back and look at Parolo et al 2016 Figure 1, Sazzini et al 2016 Figure 2, and Raveane et al 2019 Figures 1 and 2. My results vs all Modern Italian populations mirror the plots in those papers. A basic pattern of Clustering, 1)Southern Italian/Sicily cluster with some Central regions drifting South, 2) A Central Italian Cluster, 3) a Northern Italian Cluster with Tuscany drifting North and 4) Sardinia.

Distance to:PalermoTrapani
3.56228943Italy_Campania
3.83059103Italy_Abruzzo
4.03985458Italy_Sicily
5.85659507Italy_Calabria
6.30021809Italy_Apulia
7.04316676Italy_Marche
7.65791512Italy_Lazio
10.45749014Italy_Romagna
13.19014973Italy_Tuscany
14.95868467Italy_Emilia
15.59318961Italy_Liguria
18.60377267Italy_Lombardy
18.76349355Italy_Piedmont
18.91861555Italy_Veneto
20.12508375Italy_FriuliVG
22.89652812Italy_Trentino
25.68679815Italy_Aosta_Valley
44.65843593Sardinian

Target: PalermoTrapani
Distance: 1.7225% / 1.72250811 | ADC: 0.25x

56.4Italy_Sicily
24.2Italy_Campania
15.2Italy_Marche
4.2Sardinian
 
Don't you guys find it a bit weird that, as per Lazaridis' qpAdm model, Isparta_EBA had only 6% Levantine Neolithic ancestry, but it is unmistakably shifted eastward towards the Levant? I mean, if Levant_N was extremely divergent from Anatolia_N, any even reasonably small percentage of that admixture would really cause a quite significant shift, but that usually doesn't happen in any non-negligible way when the additional admixture is from a not so distant source, not in very low proportions.


This is what I'm trying to say:


https://imgur.com/a/BGTB4Ox
Well, as Moesan once said, there're PCAs and PCAs.:) The same authors claim the qpAdm estimates corespond to the PCA.

"We test the robustness of the qpAdm estimates by plotting populations with the inferred mixture proportions (Table S2.26) in the weighted average position of their source populations in PCA space. The results (Extended Data Fig. 5) indicate a close correspondence between the two in all models considered."

They also say SW AN is shifted to the East, toward Levant, but that compared to Minoans and Mycenaeans.

Here's Figure 5:
https://www.nature.com/articles/nature23310/figures/7
 
I've tried something similar, using other ANFs (don't remember if all), with the difference I grouped them in only one cluster (TUR_N or something). Since the results were similar to those using Barcin only, I used just this one (Barcin) to make it easier to reproduce. But seeing the numbers you posted, it looks like keeping them divided is better for that purpose. Thank you!
As for the other issue, well, that paper on Epipaleo Anatolian showed that ANF, while mostly AHG (~80%?), already had a bit of Iran Neo and Early Holocene Levantines, and Levant Neo itself had AHG or AAF*, so it can help to explain perhaps? Also, AHG could be modeled as half WHG-like and half Natufian-like:
https://www.biorxiv.org/content/biorxiv/early/2018/09/20/422295.full.pdf
*"In turn, Levantine early farmers (Levant_Neol) that are temporally intermediate between AAF and ACF could be modeled as a two-way mixture of Natufians and AHG or AAF (18.2 ± 6.4 % AHG or 21.3 ± 6.3 % AAF ancestry; tables S4 and S8 and data table S4), confirming previous reports of an
Anatolian-like ancestry contributing to the Levantine Neolithic gene pool."

I know about that model, but I found it really strange though it's of course clear AHG was intermediate between WHG and Natufian, but satisfyingly modeled as half WHG + half Natufian? Did AHG have so much less Basal Eurasian than Natufians (considering WHG lacked it), and conversely did they also have that much Taforalt and therefore ANA ancestry? I have never read any other study claim that. Besides WHG though remotely related to the Common West Eurasian HG that also originated the non-BE portion of AHG, it seems to be very divergent from it, with no recent commonality. I confess I found that claim quite unconvincing at least until I see other studies confirming it.
 
Well, as Moesan once said, there're PCAs and PCAs.:) The same authors claim the qpAdm estimates corespond to the PCA.
"We test the robustness of the qpAdm estimates by plotting populations with the inferred mixture proportions (Table S2.26) in the weighted average position of their source populations in PCA space. The results (Extended Data Fig. 5) indicate a close correspondence between the two in all models considered."
They also say SW AN is shifted to the East, toward Levant, but that compared to Minoans and Mycenaeans.
Here's Figure 5:
https://www.nature.com/articles/nature23310/figures/7

Did they test that comparison between hypothetical weighted PCA position and actual PCA plot only for Mycenaeans? What a shame... :-/

Anyway, I don't doubt amateur calculators may skew results underestimating or overestimating them, but it seems in this case the models by professional geneticists also prove the main point: the signal is there. More, less, it ultimately doesn't matter as much that it is there in sufficient proportions to cause the population samples to drift eastward toward the Levant.
 
I've tried something similar, using other ANFs (don't remember if all), with the difference I grouped them in only one cluster (TUR_N or something). Since the results were similar to those using Barcin only, I used just this one (Barcin) to make it easier to reproduce. But seeing the numbers you posted, it looks like keeping them divided is better for that purpose. Thank you!
As for the other issue, well, that paper on Epipaleo Anatolian showed that ANF, while mostly AHG (~80%?), already had a bit of Iran Neo and Early Holocene Levantines, and Levant Neo itself had AHG or AAF*, so it can help to explain perhaps? Also, AHG could be modeled as half WHG-like and half Natufian-like:
https://www.biorxiv.org/content/biorxiv/early/2018/09/20/422295.full.pdf
*"In turn, Levantine early farmers (Levant_Neol) that are temporally intermediate between AAF and ACF could be modeled as a two-way mixture of Natufians and AHG or AAF (18.2 ± 6.4 % AHG or 21.3 ± 6.3 % AAF ancestry; tables S4 and S8 and data table S4), confirming previous reports of an
Anatolian-like ancestry contributing to the Levantine Neolithic gene pool."
Really complicated. It's like Taforalt x Natufian x SSA. An older paper stated Taforalt had Natufian and SSA, but an updated one supported the opposite: Taforalt contributing to both Natufian and SSA.

We'll likely learn much more on this subject in the future, with even better tools and morr ancient DNAs. :)

@Palermo
Regarding G25, see Ygorcs post. Using other Anatolian Neo as sources, the Levant Neo drops even more. So it's possible G25 is not that off. We can get substantially different results depending on how we model.

I confess I haven't followed all the thread due to lack of time. It's growing fast, so I did just "speed reading". I'll try to read it more carefully later, in order to understand better what's being discussed.

@Ygorcs
Ops. Always hurry, I didn't realize it was just for Mycenaeans. Sorry!
In my quoting I assumed they did for others. My bad. Thanks for pointing it out.
 
I checked the supplement of Raveane et al 2019, they only use I2683_Anatolia_BA for ABA. Thus it actually has the least amount of Levantine admixture among them. Moreover, If G25 is inflating it, which it seems so, then it truly negligible. Especially, by the time it mixed with other source populations in Italy.

Then it totally makes more sense if the Anatolia_N models are using takes into consideration the internal genetic structure of Neolithic Anatolia: Kumtepe and Tepecik-Ciftlik already had much more Natufian and CHG than Bonkuclu and especially Barcin even as early as the Neolithic. You get really low percentages of Levant_N for South Italy, Sicily and even Malta if you model them using those other Anatolia_N samples apart from Barcin, but the strange thing is that they also "eat" most of the CHG (probably because it was already found in Tepecik-Ciftlik). But this hypothetical ancestral makeup would require something that I think is quite improbable: that there were two clearly different (genetically) waves of Neolithic expansion into Europe, one very Barcin-like and the other much more Tepecik-Ciftlik-like even back in the Neolithic. How likely is that?

TargetDistance | ADC: 0.25xGEO_CHGIRN_Ganj_Dareh_NLevant_PPNBMAR_ENTUR_Barcin_NTUR_Boncuklu_NTUR_Kumtepe_NTUR_Kumtepe_N_low_resTUR_Tepecik_Ciftlik_NWHGYamnaya_RUS_Kalmykia
Italian_Abruzzo0.016736730.00.00.00.01.47.89.60.053.20.028.0
Italian_Aosta_Valley0.029097770.00.00.00.019.035.80.00.02.25.038.0
Italian_Apulia0.017003250.00.60.00.00.02.216.00.055.40.025.8
Italian_Basilicata0.015769800.01.80.00.01.04.411.00.056.60.025.2
Italian_Bergamo0.023992350.00.00.00.08.643.60.00.013.02.832.0
Italian_Calabria0.017125460.02.80.40.80.02.619.61.651.20.021.0
Italian_Campania0.017259350.02.82.00.00.64.014.40.053.20.023.0
Italian_Jew0.014629700.05.417.00.40.00.017.60.041.20.018.4
Italian_Lazio0.019164550.00.00.00.05.415.40.60.050.60.627.4
Italian_Liguria0.024660380.00.00.00.04.033.06.22.818.63.032.4
Italian_Lombardy0.024223400.00.00.00.00.049.80.00.016.61.432.2
Italian_Marche0.017152610.00.00.00.03.814.48.60.042.80.230.2
Italian_Molise0.015102360.00.00.00.00.011.62.80.057.20.028.4
Italian_Northeast0.023689510.00.00.00.011.242.60.00.04.23.238.8
Italian_Piedmont0.019836900.00.00.00.04.835.80.00.025.01.233.2
Italian_Trentino-Alto-Adige0.021256840.00.00.00.014.440.80.00.06.24.234.4
Italian_Tuscany0.017496310.00.00.00.00.231.89.20.027.00.831.0
Italian_Umbria0.017174020.00.00.00.00.023.89.20.037.00.030.0
Italian_Veneto0.022119770.00.00.00.010.041.60.00.010.43.234.8
Sicilian_East0.017533900.00.65.20.40.00.025.82.042.20.223.6
Sicilian_West0.018955940.00.41.02.80.02.015.40.052.42.423.6
Maltese0.016373740.00.26.23.60.01.626.61.237.00.023.6
Greek_Central_Anatolia0.020185209.611.00.00.00.00.00.01.070.00.08.4
Greek_Central_Macedonia0.022415670.00.00.00.00.227.60.00.037.80.234.2
Greek_Crete0.016257790.06.80.00.00.00.00.00.072.20.021.0
Greek_Izmir0.017646610.00.40.00.00.09.60.40.061.40.028.2
Greek_Kos0.017660850.08.60.00.00.80.80.20.072.80.016.8
Greek_Peloponnese0.018883260.00.00.00.01.216.47.20.047.40.027.8
Greek_Thessaly0.018909020.00.00.00.02.222.49.20.033.20.033.0
Greek_Trabzon0.0229159628.09.60.00.00.00.00.03.059.40.00.0
Average0.019374301.31.71.10.33.017.47.00.440.20.926.8
 
I think I also found why using Yamnaya will take away a lot of CHG in these models and become inflated. Steppe ancestry in most of Europe is not Yamnaya-like, but more EHG-shifted than Yamnaya was, more like a half-way between the most CHG-shifted Yamnaya samples and the Khvalynsk (who had much less CHG). So, using Yamnaya samples will forcefully cause the software to look for extra CHG to get a better fit. See what happens when you use Khvalynsk as a steppe proxy instead of Yamnaya:

TargetDistanceGEO_CHGIRN_Ganj_Dareh_NLevant_PPNBLevant_PPNCMAR_ENRUS_Khvalynsk_EnTUR_Barcin_NTUR_Kumtepe_NTUR_Kumtepe_N_low_resTUR_Tepecik_Ciftlik_NWHG
Italian_Abruzzo0.016834921.27.60.00.80.021.836.40.00.031.40.8
Italian_Aosta_Valley0.035124070.00.00.00.00.032.055.20.00.03.29.6
Italian_Apulia0.016020283.27.40.00.00.019.230.00.60.439.20.0
Italian_Basilicata0.015341383.48.20.02.40.019.233.00.00.033.80.0
Italian_Bergamo0.027233910.80.20.00.00.028.056.20.00.07.47.4
Italian_Calabria0.016052342.810.01.20.01.216.234.80.02.231.60.0
Italian_Campania0.015623242.29.80.04.80.018.034.80.60.828.80.2
Italian_Jew0.013350011.012.42.016.01.415.024.60.60.626.40.0
Italian_Lazio0.020166860.06.60.01.60.021.238.80.00.028.63.2
Italian_Liguria0.029334435.20.00.00.01.025.249.01.21.69.67.2
Italian_Lombardy0.026151970.80.00.00.00.028.452.80.00.011.66.4
Italian_Marche0.020847571.83.20.00.00.024.034.21.80.032.42.6
Italian_Molise0.016590202.26.40.00.00.022.234.40.00.034.40.4
Italian_Northeast0.028877942.00.00.00.00.031.449.60.00.010.46.6
Italian_Piedmont0.023818482.22.20.00.00.027.049.60.00.014.05.0
Italian_Trentino-Alto-Adige0.025675472.40.00.00.00.028.658.00.00.02.88.2
Italian_Tuscany0.021107251.82.40.00.00.025.444.66.60.015.04.2
Italian_Umbria0.019355992.25.00.00.00.023.845.61.00.019.82.6
Italian_Veneto0.026133891.80.40.00.00.028.855.00.00.06.67.4
Sicilian_East0.017820132.87.25.20.01.016.825.84.61.033.22.4
Sicilian_West0.017704680.010.01.40.02.416.223.62.00.039.84.6
Maltese0.014584050.07.80.05.24.417.625.69.22.826.01.4
Greek_Central_Anatolia0.0182289412.613.23.60.60.06.815.80.02.445.00.0
Greek_Central_Macedonia0.022902733.82.60.00.00.025.634.00.02.230.41.4
Greek_Crete0.015850704.411.00.02.80.015.218.60.00.647.40.0
Greek_Izmir0.016665375.06.60.60.00.019.836.00.02.030.00.0
Greek_Kos0.013586567.611.45.03.80.012.231.40.01.826.80.0
Greek_Peloponnese0.016985015.65.20.00.00.020.643.20.01.424.00.0
Greek_Thessaly0.021564384.43.00.00.00.026.042.64.20.218.21.4
Greek_Trabzon0.0215180329.411.25.60.00.00.88.60.02.442.00.0
Average0.020368363.85.70.81.30.421.137.41.10.725.02.8

 
Riverman: Let's Ignore the Imperial Roman Samples from Antonio et al 2019. I ran my model fits using just the 11 Iron Age Romans from Antonio et al 2019. I presented the model fits using 3 different calculators in post #137 (Dodecad 12B, MDLP16 and Eurogenes K13). The Dodecad 12B and Eurogenes K13 has the Roman samples there. Jovialis and the Eupedia Team produced the MDLP16 and I think also the Dodecad K7 coordinates for the 127 ancient Roman samples in the Antonio et al 2019 paper.

I am curios what your distances would be using the Iron Age samples only since you are so adamant about the Imperial Roman Samples and inferences being drawn from them, which I don't agree with.

If I don't use Imperial_Roman the unique signal gets lost in more mixed population averages. In individuals, like stated before, I can even work with the Latini sample, but then the supposedly "Imperial Roman citizens" from all over the Western empire need, as a rule, additional Greek, Anatolian, Levantine and individually North African, slight SSA. Just play around yourself, you get an "Imperial Roman" signal much beyond Italy and what do you think this is? Its primarly present in people which home area was a Roman province, both in Western and Eastern Europe. If you leave this sample out of a run, it just gets more complicated, because you then get especially the typical Greek, Levantine and North-Central European, secondary North African, Near Eastern and Subsaharan components like in many Imperial Romans of that time. So even if you don't accept Imperial_Roman being a useful sample, in Roman times the very components which it contains got spread, up to Britain and Belgium.

For Britain the Greek contribution gets, in my opinion, even inflated without using Imperial_Roman. Because for accounting for the 1-2 percent of Imperial Roman, even if using the England_Roman sample, you need a much higher percentage of Greek ancestry, for which we, by the way, have still no ideal samples for Roman times, like for Roman time Anatolians, Mesopotamians, Southern Italians for that matter, North Africans, Celts etc. Its not like we have all the ideal samples, so its obviously a good stand in.

Riverman: As an addendum to my Post #137, here is my model fit using the more parsimonious Dodecad K7. Just to reiterate, the 3 models in my post 137 and the one in this thread are only using the 11 Iron Age Roman samples to estimate the distances between Me and Iron Age Rome.

Target: PalermoTrapani
Distance: 0.5866% / 0.58658612 | ADC: 0.25x
95.4R437_Iron_Age_Palestrina_Selicata
3.0R475_Iron_Age_Civitavecchia
1.6R850_Iron_Age_Ardea


Your main component R437 with Imperial_Roman but without Greek (anachronistic I know):

62.0ITA_Rome_Imperial
34.8ITA_Rome_Latini_IA
3.2TUR_Alalakh_MLBA

Without Imperial_Roman, but with Greek samples:
25.0ITA_Rome_Latini_IA
17.6GRC_Minoan_Lassithi
17.0GRC_Mycenaean
14.2TUR_Alalakh_MLBA
14.0CZE_Early_Slav
12.2Levant_LBN_Roman
On the West Eurasian PCA this sample is closest, if looking at modern Italians, to a sample from Campania. In the European context to Campanians, Sicilians (East) and Maltese. Check:
https://vahaduo.github.io/g25views/#Europe1
I also checked for R475_Iron_Age_Civitavecchia from the Etruscan panel, this sample is North African shifted. It doesn't pick up Imperial_Roman, even if offered.

32.4Berber_Tunisia_Chen
31.6GRC_Minoan_Lassithi
27.2CZE_Hallstatt_Bylany
8.8ITA_Rome_Latini_IA

I checked both "Prenestini tribe" samples and its amazing how different they are. ITA_Prenestini_tribe_IA:RMPR435b is rather Northern-Western shifted and closest to Basque/Southern French on the PCA. These are really completely different people. Same for the ITA_Ardea_Latini_IA sample. So its not just Imperial_Roman being a wild mix, but the other Italic and Etruscan samples to a large degree as well.
 
Then it totally makes more sense if the Anatolia_N models are using takes into consideration the internal genetic structure of Neolithic Anatolia: Kumtepe and Tepecik-Ciftlik already had much more Natufian and CHG than Bonkuclu and especially Barcin even as early as the Neolithic. You get really low percentages of Levant_N for South Italy, Sicily and even Malta if you model them using those other Anatolia_N samples apart from Barcin, but the strange thing is that they also "eat" most of the CHG (probably because it was already found in Tepecik-Ciftlik). But this hypothetical ancestral makeup would require something that I think is quite improbable: that there were two clearly different (genetically) waves of Neolithic expansion into Europe, one very Barcin-like and the other much more Tepecik-Ciftlik-like even back in the Neolithic. How likely is that?

TargetDistance | ADC: 0.25xGEO_CHGIRN_Ganj_Dareh_NLevant_PPNBMAR_ENTUR_Barcin_NTUR_Boncuklu_NTUR_Kumtepe_NTUR_Kumtepe_N_low_resTUR_Tepecik_Ciftlik_NWHGYamnaya_RUS_Kalmykia
Italian_Abruzzo0.016736730.00.00.00.01.47.89.60.053.20.028.0
Italian_Aosta_Valley0.029097770.00.00.00.019.035.80.00.02.25.038.0
Italian_Apulia0.017003250.00.60.00.00.02.216.00.055.40.025.8
Italian_Basilicata0.015769800.01.80.00.01.04.411.00.056.60.025.2
Italian_Bergamo0.023992350.00.00.00.08.643.60.00.013.02.832.0
Italian_Calabria0.017125460.02.80.40.80.02.619.61.651.20.021.0
Italian_Campania0.017259350.02.82.00.00.64.014.40.053.20.023.0
Italian_Jew0.014629700.05.417.00.40.00.017.60.041.20.018.4
Italian_Lazio0.019164550.00.00.00.05.415.40.60.050.60.627.4
Italian_Liguria0.024660380.00.00.00.04.033.06.22.818.63.032.4
Italian_Lombardy0.024223400.00.00.00.00.049.80.00.016.61.432.2
Italian_Marche0.017152610.00.00.00.03.814.48.60.042.80.230.2
Italian_Molise0.015102360.00.00.00.00.011.62.80.057.20.028.4
Italian_Northeast0.023689510.00.00.00.011.242.60.00.04.23.238.8
Italian_Piedmont0.019836900.00.00.00.04.835.80.00.025.01.233.2
Italian_Trentino-Alto-Adige0.021256840.00.00.00.014.440.80.00.06.24.234.4
Italian_Tuscany0.017496310.00.00.00.00.231.89.20.027.00.831.0
Italian_Umbria0.017174020.00.00.00.00.023.89.20.037.00.030.0
Italian_Veneto0.022119770.00.00.00.010.041.60.00.010.43.234.8
Sicilian_East0.017533900.00.65.20.40.00.025.82.042.20.223.6
Sicilian_West0.018955940.00.41.02.80.02.015.40.052.42.423.6
Maltese0.016373740.00.26.23.60.01.626.61.237.00.023.6
Greek_Central_Anatolia0.020185209.611.00.00.00.00.00.01.070.00.08.4
Greek_Central_Macedonia0.022415670.00.00.00.00.227.60.00.037.80.234.2
Greek_Crete0.016257790.06.80.00.00.00.00.00.072.20.021.0
Greek_Izmir0.017646610.00.40.00.00.09.60.40.061.40.028.2
Greek_Kos0.017660850.08.60.00.00.80.80.20.072.80.016.8
Greek_Peloponnese0.018883260.00.00.00.01.216.47.20.047.40.027.8
Greek_Thessaly0.018909020.00.00.00.02.222.49.20.033.20.033.0
Greek_Trabzon0.0229159628.09.60.00.00.00.00.03.059.40.00.0
Average0.019374301.31.71.10.33.017.47.00.440.20.926.8

According to this study, https://www.cell.com/current-biology/fulltext/S0960-9822(16)30850-8, early farmers from Europe were more similar to samples from Tepecik-?iftlik and Barcın taht to ones from Boncuklu.

We further evaluated genetic differentiation among Boncuklu, Tepecik-?iftlik, Barcın, European Mesolithic, and Neolithic populations by calculating F[SUB]st[/SUB] (Supplemental Experimental Procedures; Data S3). The results were consistent with the pattern of differentiation in the PCA; particularly, Boncuklu appeared to be distinct from both Tepecik-?iftlik and Barcın (F[SUB]st[/SUB] = 0.020 and 0.030, respectively; Z > 4). A multidimensional scaling (MDS) plot summarizing pairwise F[SUB]st[/SUB] values revealed clustering of Tepecik-?iftlik and Barcın with European Neolithic populations, whereas Boncuklu attained a peripheral location (Figure 2C). This peripheral location is most likely due to high genetic homogeneity and drift in Boncuklu, as such a pattern was not observed in an MDS analysis of mean f[SUB]3[/SUB] statistics (Figure S3C).

We also found that Tepecik-?iftlik individuals were consistently closer to Iceman/Remedello and to Kumtepe than to any other Anatolian or European early Neolithic population, including their contemporary Barcın and the neighboring Boncuklu (Figure 3D). These results point to gene flow from an eastern source into Chalcolithic Kumtepe and later into Europe, which could have crossed central Anatolia already before the Chalcolithic.

As we have shown in this paper, individuals of the Chalcolithic Remedello group [24
] from northern Italy also share strong affinity with Kumtepe.
 
No doubt, I2683 is the one that overlaps with modern SItaly1.

SaOtiQo.png


qZewmh7.png

This study is specifically using Bar8 with I guess is Barcin8_N, for Anatolian_N.
 
The first traces of Iran_N were found in the Neolithic age in Italy and there is a complete lack of samples from the Iron and Bronze Ages in southern Italy in order to draw conclusions.

Considering that Davidski is constantly modifying and adjusting his averages and values of ancient and modern samples on G25 so that are in line with his ideas, considering that these instruments are very inaccurate, and that we are talking about minimum percentages, which become really significant only in the extreme south of Italy, it seems to me quite strange that many users are only interested in discussing this when it comes to Italians genetics.

If indeed he's using a bunch of samples including people from all over the world who just happened to have died and been buried in Rome to stand for "Imperial Romans", a decision based on a vast and unwarranted assumption, and has, for goodness' sakes averaged or at least is using "all" those "Philistine" samples of which only one is actually a Philistine, then it's a useless tool for these purposes. That's not to mention he has picked "Sizzi" chosen samples and samples handpicked by Sikeliot.

It should be no surprise he tinkers with the tool. I once published a whole discussion he had with one of his Russian friends on Stormfront where they were trying to figure out which samples to use to get the results they wanted.

That's another episode in the history of this "hobby" which most people would prefer to forget, or with which they agree, and so they want to hide it.

Once a liar, always a liar. It's called being an impeachable witness.:)

Would you hire a scientist for your team who was proven to have fudged data? It's happening right now in the Social Sciences, especially psychology, and heads are rolling. About time, too.
 
Is there a selection from G25 Imperial Roman without outliers to the NCE or Near East? Possibly even averaged for a comparison? Would like to compare the position and results.
 
The main differences between most West Eurasians are the ancestral proportions, not the basic ancestral components, especially if you take this very basic components as a reference. An "Imperal_Roman-like" ancestry is fairly widespread in the Imperial Roman time. Its just there. I know its just an approximation, and we can model the same profile differently. But its just more comfortable to use, because it captures and ancestral profile of importance probably the best - so far and until better samples are out there. You just find that profile again and again, even in the samples from Szolad and Collegno.

You find samples like what you call the more eastern "Imperial Roman" samples in Szolad because that area was the site of a pocket of isolated Romance speakers who attracted craftsmen from all over the ancient Roman world. People came from more west and more east of Pannonia. At one point it was also re-taken by the Byzantines and Byzantines naturally lived and died there.

Would you take all those samples and say they're representative of Pannonia as a whole?

That would be beyond absurd, and yet that's what you're doing for Rome.
 
Regarding G25 PCA, perhaps a way of "judging" it is comparing its results to those from researchers?
By the way, afaik the author of the G25 PCA himself sometimes uses qpAdm to reinforce certain results, to see if the different tools "agree" in something, and researchers may do the same.

From the paper Genetic origins of the Minoans and Mycenaeans (if you guys know about a better one for this purpose, please tell me), from 2017 - as the G25 PCA release, if I'm not mistaken:
https://static-content.springer.com...jects/41586_2017_BFnature23310_MOESM1_ESM.pdf
"In this section, we model these populations using the qpWave/qpAdm framework1 which allows us to model a Left set consisting of a Test population (whose history of admixture we are investigating) and Reference populations (potential sources of ancestry) in relation to a Right set of outgroups.
(...)
We test for rank=N-1 using qpWave and estimate mixture proportions using qpAdm1 . We report only feasible mixture proportions (in interval [0, 1]). We use a significance level of p=0.05 for rejecting models and mark p-values greater than 0.05 (that represent feasible models) in red.
(...)
We test the robustness of the qpAdm estimates by plotting populations with the inferred mixture proportions (Table S2.26) in the weighted average position of their source populations in PCA space. The results (Extended Data Fig. 5) indicate a close correspondence between the two in all models considered.
(...)
The population from Bronze Age southwestern Anatolia does not form a clade with any single (N=1) population of the All set (p-value for rank=0 < 1e-25). It cannot be modelled as any 2-way mixture (Table S2.8), with the best ones involving a mixture of Anatolian Neolithic and either Iran Neolithic or Caucasus hunter-gatherers. This population can be modelled as a 3-way mixture (Table S2.9) of ~62% Neolithic Anatolian, ~32% Caucasus hunter-gatherer (CHG), and ~6% Levantine Neolithic ancestry. This extra Levantine Neolithic ancestry parallels the PCA (Fig. 1b) that shows that the Bronze Age Anatolian sample is to the “east” (towards the Levant) relative to the Minoans and Mycenaeans."

So, according to this tool, the best 3-way mixture for SW Anatolian BA is ~62% Neolithic Anatolian, ~32% CHG and ~6% Levantine Neolithic.

I checked G25 results using R, and supposedly using the same components.
The target is the average for SW Anatolia EBA (Isparta).

Using individuals as source / Using Averages

Scaled and penalty = 0 (default in Vahaduo)
TUR_Barcin_N,53 / 53.4
GEO_CHG,28.6 / 29.6
Levant_PPNB,18.4 / 17

Scaled and penalty = 0.001 (as some people prefer)
TUR_Barcin_N,71.4 / 53.8
GEO_CHG,18.2 / 29
Levant_N,10.4 / 17.2

Unscaled and penalty = 0
Anatolia_Barcin_N,52.2 / 54.6
GEO_CHG,25.8 / 25.2
Levant_PPNB,22 / 20.2

Unscaled and penalty = 0.001
Anatolia_Barcin_N,72.2 / 55
Levant_PPNB,15.2 / 20.2
GEO_CHG,12.6 / 24.8

Substantial differences between the results using individuals (rather than averages) and different penalties.
Vahaduo scaled would return 17-18.4% for Levant_PPNB in SW Anatolia EBA average then. Considering samples from Isparta individually as target (rather than the average), the variation would be something like 14-20.6 in Vahaduo scaled (pen = 0), against those 6% from qpAdm.

Conclusion is: The reference samples are supposedly the same, but for whatever reason there's no agreement between that qpAdm's and G25's regarding proportions. If we use researchers' tools as reference, the Levant Neo contribution for SW Anatolia EBA is possibly overestimated by G25. Whilst using pen = 0.001 may decrease the difference in Levant Neo %, it'd affect negatively the other two, especially CHG.

Excellent analysis, Regio, sorry I'm out of juice.:)
 
You find samples like what you call the more eastern "Imperial Roman" samples in Szolad because that area was the site of a pocket of isolated Romance speakers who attracted craftsmen from all over the ancient Roman world. People came from more west and more east of Pannonia.

That's exactly what I want to capture: The Roman Imperial era migrations! Like you said, they came to this place from different parts of the empire and approach the "Imperial_Roman" profile. That's why the Imperial_Roman sample is useful, even as an average, because its close to the usual "colonial" admixture in many parts of the empire and of course, Italy and Rome itself. Actually, the Lombards, since some of them were admixed, brought not just Germanic, but also exactly this ancestry to Italy too, even though their immigration as a whole surely caused the Northern shift I talked about as well. And in Collegno they again met similar local people.

Would you take all those samples and say they're representative of Pannonia as a whole?

If the regional samples look quite different, no. I'm hoping for more Eastern-Central European samples and even in the Szolad samples there is this non-Germanic component close to Eastern Celts and Early Slavs, which I would assume is either local (Illyro-Celtic?), Scythian-Sarmatian or a mixture with actual Slavs. I want to have more samples to explore that, as well as how big the Roman colonisation influence was. The "Roman colonisation" being probably not always accurately, but comfortably captured with the Imperial Roman sample. Because its exactly this kind of wild mix of Greek, Italian Roman and additional NCE, NA and Levantine-Near Eastern influences which seems to have been attracted especially by the larger and flourishing Roman settlements.

That would be beyond absurd, and yet that's what you're doing for Rome.

Actually I try to bridge the gap from Italics-Etruscans to Medieval Italians. And looking at the results so far, there was an obvious influx first from the East/South (Imperial Era), than from the North, producing modern Italians. The Imperial Roman captures the Eastern/Southern flow, the later samples from Late Antiquity on the Northern one, both from within Italy as well as outside. To me Imperial Rome is no homogeneous population to begin with. That's also clear if looking at the classical and modern literature, as well as the biographies and personal data. However, the elements which appeared in Imperial Roman times were not new, you find them in pre-Imperial samples too, but they increased in a disproportional way which can't be explained by internal Italian migration alone. That's not feasible. Like the later Northern-Central European shift too can't be explained just by internal migration, that's not feasible neither. Italy as a whole got shifted towards the Levante/East-South in Imperial times and North afterwards, after the collapse of Rome. That's also what the recent studies told us, or at least how I understood them.

If someone does purify the Imperial Roman sample from outliers, I'm pretty sure this will be a better fit on the one hand, but on the other it will just result in modern Italians needing more Northern-Central European, Levantine, Near Eastern and North African admixture, because a large part of this for Central-Southern Italy being accounted for if using the average. But I'm curious how it will work out and hope somebody has done it (y)
 
Last edited:
No, I can't, and that's exactly the reason I am asking people about their ideas on what happened to them. Are you triggered by it? Why? Is it the aforementioned unexplained fear of having some even tiny ancestral relationship with those "Arab" Levantines?

Oh, and Crete by the IA was also a Greek island, and a very important one at that. Cyprus is another, and the Levantine ancestry is also there in significant percentages.

I have no problem with Levantine ancestry. It was probably there in Crete in 2000BC and earlier. Except the Philistines were in the Levant between 1200BC and 600BC and then we don't hear about them anymore so Crete of 2000BC did not owe their Levantine ancestry to the Philistines. Could there have been additional influx from the Philistines? Sure but we don't have any samples to prove or disprove it. So you're totally speculating with no history, no archaeology and no genetics.
 
That's exactly what I want to capture: The Roman Imperial era migrations! Like you said, they came to this place from different parts of the empire and approach the "Imperial_Roman" profile. That's why the Imperial_Roman sample is useful, even as an average, because its close to the usual "colonial" admixture in many parts of the empire and of course, Italy and Rome itself. Actually, the Lombards, since some of them were admixed, brought not just Germanic, but also exactly this ancestry to Italy too, even though their immigration as a whole surely caused the Northern shift I talked about as well. And in Collegno they again met similar local people.



If the regional samples look quite different, no. I'm hopeing for more Eastern-Central European samples and even in the Szolad samples there is this non-Germanic component close to Eastern Celts and Early Slavs, which I would assume is either local (Illyro-Celtic?), Scythian-Sarmatian or a mixture with actual Slavs. I want to have more samples to explore that, as well as how big the Roman colonisation influence was. The "Roman colonisation" being probably not always accurately, but comfortably captured with the Imperial Roman sample. Because its exactly this kind of wild mix of Greek, Roman and additional NCE, NA and Levantine-Near Eastern influences which seems to have been attracted especially by the larger and flourishing Roman settlements.



Actually I try to bridge the gap from Italics-Etruscans to Medieval Italians. And looking at the results so far, there was an obvious influx first from the East/South (Imperial Era), than from the North, producing modern Italians. The Imperial Roman captures the Eastern/Southern flow, the later samples from Late Antiquity on the Northern one, both from within Italy as well as outside. To me Imperial Rome is no homogeneous population to begin with. That's also clear if looking at the classical and modern literature, as well as the biographies and personal data. However, the elements which appeared in Imperial Roman times were not new, you find them in pre-Imperial samples too, but they increased in a disproportional way which can't be explained by internal Italian migration alone. That's not feasible. Like the later Northern-Central European shift too can't be explained just by internal migration, that's not feasible neither. Italy as a whole got shifted towards the Levante/East-South in Imperial times and North afterwards, after the collapse of Rome. That's also what the recent studies told us, or at least how I understood them.

If someone does purify the Imperial Roman sample from outliers, I'm pretty sure this will be a better fit on the one hand, but on the other it will just result in modern Italians needing more Northern-Central European, Levantine, Near Eastern and North African admixture, because a large part of this for Central-Southern Italy being accounted for if using the average. But I'm curious how it will work out and hope somebody has done it (y)

You completely miss the point, again! The point is, should you be using a group of locals plus these Byzantines as a source for the modern Hungarian population. The answer is emphatically NO. When the Byzantines left most of these people undoubtedly fled with them or scattered to safer and more profitable areas.

If you don't see the lack of logic in your argument there's nothing more to be said.
 
I have no problem with Levantine ancestry. It was probably there in Crete in 2000BC and earlier. Except the Philistines were in the Levant between 1200BC and 600BC and then we don't hear about them anymore so Crete of 2000BC did not owe their Levantine ancestry to the Philistines. Could there have been additional influx from the Philistines? Sure but we don't have any samples to prove or disprove it. So you're totally speculating with no history, no archaeology and no genetics.

Amen to that.
 
It is like trying to determine the genetics of Americans by digging in the ethnic cemeteries in Chicago or New York 2000 years from now. The average American would be a lot more admixed than the people buried in the Jewish cemeteries, the Lithuanian cemeteries or the Serbian cemeteries. With limited samples at that.

Also I have to caution against using statistical best fit models particularly for an admixed population. Take 4 population Oracles in Gedmatch. Yeah those provide the best fit but some of them don't make any sense. In my case if you give me some percentage of Bulgarian & Greek & Romanian it make sense since my ancestors have lived in Eastern Thrace as far back as we can trace. But if you give me a mix of Galicians and Russians it would not make sense unless you have archaeological and historical evidence that the admixture happened in the geographical area of interest.
 

This thread has been viewed 188541 times.

Back
Top