I had to make this post since people were posting outdated or inaccurate models from shitty calculators that are even considered as inaccurate by their creators. The last post on this subreddit seems to have used Maronite Cypriots to represent Greek Cypriots.
**Admixtools** is an open-source software package developed by researchers at Harvard University for analyzing and visualizing population admixture and structure in genetic data. The software provides a suite of tools for estimating admixture proportions, identifying population structure, and conducting hypothesis tests for various demographic scenarios.[https://academic.oup.com/genetics/article/217/4/iyaa045/6070149](https://academic.oup.com/genetics/article/217/4/iyaa045/6070149)
The fixation index (FST) is **a measure of population differentiation due to genetic structure**. It is frequently estimated from genetic polymorphism data, such as single-nucleotide polymorphisms (SNP) or microsatellites.
[admixtools Fst distance ](https://preview.redd.it/s6xhg9l81hta1.png?width=1624&format=png&auto=webp&v=enabled&s=e6f0a2519acc6cc6f46d66516db5ba7fe0ffc95c)
I’m too lazy to run Cypriots on admixtools again so here’s the distance to them on Global25 instead since its scaled on fst (still not perfect). Aegean Islanders are the closest, then Lebanese. Jews are mixed with South Italians and the Levant, thats why they are so close.
[G25 distance](https://preview.redd.it/ofdoqoaf1hta1.png?width=330&format=png&auto=webp&v=enabled&s=f49d870e3e6bfd4626a33fcc5e85e752a5b1c80d)
[G25 model](https://preview.redd.it/j9vyycyy3hta1.png?width=728&format=png&auto=webp&v=enabled&s=547181b43551a08e573b6c0c1623314f97e3f322)
When using too many proxies to model people it can create overfits, Anatolian Iron Age and Mycenaeans were similar genetically so its hard to tell those ancestries completely apart.
I removed Anatolian in this model to show total “Mycenaean like” ancestry/shift , it does NOT imply actual Mycenaean ancestry.
[G25 model](https://preview.redd.it/vx8u0ah84hta1.png?width=670&format=png&auto=webp&v=enabled&s=d96c77f6d9c7c77ed4fdc1134d1692530c24e172)
Like above but without the South Slavic proxy because it has Balkan ancestry, this uses a Slavic proxy from medieval Hungary. So this is the total Mycenaean like shift/ancestry from mixing with related populations.
[G25](https://preview.redd.it/6w9n66ve6hta1.png?width=597&format=png&auto=webp&v=enabled&s=6a66e0f6ed419704fec99d55e3ae21312e99f3eb)
​
https://preview.redd.it/6h0vot8ykita1.png?width=713&format=png&auto=webp&v=enabled&s=d30771b7edc29395ee39c033386d16cceb3028e5
The main Early Bronze Age shifts.
https://preview.redd.it/uiqikt4pvita1.png?width=813&format=png&auto=webp&v=enabled&s=18dc3b82dc55760fed031fc6d2c3fe2e75f5fbb4
You can see below that Mycenaeans and Anatolians were similar.
[Neolithic qpAdm admixtools model](https://preview.redd.it/o3vysmnd4hta1.png?width=1466&format=png&auto=webp&v=enabled&s=ec6ca144e52638f5a48f58226c7ed06b10350580)
​
[Mesolthic qpAdm admixtools model](https://preview.redd.it/425i6lue4hta1.png?width=1532&format=png&auto=webp&v=enabled&s=4c64c8adc48cec447eaa9c27155dfb7e36ec32d8)
MyHeritage is extremely inaccurate, do not trust their results. They do not use Cypriots as a reference and they use other modern MIXED people as references. You cant just use modern Greeks who have Slavic ancestry as a Greek reference for Cypriots.
Another example is Jews. South Europeans score Jewish on myheritage, those Jews have literal 50-60% Italian ancestry + 20-30% Levantine.
6 comments
Thanks a lot for the post and the explanations.
My admixture is mix with tashi
Gorilla’s got data for days
Mycenaeans are pretty much 80%+ old Aegean (which means Minoan-like, which in return means a mix of Anatolian Neolithic with Caucasus HG/Iran Neo and a little bit of natufian probably) plus the rest yamnaya (EHG plus CHG for the most part). Indeed using Anatolian+Mycenaean in these models is redundant. Cypriot Greeks being closest to Dodecanese Greeks makes perfect sense of course (same dialect group too).
What confuses me is your contemporary mainland Greek qpADM results. They return more natufian and less WHG (from absorbing Slavs) than I would expect.
Sawtooth favorite Levant people, the Druze.
On a more serious note, I’ve read somewhere that Cypriots were relatively heterogenous pre-1974 and different areas were somewhat different genetically for a small island. Anecdotally, moprphou and kyrenia people were usually more dark skinned, karpasia people were more fair etc.
Do we know how much that is reflected in the data if at all?