Slavic G25 (ancient) calculator

Usual discrepancy between my parents results and mine as in every Gedmatch or G25 calc:) I think FTDNA tested them using different chip, although it was not officially stated. SNP number is the same anyway.


To have higher Baltic and Greco-Roman in the same time than both of them is not very probable...

Target: Lukasz_scaled
Distance: 2.5041% / 0.02504113 | ADC: 0.25x
52.2 Slavic
39.6 Baltic
8.2 Greco-Roman

Target: mother
Distance: 1.5541% / 0.01554147 | ADC: 0.25x
80.4 Slavic
11.8 Baltic
5.0 Greco-Roman
2.2 Balkan
0.6 Celtic

Target: father
Distance: 2.3031% / 0.02303099 | ADC: 0.25x
59.0 Slavic
36.2 Baltic
3.4 Greco-Roman
1.4 Celtic

This calculator could use a reboot. I am thinking maybe I should not inlude samples as old as bronze age, only medieval and iron age. Could improve accuracy, maybe.
And add few more samples, remove less clean ones ?

Scythian Moldova, I have it as one of Paleo-Balkan samples, which exact time frame is that?

Need a piece of advice: could some of IA French samples be added as Celtic proxy, if they are any good, and if yes, which?
Also thinking to include some Hungarian Scythian for Paleo-Balkan (very southern shifted results) plus Empiures as another Greek sample.

Opinions?

I'd remove BA Balts and BA ''Croatians'', probably it is true they have no yet specific Baltic or Illyrian drift so they skew some results. Also Lech because it's also BA. And I am thinking to remove early Czech Slav, but didn't decide yet because it's local admixture is very small.

Since I have only one Germanic sample, I was thinking to add Iron Age Danish sample too, if it's clean (it should be, right?)
so many questions.... :)))

Quote:

---

Originally Posted by Feiichy

Since I have only one Germanic sample,

---

Which sample it is?

You can use Poprad or those Szolad samples which are highly Germanic, as good proxy of early medieval Germanics in Central Europe.

Quote:

---

Originally Posted by Lukasz

Which sample it is?

---

Germany Medieval

I just wonder Lukasz, do you think you really have significant Baltic admix or is this just shared Balto-Slavic ancestry in BA before groups separated? I did thest those samples I chosed and they are all firmly closest to modern day Balts, but I am suspicious you score less Slavic than I do.

That is why I start to doubt how much reliable these samples are.

Quote:

---

Originally Posted by Feiichy

This calculator could use a reboot. I am thinking maybe I should not inlude samples as old as bronze age, only medieval and iron age. Could improve accuracy, maybe.
And add few more samples, remove less clean ones ?

Scythian Moldova, I have it as one of Paleo-Balkan samples, which exact time frame is that?

---

looking forward to the reboot! yes i agree remove the samples wich are a not so clean. Because with my results i dont get my 1/4 Germanic because there was already some Germanic in the Balkan sample and when i removed it i got my 1/4 Germanic.

Quote:

---

Originally Posted by 17571imre

looking forward to the reboot! yes i agree remove the samples wich are a not so clean. Because with my results i dont get my 1/4 Germanic because there was already some Germanic in the Balkan sample and when i removed it i got my 1/4 Germanic.

---

There is no Germanic in Balkan samples, but there is western European drift in it (bell beaker and even "celtic" like-not actual celtic at all though, just genetic overlap).

These samples are not mixed if that is what you mean, entire Croatia was homogemously North Italian like in bronze age, both in Dalmatia and Slavonia and very similar to early Italics/Latins, Etruscans and Republican Romans.

But, I doubt such genetic profile remained in Balkans for long because modern Balkanites (even most Croats) usually can't be modeled with it so well. Maybe it became already more southeastern like in iron age. IA Croatian sample is mislabeled as far as I know and actually just late BA one (someone please correct me if wrong, it's important for bigger picture)

I can be modeled with it, but well, I think that is because I am part west European so it eats my western admixture similary like yours.

My mom who is 100% pure Slav


Target: Daco_mom_scaled
Distance: 2.7529% / 0.02752944
86.2 Germanic
9.2 Celtic
2.4 Balkan
2.2 Baltic

Target: Daco_mom_scaled
Distance: 2.7610% / 0.02761014 | ADC: 0.25x
90.6 Germanic
8.8 Celtic
0.6 Slavic

Distance to: Daco_mom_scaled
0.02806770 Germanic
0.04982111 Celtic
0.05778298 Slavic
0.06183121 Slavic
0.07380064 Slavic
0.08691940 Celtic
0.09381964 Balkan
0.09677089 Balkan
0.09919664 Balkan
0.10003828 Baltic
0.10391231 Balkan
0.12193822 Baltic
0.12940293 Baltic
0.14208524 Balkan
0.16246084 Greco-Roman
0.16899246 Greco-Roman
0.20126000 Finnic
0.24957585 Turkic
0.24995627 Turkic

Quote:

---

Originally Posted by Daco Celtic

My mom who is 100% pure Slav


Target: Daco_mom_scaled
Distance: 2.7529% / 0.02752944
86.2 Germanic
9.2 Celtic
2.4 Balkan
2.2 Baltic

Target: Daco_mom_scaled
Distance: 2.7610% / 0.02761014 | ADC: 0.25x
90.6 Germanic
8.8 Celtic
0.6 Slavic

Distance to: Daco_mom_scaled
0.02806770 Germanic
0.04982111 Celtic
0.05778298 Slavic
0.06183121 Slavic
0.07380064 Slavic
0.08691940 Celtic
0.09381964 Balkan
0.09677089 Balkan
0.09919664 Balkan
0.10003828 Baltic
0.10391231 Balkan
0.12193822 Baltic
0.12940293 Baltic
0.14208524 Balkan
0.16246084 Greco-Roman
0.16899246 Greco-Roman
0.20126000 Finnic
0.24957585 Turkic
0.24995627 Turkic

---

No point in this calc for NW Euros, especially not for Irish since their mix requires NW Euro Beaker samples, which for Slavs/eastern euros aren't relevant.

Quote:

---

Originally Posted by Feiichy

No point in this calc for NW Euros, especially not for Irish since their mix requires NW Euro Beaker samples, which for Slavs/eastern euros aren't relevant.

---

I know it's not intended for NW Euros but I just like to see what comes up. Sometimes the fits are better than you might expect.

Quote:

---

Originally Posted by Feiichy

That is why I start to doubt how much reliable these samples are.

---

Here is what some Wielkopolskans score in your calculator:

(can you reveal what are your 3 Slavic reference samples - I suppose modern Slovaks are among them)

https://i.imgur.com/GGsGofp.png

For comparison - here is how we plot in the Global25 PCA:

https://i.imgur.com/B9WNHjm.png

Quote:

---

Originally Posted by Feiichy

These samples are not mixed if that is what you mean, entire Croatia was homogemously North Italian like in bronze age, both in Dalmatia and Slavonia and very similar to early Italics/Latins, Etruscans and Republican Romans.

But, I doubt such genetic profile remained in Balkans for long because modern Balkanites (even most Croats) usually can't be modeled with it so well.

.

---

Albanians in many calcs have central Italian affinity.

Quote:

---

Originally Posted by Lukasz

Usual discrepancy between my parents results and mine as in every Gedmatch or G25 calc:) I think FTDNA tested them using different chip, although it was not officially stated. SNP number is the same anyway.


To have higher Baltic and Greco-Roman in the same time than both of them is not very probable...

Target: Lukasz_scaled
Distance: 2.5041% / 0.02504113 | ADC: 0.25x
52.2 Slavic
39.6 Baltic
8.2 Greco-Roman

Target: mother
Distance: 1.5541% / 0.01554147 | ADC: 0.25x
80.4 Slavic
11.8 Baltic
5.0 Greco-Roman
2.2 Balkan
0.6 Celtic

Target: father
Distance: 2.3031% / 0.02303099 | ADC: 0.25x
59.0 Slavic
36.2 Baltic
3.4 Greco-Roman
1.4 Celtic

---

Not very probable, but at the same time it's entirely possible. You should look at more results of parents and offspring. This is more common than you think.

https://blogs.ancestry.com/ancestry/...hy-it-matters/

Quote:

---

Originally Posted by Ion Basescul

Not very probable, but at the same time it's entirely possible. You should look at more results of parents and offspring. This is more common than you think.

https://blogs.ancestry.com/ancestry/...hy-it-matters/

---

Those results aren't the best examples
https://blogs.ancestry.com/ancestry/...ween-sista.png

Because Ireland, Britain and West Euro ancestry are similar for those family as we see, and could be interchangeable to large extent. Look that their Finnish level is stable for example

Quote:

---

Originally Posted by Peterski

Here is what some Wielkopolskans score in your calculator:

(can you reveal what are your 3 Slavic reference samples - I suppose modern Slovaks are among them)

https://i.imgur.com/GGsGofp.png

For comparison - here is how we plot in the Global25 PCA:

https://i.imgur.com/B9WNHjm.png

---

It's ancient calculator as title say, I never use nodern samples for such calculations.

Slavic inclides early Czech Slav, Avar Szolad and Sunghir medieval Russian.

I think in new version I 'll use only individual samples instead of averages, for cleaner results.

Quote:

---

Originally Posted by Peterski

Here is what some Wielkopolskans score in your calculator:

(can you reveal what are your 3 Slavic reference samples - I suppose modern Slovaks are among them)

https://i.imgur.com/GGsGofp.png

For comparison - here is how we plot in the Global25 PCA:

https://i.imgur.com/B9WNHjm.png

---

She used Czech early Slav, szolad avar, and Rus sunghir.

Share your results bro!

Sent from my ANE-LX1 using Tapatalk

Quote:

---

Originally Posted by Karol Klačansky

Share your results bro!

---

I did, Wlkp1a and Wlkp1b is me with two different raw data files.

Quote:

---

Originally Posted by Peterski

I did, Wlkp1a and Wlkp1b is me with two different raw data files.

---

Well you are very Slavic on this and less Germanic than I would expect. :)

Quote:

---

Originally Posted by Polrebby

Albanians in many calcs have central Italian affinity.

---

Yes, but they prefer Bulgarian IA over Croatian for modeling. All Balkanites do actually except Slovenians/NW Croats and that's just probably genetic similarity not direct ancestry.

These BA Balkan samples are problematic so Sorbs and Czechs score substantially because it's similar to some celtic samples. I will most likely remove them in next run.

Quote:

---

Originally Posted by Feiichy

I just wonder Lukasz, do you think you really have significant Baltic admix or is this just shared Balto-Slavic ancestry in BA before groups separated? I did thest those samples I chosed and they are all firmly closest to modern day Balts, but I am suspicious you score less Slavic than I do.

That is why I start to doubt how much reliable these samples are.

---

No samples are good, I'm just Baltic shifted more than average.

I always score lot of BA Croatian btw, unlike other Croats. When modeling with moderns I always get something Basque like/SW French/Iberian, unlike other south slavs.

Now question is do I have more pressrved BA Croatian ancestry than other south slavs because BA samples were very SW Euro like, or it is just my western euro side which makes it so?

I don't know the answer and even played with idea my Alsatian German side has some French roots, maybe from SW France or something so that is why I score like that.

But that is too small part of my ancestry to explain why my non Slavic side is so SW and not SE like in most other people.

Maybe Dalmatian islanders really preserved more of this BA Croatian ancestry than mainland south slav groups.

Quote:

---

Originally Posted by Lukasz

Those results aren't the best examples
https://blogs.ancestry.com/ancestry/...ween-sista.png

Because Ireland, Britain and West Euro ancestry are similar for those family as we see, and could be interchangeable to large extent. Look that their Finnish level is stable for example

---

Look at Asia then. Well, in any case, this is far from being uncommon. Just find some kits of families and you will see.
Another example is how I shift towards W. Europe compared to my parents' average, and their average pulls towards E. Europe instead. I also score as much East Asian as my mom, whereas my dad has less, so the average should theoretically be lower.
I recall off the top of my head that another Polish user had a similar inconsistency. In his case, he was scoring more Med than his parents' average.

This is the breakdown for Slavic samples (I see that you used Av1 and Av2 average, rather than just one of them):

https://i.imgur.com/eBy9zvY.png

For comparison here is what I scored in a model Aren proposed, where he used only Sunghir6 as the Slavic sample:

(the last one is Polish average)

https://i.imgur.com/GaO559x.png

^^^
In this model Wlkp3 scores 29% DEU_MA, while in your model all of this goes into Czech Slav and Avars Hungary.

I already said in next update only individual samples will be used instead of averages for very same reason.

But you know what is interesting? That in PCA, none of the Wielkopolskans plot in this area:

https://i.imgur.com/j9ebm8p.png

I wonder which regions of Poland are the Polish samples that plot in this area on the PCA.

Quote:

---

Originally Posted by Peterski

This is the breakdown for Slavic samples (I see that you used Av1 and Av2 average, rather than just one of them):

https://i.imgur.com/eBy9zvY.png

For comparison here is what I scored in a model Aren proposed, where he used only Sunghir6 as the Slavic sample:

(the last one is Polish average)

https://i.imgur.com/GaO559x.png

^^^
In this model Wlkp3 scores 29% DEU_MA, while in your model all of this goes into Czech Slav and Avars Hungary.

---

Can you share the source pops here.

Sent from my ANE-LX1 using Tapatalk

Quote:

---

Originally Posted by Karol Klačansky

Can you share the source pops here.

---

The source pops for what? Aren's model? I think it was only those 4 visible (Bylany, DEU_MA, Rome_Imperial, Sunghir_MA).

Quote:

---

Originally Posted by Peterski

The source pops for what? Aren's model? I think it was only those 4 visible (Bylany, DEU_MA, Rome_Imperial, Sunghir_MA).

---

yea man im just a lazy bastard.

del

I gave up from reboot, been working on it whole day until now but I wasn't satisfied with test runs.
Basically I can't make flawless model, including this one of course. Too much time and effort for little return.

Best to leave it to scientists. :)

Quote:

---

Originally Posted by Feiichy

I gave up from reboot, been working on it whole day until now but I wasn't satisfied with test runs.
Basically I can't make flawless model, including this one of course. Too much time and effort for little return.

Best to leave it to scientists. :)

---

It's a good model. Wont get much more accurate until we get more ancient samples. Best we can do is come up with ranges for our ancestry. Basically I'm anywhere from 50-60 percent Slavic from what I'm seeing genetically. I'm cool with this. I wish they would get more samples from great Moravian times so I could confirm I'm a direct discendent. That's all I'm waiting for.

Sent from my ANE-LX1 using Tapatalk

Model for me which gives me lowest possible fit (ancients)

Target: Feiichy
Distance: 1.0407% / 0.01040694

48.0 Balto-Slavic
33.2 Old_Balkan
13.2 North_Caucasian
5.4 German
0.2 Mongoloid

After I added ancient north caucasian samples from russia my fit improved significantly
No idea where it is from but everywhere I have very high west asian/caucasus/chg affinity.

And it seems my Slavic input is not anything special but rather low for a Croat.

This is first time I get such low distance fit, previous model was 2 times higher distance.

Quote:

---

Originally Posted by Feiichy

Model for me which gives me lowest possible fit (ancients)

Target: Feiichy
Distance: 1.0407% / 0.01040694

48.0 Balto-Slavic
33.2 Old_Balkan
13.2 North_Caucasian
5.4 German
0.2 Mongoloid

After I added ancient north caucasian samples from russia my fit improved significantly
No idea where it is from but everywhere I have very high west asian/caucasus/chg affinity.

And it seems my Slavic input is not anything special but rather low for a Croat.

This is first time I get such low distance fit, previous model was 2 times higher distance.

---

Yes, even in my oracle to Eurogenes K10 you got something like 87% Slovenian and 13% Armenian in mixed mode.

Quote:

---

Originally Posted by Feiichy

Model for me which gives me lowest possible fit (ancients)

Target: Feiichy
Distance: 1.0407% / 0.01040694

48.0 Balto-Slavic
33.2 Old_Balkan
13.2 North_Caucasian
5.4 German
0.2 Mongoloid

After I added ancient north caucasian samples from russia my fit improved significantly
No idea where it is from but everywhere I have very high west asian/caucasus/chg affinity.

And it seems my Slavic input is not anything special but rather low for a Croat.

This is first time I get such low distance fit, previous model was 2 times higher distance.

---

The Baltic is probably ''sucking'' lots of Slavic since Balts were more northeastern shifted from the original Slavs. Slavic subclades of I2a and R1a are over 60% in Croatia and Slavs were not pure R1a and I2a, they had other haplogroups. Croatia is really close to Poland in terms of Slavic Y-DNA R1a and I2a who are around 80%-85% Slavic. (The territory of Poland was depopulated the Gemanic tribes and repopulated by Slavs).

Around 65% to 75% Slavic ancestry in Croats, and 70% as the main average.

Quote:

---

Originally Posted by -Scar-

The Baltic is probably ''sucking'' lots of Slavic since Balts were more northeastern shifted from the original Slavs. Slavic subclades of I2a and R1a are over 60% in Croatia and Slavs were not pure R1a and I2a, they had other haplogroups. Croatia is really close to Poland in terms of Slavic Y-DNA R1a and I2a who are around 80%-85% Slavic. (The territory of Poland was depopulated the Gemanic tribes and repopulated by Slavs).

Around 65% to 75% Slavic ancestry in Croats, and 70% as the main average.

---

G25 Croatian average is from Zagreb, lot of Croats are more Med than that. I am mixed, not fully south slavic.

And I get much better fit by using pure Baltic samples (Iron Age Estonia and Golden Horde Euro, autosomally Lithuanian), than medieval Slavic samples.

Ph2ter told me same, that my Slavic part is more northwestern compared to his. So maybe that is why I plot quite north, my Slavic input is not big but is very purely NE Euro/Baltic like.

Quote:

---

Originally Posted by Feiichy

Ph2ter told me same, that my Slavic part is more northwestern compared to his. So maybe that is why I plot quite north, my Slavic input is not big but is very purely NE Euro/Baltic like.

---

The pagan Slavs that occupied Croatia were more I2a than R1a, so I doubt they resemebled genetically Balts more than those medieval Slavic samples that we have.

Quote:

---

Originally Posted by -Scar-

The pagan Slavs that occupied Croatia were more I2a than R1a, so I doubt they resemebled genetically Balts more than those medieval Slavic samples that we have.

---

I am not speaking for average Croats but for what I get :)
Maybe that expains why on myheritage I get East Euro + Greek/South Italian and no Balkan at all.
It seems my slavic component is very northern while my paleo part is very med, no much in between.
Also on 23andme I get very high east euro and low balkan, despite balkan is based on modern south slavs.

Quote:

---

Originally Posted by Feiichy

I am not speaking for average Croats but for what I get :)
Maybe that expains why on myheritage I get East Euro + Greek/South Italian and no Balkan at all.
It seems my slavic component is very northern while my paleo part is very med, no much in between.
Also on 23andme I get very high east euro and low balkan, despite balkan is based on modern south slavs.

---

It isn't meant to be taken literally, the key is to use the right samples. A Sicilian and be modeled as half Lebanese and half central European (I suppose).

Quote:

---

Originally Posted by Ion Basescul

Not very probable, but at the same time it's entirely possible. You should look at more results of parents and offspring. This is more common than you think.

https://blogs.ancestry.com/ancestry/...hy-it-matters/

---

Quote:

---

Originally Posted by Ion Basescul

Look at Asia then. Well, in any case, this is far from being uncommon. Just find some kits of families and you will see.
Another example is how I shift towards W. Europe compared to my parents' average, and their average pulls towards E. Europe instead. I also score as much East Asian as my mom, whereas my dad has less, so the average should theoretically be lower.
I recall off the top of my head that another Polish user had a similar inconsistency. In his case, he was scoring more Med than his parents' average.

---

Exactly.