German GEDmatch results

[vbnetkhio]

Baltic component peaks in samples like these



So it is essentially a Baltic Bronze Age or a vague EHG component (composed in various shades of WHG+ANE).





If only Global25 was free and easily accessible like GEDmatch....

K13 can model Nordic-like, Balto-Slavic-like and Italian-like groups properly, and coincidentaly most of Eastern/Central Europe and the Balkans are a mix of exactly such groups, so it's a useful calculator for us.
When used for this area, the Baltic components really signifies Balto-slavic influence.

For most other things it fails, and the components act as bad proxies. I don't mean ancient samples only.

For example, it shows the Finnish as Nordic+ Balto-Slavic, but they actually have addtional WHG and EHG from Sami-like admixture, which in k13 goes into NA and Baltic and probably messes up the ratio.

[vbnetkhio]

So essentially 70% EHG and 30% EEF.

the Baltic component is based on modern Latvians or Lithuanians, however because of the "calculator effect", Latvians and Lithuanians themselves get only 50-60% Baltic.
Same goes for the British and the North Atlantic component.

So they are basically nothing, and it's better to treat them as coordinates (like those in G25) than actual components.

[rothaer]

(…)
So they are basically nothing, and it's better to treat them as coordinates (like those in G25) than actual components.

I agree with exclamation mark. This is applicable to a number of GEDmatch calculators. How much unnecessary texting could have been avoided if f. i. the “Iberian” component of K36 just would have been called „component 7“ (figuratively)?!

[vbnetkhio]

G25 cost too much and their modern references are terrible. I don't see it replacing gedmatch anytime soon.

k13 results can be converted to G25 with Genoplot.
somebody on anthrogenica converted the Macedonian Greek average like this, and it was almost identical to the one already in G25.
so it's realiable, at least in some cases.

[rothaer]

(…)

For example, it shows the Finnish as Nordic+ Balto-Slavic, but they actually have addtional WHG and EHG from Sami-like admixture, which in k13 goes into NA and Baltic and probably messes up the ratio.

This is one of the most insightful and important comments that I’ve read for a year. Exactly this problem is existent in Eurogenes calculators up to G25 inclusively even. But not in K36.

[vbnetkhio]

This is one of the most insightful and important comments that I’ve read for a year. Exactly this problem is existent in Eurogenes calculators up to G25 inclusively even. But not in K36.

I don't think g25 has this problem.. k36 does, for example it has a fennoscandian component based on modern Finns, so it can model Finns properly, but it still can't deal with Sami (because of a lack of WHG and EHG components)

[Peterski]

What I learned while collecting various verified samples? Don't trust academic samples, they are often random people from some city with mixed ancestry (either ethic or if not, than regional).

"German East" everyone used for years is a great example (and we use it still, just renamed it to German_Saxony). Now when we got decent number of fully verified kits from Saxony, I can see that average wasn't good and probably included people with ancestry not only from Saxony, but other places. Nobody seems to check 4/4 grandparents in academic papers it seems, their reliability is really low.

New Saxony average we will publish soon to replace academic one (won't mix them but totally drop Leipzig academic kits) is pretty different and quite less eastern.

It's even close to "central German" genetic profile in a way, something many argued doesn't exist. All academic averages on Vahado need to be replaced with better ones, they are simply not reliable.

About not trusting academic samples I told you already in 2016 but then you laughed and believed in academic data like in a holy grail, and almost called me names for wanting to use non-academic data.

But now you are going from one extreme to another. You shouldn't dismiss academic samples entirely. There are no obvious outliers in the Leipzig sample in my opinion (they all look like they could be from Saxony, just from different parts of Saxony).

Check for comparison kits from Thuringia and from Saxony-Anhalt before you claim that your new average for Saxony is representative (it can be too western-shifted).

The GEDmatch kits I have for Saxony are not much different than the Leipzig academic average. So maybe you just found kits from a specific sub-region of Saxony.

Here is a German sample with all 4/4 grandparents from Thuringia for comparison, and still very eastern-plotting:

G25 coordinates (scaled):

DE_Thuringia1,0.132035,0.13405,0.056945,0.047804,0.036622,0.014223,0.010105,0.00923,-0.00225,-0.016583,-0.001786,0.00015,0.0055,0.01679,-0.000543,0.003713,0.001304,-0.003421,-0.002514,-0.001751,-0.002995,0.001855,0.003204,0.003253,-0.001197

Distance to: DE_Thuringia1
0.01788853 Czech
0.02052746 Ukrainian_Zakarpattia
0.02251585 Hungarian
0.02336433 Ukrainian_Lviv
0.02427786 German_East
0.02446324 Slovakian
0.02520100 Croatian
0.02579155 Polish
(...)

[vbnetkhio]

I agree with exclamation mark. This is applicable to a number of GEDmatch calculators. How much unnecessary texting could have been avoided if f. i. the “Iberian” component of K36 just would have been called „component 7“ (figuratively)?!

I meant "nothing" in the sense that is very hard to tell to which EHG/WHG/steppe ratio they correspond, especially after the calculator effect.
The Iberian component still peaks in Iberians, but people without Iberian ancestry can get it too, either because of the calculator effect, or as a proxy because of a lack of components.

[Peterski]

Sachsen-Anhalt kit from south-western part of Sachsen-Anhalt, which has no or very few Slavic toponyms (but Slavic admixture is there anyway):

Population
North_Atlantic 41.2 Pct
Baltic 32.61 Pct
West_Med 8.74 Pct
West_Asian 5.02 Pct
East_Med 9.88 Pct
Red_Sea 0.29 Pct
South_Asian 1.26 Pct
East_Asian -
Siberian -
Amerindian -
Oceanian 0.64 Pct
Northeast_African -
Sub-Saharan 0.35 Pct

=====

And here is that Thuringian's K13 (from eastern Thuringia):

Population
North_Atlantic 33.73 Pct
Baltic 37.38 Pct
West_Med 8.85 Pct
West_Asian 8.81 Pct
East_Med 8.33 Pct
Red_Sea 0.92 Pct
South_Asian 0.17 Pct
East_Asian -
Siberian 0.51 Pct
Amerindian 0.11 Pct
Oceanian 1.17 Pct
Northeast_African -
Sub-Saharan -

I meant "nothing" in the sense that is very hard to tell to which EHG/WHG/steppe ratio they correspond, especially after the calculator effect.
The Iberian component still peaks in Iberians, but people without Iberian ancestry can get it too, either because of the calculator effect, or as a proxy because of a lack of components.

The Iberian component also existed e.g. in Neolithic British samples.

And in general in Neolithic Farmer populations throughout Europe.

[Jana]

The GEDmatch kits I have for Saxony are not much different than the Leipzig academic average. So maybe you just found kits from a specific sub-region of Saxony.

No, we have kits from multiple regions of Saxony, and several from Czech border. Old average was innacurate and shit.

Leipzig samples look east German, not specifically upper Saxon. They can be from wherever and they will be put into trashbin in new update.