G25 PCA vs Vahaduo distances :

[JamesBond007]

These two results don't seem to match it is like Vahaduo distances Germanify me but the G25 PCA plot rightly puts me nearest to the Scots :



https://i.postimg.cc/5tG0mx2m/Polish-20220830-090230322.png

Distance to: Kevin_scaled
0.02829386 Dutch
0.02982029 Norwegian
0.03080579 Danish
0.03097093 Welsh
0.03119617 Irish
0.03153596 Scottish
0.03173767 English
0.03264251 Orcadian
0.03323573 French_Brittany
0.03424261 English_Cornwall

What is the reason for this discrepancy? I would say G25 is meant to be used in a PCA with wisely chosen samples otherwise it is trash or something.

[Nmiito]

I would say PCAs are actually more biased and inaccurate.

[JamesBond007]

I would say PCAs are actually more biased and inaccurate.

Did you know that G25 was conceived as a PCA technology so that everything else is an afterthought technically?

[Nmiito]

Did you know that G25 was conceived as a PCA technology so that everything else is an afterthought technically?

For the global population yes, not to differentiate between highly related Northwestern European populations.

[JamesBond007]

For the global population yes, not to differentiate between highly related Northwestern European populations.

Good point the keyword is global. Celtic vs Germanic coordinate PCA :


https://i.postimg.cc/prxcMr8R/Polish-20220830-150922503.png

^ looking English here not Scottish

[Batman_]

Same here, I get one Illyrian from Northern Albania sample first yet plot closer to one sample from Macedonia. The difference aren't huge but still weird. But I managed to find a PCA Map where it is more in line with the calculator distances and most calculators follow what PCA Maps show for most samples it seems, depends on what PCA Map you use too. One sample from Serbia which has a typical Illyrian profile clusters next to me and that's what I get first too for example.

[Batman_]

What I really learnt is that PCA Maps can be misleading, calculator distances are the most accurate.