What is ideal distance on gedmatch calcs/G25?

[Jana]

What do academic researches say about it?

[Impaler]

I think depends on the calculator. In my opinion between 1.3-1.8 would be ideal.

[Luke35]

I have had good results anywhere between 1.5 and 3.

[ph2ter]

Gedmatch and G25 are not used in academic research. They use qpAdm and similar tools and they analyse row DNA data (SNPs) using statistical and probabilty functions.

Distance depends on calculator: how many components it has and what is the possible range of values.
Distance is actually distance between two points in a space with n-dimensions where n is the number of components (for example in G25 there are 25 dimensions, in K13 there are 13 dimensions).

[Jana]

Gedmatch and G25 are not used in academic research. They use qpAdm and similar tools and they analyse row DNA data (SNPs) using statistical and probabilty functions.

Distance depends on calculator: how many components it has and what is the possible range of values.
Distance is actually distance between two points in a space with n-dimensions where n is the number of components (for example in G25 there are 25 dimensions, in K13 there are 13 dimensions).

So what is ideal distance on these non academic tools in your opinion?

[ph2ter]

So what is ideal distance on these non academic tools in your opinion?

It is ideal to get as low as possible distance, but at the same time to use a maximum of only 4 or 5 populations.
If you get very low distance but you at the same time are using many populations (or references) it is called overfitting - and it is not good.

[gixajo]

Gedmatch and G25 are not used in academic research. They use qpAdm and similar tools and they analyse row DNA data (SNPs) using statistical and probabilty functions.

Distance depends on calculator: how many components it has and what is the possible range of values.
Distance is actually distance between two points in a space with n-dimensions where n is the number of components (for example in G25 there are 25 dimensions, in K13 there are 13 dimensions).

Distance in G25 results are "measured" in an space with n dimensions or is measured in a conversion make to a simplier 2 dimensions plane, a rectangular coordinate system?

And, do you know wich component distance makes reference every G25 coordinate?

[ph2ter]

Distance in G25 results are "measured" in an space with n dimensions or is measured in a conversion make to a simplier 2 dimensions plane, a rectangular coordinate system?

And, do you know wich component distance makes reference every G25 coordinate?
Distance is measured in a space with n orthogonal dimensions. It is normal Euclidean distance, not some projection in 2D plane or whatever.

G25 coordinates are taken from some form of transformation or from PCA made from unknown components (I believe).
Somewhere Davidski has written about this process but I forgot the details.

[gixajo]

Distance is measured in a space with n orthogonal dimensions. It is normal Euclidean distance, not some projection in 2D plane or whatever.

There is not much difference between those 2 ways,at least the proportions are maintained in a certain way but ok.

[/QUOTE]G25 coordinates are taken from some form of transformation or from PCA made from unknown components (I believe).
Somewhere Davidski has written about this process but I forgot the details.[/QUOTE]

I think keeping that as secret is a great idea.:thumb001:

Some coordinates seem to indicate a certain relationship with some known components, but this is something than anyone could know simply playing with them.

[Kenshiro]

125718
For only 2-3 population i think this Is a very good distance, i love my dna it's litterally between the Western Roman Empire and the Eastern Roman Empire, i'm in the good fit between this two beautiful history, In fact, I'm very interested in Greek-Roman paganism, I also appreciate other cultures obviously, but I feel a very particular connection with ancient Byzantium, ancient Athens and ancient Rome