Can you provide the citation to a paper published in for example Nature, Science Advances, Cell, etc, that documents G25 has been used to model the populations that were analyzed in those respective studies. G25 being designed based on SMARTPCA method (Nick Patterson) does not mean that Nick Patterson endorses or suggest he had a hand in developing it.
You are trying to justifying your support for G25 by what is referred to in the circles I run in as "name dropping".
G25, as with any other model developed by non academicians (we are talking about people with training in Mathematics, Statistics, computer science) in my view is only as good as how robust it is to the extant published research. If there are 5, then 10 published papers from various research teams in various leading academic journals that document A and then G25, or Dodecad, or MDLP document B, or C, etc. Then I will defacto question something regarding the non-Professional calculators.
Now to be fair, I think the G25 Eurogenes is much better than the Old Eurogenes K13/K15 models regarding Southern Europeans. Dodecad, from my perspective, has always produced results that are fairly consistent with what has been published. I would say G25 has in terms of "single distances" performed better for me versus Eurogenes K13/K15. I would add that the problem with the Eurogenes K13/K15 models might not be a problem with the models themselves, but rather, the samples used in those spreadsheets. Regarding G25 and modeling source populations, I have no clear opinion yet. I will note that I only have G25 simulated coordinates.
As for modeling Southern Italians, Central and Northern as well, It is my view that the Dodecad 12B K8 model that Jovialis produced is the most accurate for me personally, and other individuals that are Americans of Italian ancestry with all of their ancestry from 1 of the 20 political regions of modern Italy as well as those Italians here whose ancestors are ethnically Italian with long standing histories in 1 of those same 20 political regions of Modern Italy. The Jovialis K8 model also matches up with several recent academic papers related to Italian genetics.