Vahaduo Online tool for G25

[Björnsson]

It's not too bad at all. Literally, you can start with cutting and pasting existing calculators on this sub-forum into "source," then cut and paste other coordinates into "target," go to Multi and click "run."

Gedmatch should have it as an embedded app. I'm not interested in programming. I use Windows for a reason. Then again, this is kind of like AutoCAD and game level creation.

[firemonkey]

The PCA runner is interesting , but I can't see the point of downloading a screenshot from it. I would if I could still see the rectangle with the population in it .

[evon]

This is the Linux of DNA apps. Nothing more user-friendly?

For Linux standard it is very user friendly, and I use Linux on my now neglected desktop computer. The only thing I would say is negative is that I am unable to use the individual samples list due to the size of the list, but otherwise it is great. The "single" output didn't work so well, but I was not expecting it to, as G25 is no miracle tool.

[evon]

Gedmatch should have it as an embedded app. I'm not interested in programming. I use Windows for a reason. Then again, this is kind of like AutoCAD and game level creation.

Since G25 is a paid tool, I don't think Gedmatch will implement such a feature any time soon. Gedmatch is also receiving allot of criticism at the moment due to security issues and privacy issues. So I have a feeling that unless they can fix these issues soon, gedmatch will be history.

[evon]

The only thing I would say is negative is that I am unable to use the individual samples list due to the size of the list, but otherwise it is great. The "single" output didn't work so well, but I was not expecting it to, as G25 is no miracle tool.

I got it working with some help from Finn the solution is to open the document in google docs and then copy paste from there. But I must say I see the same old "problems when I run the individual versus the pop avg:

Pop avg:

 

Target: Grandmother_scaled
Distance: 0.9933% / 0.00993252
35.6 Norwegian
27.2 Icelandic
12.4 Irish
12.0 Shetlandic
8.2 Lithuanian_VZ
2.4 Italian_Northeast_o
1.0 Sorb_Niederlausitz
0.4 Berber_MAR_TIZ
0.2 Dusun
0.2 Ju_hoan_North
0.2 Lithuanian_SZ
0.2 Murut

With the pop avg it is also way off, as we dont have any Icelandic (back in the 13-1400's only), Shetlandic, Irish, Lithuanian, Italian etc ancestry. In fact the only correct one, of these, is the Norwegian, although it should be much higher than 35.6. The rest is just statistical garbage, one might be tempted to say it is worse than Oracle..

With individual samples (spoiler, it is much worse):

 

Target: Grandmother_scaled
Distance: 0.2538% / 0.00253806
Aggregated
40.2 Irish
14.2 German
10.6 Welsh
8.8 Dutch
5.6 Scottish
4.8 Estonian
4.6 Norwegian
4.0 Latvian
2.2 Karelian
1.8 Icelandic
1.8 Shetlandic
0.6 Finnish
0.4 Murut
0.2 Ju_hoan_North
0.2 Lithuanian_VZ

Again, the only "correct" ethnic groups, from among the list, are Norwegian, German and possibly some Scottish and Dutch that I have yet to find in my tree (I say possibly, because these were major immigrant groups to Norway). The rest is statistical garbage, as I see in both instances they fail to correctly assign her Balkan-west Asian and south Asian, instead giving her eastern European, Italian, north African and central African etc

[Ruderico]

We already have a thread about Vahaduo: https://anthrogenica.com/showthread....e-tool-for-G25

It is a very good tool. I had no idea how to use G25 before discovering it.

Edit: Merged threads by request

[Dorkymon]

The only thing I would say is negative is that I am unable to use the individual samples list due to the size of the list, but otherwise it is great.

https://anthrogenica.com/showthread....896#post602896

[JMcB]

I broke it even further, for those who are interested, by increasing the number of cycles to 1000 (nMonte's default).

https://drive.google.com/open?id=1EO...x9tTV2hCfbBUPN

Hello Dorkymon,

Just out of curiosity, do you know what the default number of cycles are that Vahaduo runs on?

[Dorkymon]

Hello Dorkymon,

Just out of curiosity, do you know what the default number of cycles are that Vahaduo runs on?

Hi, the cycles variable seems to be set to about 1000, although it's somehow determined dynamically. There isn't a static value set in the code itself.
However, I can tell you that I've changed it manually to 1000 and more, which produces the same result.
Lower than that worsens the fit.

If for some reason you need to change it, that can be done in row 476.

[JMcB]

Hi, the cycles variable seems to be set to about 1000, although it's somehow determined dynamically. There isn't a static value set in the code itself.
However, I can tell you that I've changed it manually to 1000 and more, which produces the same result.
Lower than that worsens the fit.

If for some reason you need to change it, that can be done in row 476.

Hello Dorkymon,

Thanks for the information. Unfortunately, I do everything on an iPad, so I can’t download the file and use it like those who have a computer. I do really like Vahaduo’s format and ease of use but because it doesn’t allow for any penalty, I don’t use it much because I usually get better results using pen 0.001.