What's your take on the way the 4-Ancestors Oracle in Eurogenes K13 does my calculations?

[gıulıoımpa]

hello, i've been inactive for very long, happy to be here again (although with most certainity momentarily because of ...life ) ,

i am making this tread because i never noticed this , but my 4 ancestors oracle always want to put in one ancestor of the 4 as Ashkenazi Jewish or italian jewish.



https://i.ibb.co/x3q4ZkM/4-pop-approx.png (https://ibb.co/Kmx1nP9)



I've no Ashkenazi ancestry in my 23 and me, should the "ashkenazi" be read as sicilian like ancestry? i know calculators sometimes swap the two or can't quite pin point them apart exactly. Speaking of Sicily my surname is actually from Naples but can also present in Sicily, although i have no ancestry from there (in the short run, and that i know of ).

For people that have it (not everybody) i need your knowledge, thanks for your input if i'll see you down here in the near future.

[JamesBond007]

Hiya, mate, long time no see. My qualifications, if you can all them that, is I took some biology/genetics in college and I've spent years with these calculators. I'm vaguely aware that Ashkenazis seem to plot somewhere in the vicinity of some Southern Europeans but common sense says Israel lies east of Italy and K13 is North/South calculator so to add East/West dimensions it would seem to me you would use K15 even though K13 seems more 'popular' that is not argument for it legitimately.

Suggestions that are not set in stone but seem worth trying to me :

The original K15 seems lackluster due to mostly limited samples sets or whatnot so you can try K15 on :

1.)genoplot.com

2.)AdmixtureStudio (only runs on Windows and perhaps macs)

3.) K15 Vahaduo updated I think there is an option now to reduce it to four populations but if you want a more even break you can use the source data in R or RStudio or whatnot.

I will quickly demonstrate what I mean, to a limited extant, real quick. I consider myself British ethnically but since K13 is straight North/South it gives me some 'Irish' and 'Welsh' crap. BTW, I don't take any of these 4 way things as gospel I pay more attention to the single 1 population and almost ignore everything else :

K15 from genoplot :

1.)England + England + Norway + South Dutch

********************
1)England 2.91
2.) Dutch 3.13
3.)Scotland

From K13 genoplot :

1.)Welsh + Welsh + Scottish + North German 2.70

2.)Welsh + Welsh + Irish + North German 2.77

1 .)English 3.91
2.)Irish 3.98
3.)Welsh 4.13

^ sorry , but there is no way I'm 50% Welsh which goes to show this is a north/south calculator like I said. I can entertain the idea of being 25% Irish or 25% Welsh but 75% Welsh and Irish or Scottish is too much man IMHO.

Also, if you notice the single population breakdown for K15 seems more logical, at least for me, too. I dunno, all I know is what works best for me so your mileage may vary. Like I said all I can do is make suggestions but I can't proffer them with 'authority'. However, it seems to me you might be using the wrong tool for the job.

What about Jtest ? Did you try that ? That would perhaps the best tool for it in the end I think it is based on EUtest and K15 is basically EUtest version 2.

[gıulıoımpa]

Hiya, mate, long time no see. My qualifications, if you can all them that, is I took some biology/genetics in college and I've spent years with these calculators. I'm vaguely aware that Ashkenazis seem to plot somewhere in the vicinity of some Southern Europeans but common sense says Israel lies east of Italy and K13 is North/South calculator so to add East/West dimensions it would seem to me you would use K15 even though K13 seems more 'popular' that is not argument for it legitimately.

Suggestions that are not set in stone but seem worth trying to me :

The original K15 seems lackluster due to mostly limited samples sets or whatnot so you can try K15 on :

1.)genoplot.com

2.)AdmixtureStudio (only runs on Windows and perhaps macs)

3.) K15 Vahaduo updated I think there is an option now to reduce it to four populations but if you want a more even break you can use the source data in R or RStudio or whatnot.

I will quickly demonstrate what I mean, to a limited extant, real quick. I consider myself British ethnically but since K13 is straight North/South it gives me some 'Irish' and 'Welsh' crap. BTW, I don't take any of these 4 way things as gospel I pay more attention to the single 1 population and almost ignore everything else :

K15 from genoplot :

1.)England + England + Norway + South Dutch

********************
1)England 2.91
2.) Dutch 3.13
3.)Scotland

From K13 genoplot :

1.)Welsh + Welsh + Scottish + North German 2.70

2.)Welsh + Welsh + Irish + North German 2.77

1 .)English 3.91
2.)Irish 3.98
3.)Welsh 4.13

^ sorry , but there is no way I'm 50% Welsh which goes to show this is a north/south calculator like I said. I can entertain the idea of being 25% Irish or 25% Welsh but 75% Welsh and Irish or Scottish is too much man IMHO.

Also, if you notice the single population breakdown for K15 seems more logical, at least for me, too. I dunno, all I know is what works best for me so your mileage may vary. Like I said all I can do is make suggestions but I can't proffer them with 'authority'. However, it seems to me you might be using the wrong tool for the job.

What about Jtest ? Did you try that ? That would perhaps the best tool for it in the end I think it is based on EUtest and K15 is basically EUtest version 2.

thanks for the thorough answer. ok , i seem to get it now also because of he references they use for the the calc it is much more efficient in North vs south kind of divide. I also agree that K15 seems more reasonable. i will surely try the option you cited. J test i did it long time ago, since i am Haplo J2 , but i forgot. i may re-do it just now.

...and ok i did it now among one of the four population approximation in the J test in most result there's always "Druze" and then a combination of either north italian - greek - tuscan. interesting to say the least. evidently i've got one strongly east pulling 25 percent of genes in me.

what i like the most among those is the 3 population approximation : 50% GR +25% GR +25% Tuscan @ 5.244004, this i think reflects what i am pretty well or at least closer than others... i even get (in the J test) the closest pop approx as GR rather than southern italian. so, that's it . me being the italo mix that i am (mainly southern with a bit of other inputs from north and center/south italy ) i think messes up a bit the calcs anyways.

thanks for the answer!

[Dušan]

hello, i've been inactive for very long, happy to be here again (although with most certainity momentarily because of ...life ) ,

i am making this tread because i never noticed this , but my 4 ancestors oracle always want to put in one ancestor of the 4 as Ashkenazi Jewish or italian jewish.


It doesnt mean that you have that ancestry, but 25% of genetic average of certian ethnicity is closest to your autosomal genetics in combination with other three quarters.

Mine:

Using 4 populations approximation:
1 Erzya + Moldavian + Sardinian + Ukrainian_Belgorod @ 3.963264


In reality, I dont have any of these ancestries.

[gıulıoımpa]

It doesnt mean that you have that ancestry, but 25% of genetic average of certian ethnicity is closest to your autosomal genetics in combination with other three quarters.

Mine:

Using 4 populations approximation:
1 Erzya + Moldavian + Sardinian + Ukrainian_Belgorod @ 3.963264


In reality, I dont have any of these ancestries.

understood. brain-fart moment for me then. i still do think my mix among italians can confuse the calcs sometimes.

[Lucas]

The original K15 seems lackluster due to mostly limited samples sets or whatnot so you can try K15 on :

1.)genoplot.com

2.)AdmixtureStudio (only runs on Windows and perhaps macs)

3.) K15 Vahaduo updated I think there is an option now to reduce it to four populations but if you want a more even break you can use the source data in R or RStudio or whatnot.

.

K15 (and K13) on Genoplot and AdmixtureStudio is also taken form K15 (and K13) Updated Vahaduo but misses some recent updates. Just for a note.

[grabielx]

The regular is an epic fail of an Oracle (for my case)

Using 4 populations approximation:
++++++++++++++++++++++++++++++++++++++++++++++++++ ++++++++++++++++++++++++++++++++++
1 Cyprian + French_Basque + Italian_Abruzzo + Southeast_English @ 1.384872
2 French_Basque + Irish + Lebanese_Christian + Tuscan @ 1.448122
3 Cyprian + French_Basque + Italian_Abruzzo + Orcadian @ 1.511925
4 Cyprian + French_Basque + Italian_Abruzzo + West_Scottish @ 1.514584
5 Cyprian + French_Basque + Italian_Abruzzo + North_Dutch @ 1.547801
6 Cyprian + French_Basque + Italian_Abruzzo + Southwest_English @ 1.565689
7 French + Italian_Abruzzo + Italian_Abruzzo + Spanish_Aragon @ 1.581117
8 French_Basque + Lebanese_Christian + North_Dutch + Tuscan @ 1.594694
9 French_Basque + Lebanese_Christian + Tuscan + West_Scottish @ 1.608546
10 Cyprian + French_Basque + Irish + Italian_Abruzzo @ 1.636616
11 French_Basque + Lebanese_Christian + Orcadian + Tuscan @ 1.658047
12 French_Basque + North_German + Sephardic_Jewish + South_Italian @ 1.707150
13 Cyprian + French_Basque + North_Dutch + West_Sicilian @ 1.708529
14 Cyprian + French_Basque + West_Scottish + West_Sicilian @ 1.732304
15 Cyprian + Danish + French_Basque + Italian_Abruzzo @ 1.750922
16 Cyprian + French_Basque + Irish + West_Sicilian @ 1.762707
17 South_Italian + South_Italian + Spanish_Aragon + West_Scottish @ 1.789104
18 Danish + French_Basque + Lebanese_Christian + Tuscan @ 1.809629
19 Cyprian + East_Sicilian + French_Basque + West_Scottish @ 1.816997
20 Italian_Abruzzo + Sephardic_Jewish + Southwest_English + Spanish_Aragon @ 1.839055

vs the genoplot.com (much more reasonable)
https://i.imgur.com/WnGdpHU.png

but again the reference populations for Italy appears to unreliable and southern shifted, but that's true for most calculators
https://i.imgur.com/sgH7vBB.png