Spatial Ancestry analysis (SPA) from UCLA

Discussions concerning Autosomal DNA.

Moderator: CeCeMoore

User avatar
Posts: 287
Joined: Mon Apr 23, 2012 6:15 pm
PostPosted: Thu Aug 22, 2013 8:06 pm
Thank you for explaining the phenomenon better.
You might be interested in this alternative reference files for SPA:
http://bga101.blogspot.com/2012/11/spat ... model.html
DNA/Admixture Central Europe (Alps, Tyrol, Dolomites, Raetia); Y-DNA J2a-L1064, J2a-L210, R1a-M17, R1b-U106 (L48-); mtDNA J1b1b, J1c1d, U5a2b2, U5b1b1. Projects : J2-M172, J2a-PF5197, Alpine DNA, ISOGG Wiki

Posts: 55
Joined: Fri May 25, 2012 8:11 pm
Location: United States of America
YDNA:
R-DF99/S11987
MtDNA:
H3ap
PostPosted: Thu Aug 22, 2013 10:27 pm
ChrisR wrote:Thank you for explaining the phenomenon better.
You might be interested in this alternative reference files for SPA:
http://bga101.blogspot.com/2012/11/spat ... model.html


No trouble, it's something that I'm still trying to understand and wish UCLA would explain better.

With the models available one has to choose which file to use based on "prior knowledge" - choose the file that is relevant to oneself. Those files from Eurogenes were designed for those whose ancestries were not adequately represented by the UCLA provided model (ex. Central & Eastern Europeans and Finnish people).

I had run the Eurasian cline but since my ancestry is not relevant to its design I did not get meaningful results no matter how many times I ran it (always places me in Bavaria in both the two way mix and single modes, I know from prior knowledge via previous tests and genealogy work that I am not 100% Bavarian so it's not surprising that the model did not produce relevant results for my case).
R-P312 clades tested for: DF99+ P312+ DF27- DF19- U152- L21- L459- Z245- L238- Z2245- Z2247-
Ysearch: VNSGW
mtDNA: H3ap
Gedmatch ID: M033622
User avatar
Posts: 287
Joined: Mon Apr 23, 2012 6:15 pm
PostPosted: Fri Aug 23, 2013 5:16 pm
Barreldriver wrote:I had run the Eurasian cline but since my ancestry is not relevant to its design I did not get meaningful results no matter how many times I ran it (always places me in Bavaria in both the two way mix and single modes, I know from prior knowledge via previous tests and genealogy work that I am not 100% Bavarian so it's not surprising that the model did not produce relevant results for my case).

Thats strange and interesting because in Nov. 2012 I had run SPA on South Tyrolean samples (all from the same valleys and with deep genealogy from there). To give an idea of the calculated origin (UCEU = UCLA Europe, EWEA = Eurogenes_West_Eurasia):
Sample 1: St. Gallen, E. Switzerland (UCEU); SW. Bavaria, S. Germany (EWEA)
Sample 2: Haute-Marne, E. France (UCEU); Vosges, E. France (EWEA)
Sample 3: Vaud, W. Switzerland (UCEU); South Tyrol, N. Italy (EWEA)
Sample 4: Zürich, NW. Switzerland (UCEU); S. Baden, SW. Germany (EWEA)

I did only one run. As you can see the results where not totally different between UCEU and EWEA admixture models and also give a possibly good prediction, as there was much migration from Southern German regions to Central Alpine regions in the Middle Ages, documented by historical sources.
But what is interesting: your sample by EWEA was placed massively in a south western direction. For South Tyrol samples there is also a slight prediction difference in a Eastern direction from EWEA, which my come frome the focus of EWEA creator on Eastern European (Balto-Slavic) populations and admixture.

I may redo the analysis in the future by running those and additional samples multiple times, as you suggest.
DNA/Admixture Central Europe (Alps, Tyrol, Dolomites, Raetia); Y-DNA J2a-L1064, J2a-L210, R1a-M17, R1b-U106 (L48-); mtDNA J1b1b, J1c1d, U5a2b2, U5b1b1. Projects : J2-M172, J2a-PF5197, Alpine DNA, ISOGG Wiki

Posts: 55
Joined: Fri May 25, 2012 8:11 pm
Location: United States of America
YDNA:
R-DF99/S11987
MtDNA:
H3ap
PostPosted: Fri Aug 23, 2013 6:34 pm
I do need to finish at least 60 runs with the Eurogenes .model just to be more certain in where it is placing me (and for consistency) but so far the points are alternating between nearby spots contained within the state of Bavaria. After this is done, I was going to try to find an average between the UCLA and Eurogenes results. But such an average may not be meaningful considering "Argument to Moderation" or Fallacy of the Golden Mean - truth does not lay in the middle ground between correct conclusions and incorrect conclusions (can I arrive at a true conclusion by looking for a middle ground between the possibly correct or incorrect results of UCLA or the possibly correct or incorrect results from Eurogenes). I would rather learn how to make my own .model files and not be dependent upon the Northwest centered files from UCLA and the Central and Eastern centered files from Eurogenes.
R-P312 clades tested for: DF99+ P312+ DF27- DF19- U152- L21- L459- Z245- L238- Z2245- Z2247-
Ysearch: VNSGW
mtDNA: H3ap
Gedmatch ID: M033622
User avatar
Posts: 287
Joined: Mon Apr 23, 2012 6:15 pm
PostPosted: Fri Aug 23, 2013 9:50 pm
Barreldriver wrote:I would rather learn how to make my own .model files and not be dependent upon the Northwest centered files from UCLA and the Central and Eastern centered files from Eurogenes.
Yes an alternative .model admixture set or an update of UCEU and/or EWEA should be possible (after a year more reference samples should be available). If you or someone else will release something, I'm interested in testing it.
DNA/Admixture Central Europe (Alps, Tyrol, Dolomites, Raetia); Y-DNA J2a-L1064, J2a-L210, R1a-M17, R1b-U106 (L48-); mtDNA J1b1b, J1c1d, U5a2b2, U5b1b1. Projects : J2-M172, J2a-PF5197, Alpine DNA, ISOGG Wiki
Previous

Return to Autosomal DNA

Who is online

Users browsing this forum: No registered users and 1 guest

cron