Is DNAgenics legit at all? Weird results

[Pheonix997]

What does everyone here make of DNAGenics? My results seemed pretty wild compared to 23andme and Ancestry (see attached screenshots). I'll also add my GEDMATCH eurogenes K13.

My ethnicity is half-Danish, half-Slovenian. But DNAGenics for example gave me 3.3% Northern Indian and some trace African and 3.3% Arab.

Is it more accurate but less biased than 23andme/ Ancestry ? Or should I disregard these results as inaccurate? Is there some way I can investigate further to find out whether I really have some small amounts of Indian, Arab, and African ancestry? I understand the 3% Arab due to the Roma-era Anatolian migration to the Balkans, including Slovenia. But the Indian seems a little exotic and was unexpected.

[Luke35]

Hey man,

Your AncestryDNA, and especially your 23andMe match basically exactly what you describe of your known ancestry. If you had Indian, Middle Eastern, or SSA - it would show on those two tests.

Your Eurogenes K13 admixture also looks totally copasetic with your background. If you really had 6.6% WANA + South Asian ancestry (~great-great grandparent level), I would expect to see a small boost in one or all of the West Asian, Red Sea, and South Asian percentages. Your admix percentages on K13 fit within the expected range for your background.

[Gannicus]

I used them to merge together all my DNA files. I don't really care for the admixture calculators on there. The reports on haplogroups are useful. They give more detail on y-haplogroups and mtdna than 23&Me.

[Figaro]

Welcome, nice to see another with partial Danish ancestry.

[Gannicus]

As for the subtle MENA admixture, it may be from the Anatolian Neolithic Farmers, or it could be from ancient MENA admixture post neolithic to the Roman period. It could be a combination of both. DNA tests from the major companies may not reveal that. It's really hard to tell without an Animus Machine from Assassin's Creed.

[ScandinavianCelt]

comparing IllDNA vs DNAGenics on chatgpt:

Answer:

Both companies:

Take your raw SNP data.
Impute missing markers.
Project you into the G25 PCA space.

Each company uses a different projection method.

IllustrativeDNA's coordinates are generally designed to be:

Stable
Conservative
Optimized for ancient population modeling

DNAgenics tends to produce coordinates that:

Show more drift toward minor ancestry components
Sometimes exaggerate eastern shifts
Often produce more "interesting" ancient fits

[Pheonix997]

Deleted.

[Pheonix997]

Hey man,

Your AncestryDNA, and especially your 23andMe match basically exactly what you describe of your known ancestry. If you had Indian, Middle Eastern, or SSA - it would show on those two tests.

Your Eurogenes K13 admixture also looks totally copasetic with your background. If you really had 6.6% WANA + South Asian ancestry (~great-great grandparent level), I would expect to see a small boost in one or all of the West Asian, Red Sea, and South Asian percentages. Your admix percentages on K13 fit within the expected range for your background.

Ok thank you. Sorry for the late reply. Is it possible the middle-eastern ancestry could be getting lumped under the "Slovenian" umbella in the same way the pre-slavic balkan does? Like 23andme doesn't differentiate slavic vs. balkan for slovenians, because that admixture is already baked in. It's the same for Scandinavians - 23andme doesn't say how much traces of uralic ancestry you might have as, say, a Swede or Norwegian. They'll just call it "swedish/ Norwegian".

On my gedmatch k13 (updated with 23andme chip upgrade) I now realize now my South Asian is basically within the expected averages, however My East-Asian is undoubtedly elevated (see screenshot).

What do you make of this? I'm thinking it may be artificially lumping what would otherwise be my Siberian and Amerindian under East-Asian. Because when I tally those 3, I fall more within the expected averages for a European (see 2nd screenshot). What do you think? I'll also post a screenshot of my distances and admix portions from Vahaduo for more information. You can see I have uralic (saami/ finland) and mongolian sources, but they don't spit out in the calculator.

[mariusz99]

No, not legit.