Lightnews — Scholar-powered news

tewoh the cat @tewoh.bsky.social · 4d

ancestry hides my west african ancestry but 23 shows it (estimated at ~0.5%)
. i think of any result, its probably the most accurate percentage wise simply because i have no other group there it could confuse it with

1

tewoh the cat @tewoh.bsky.social · 4d

for sure. even in my case a lot of the inaccuracy is coming between groups where their dna are pretty similar so the model struggles to classify precisely

1 1

tewoh the cat @tewoh.bsky.social · 4d

regarding the original topic with racial lines, its very unlikely a dna test to label someone from an ENTIRELY separate region theyre not. if someone gets east asian and they are mostly European heritage, theres almost certainly something there

1 1

tewoh the cat @tewoh.bsky.social · 4d

yes i think as much as people want to praise the accuracy of dna results i still do believe they can miss and mislabel things. especially with acadian being added, it might be obscuring other information

1 1

tewoh the cat @tewoh.bsky.social · 4d

french american groups such as the quebecois, acadians, cajuns, etc. were often endogamous, so often results are more pronounced. while i am cajun, i am also a louisiana creole, and i also have lines outside of Louisiana on both my maternal and paternal sides

1

tewoh the cat @tewoh.bsky.social · 4d

similar with me, but a lot of my ancestry came over from the 1700s (often before). its probably inversely why theres so much variance

2 1

tewoh the cat @tewoh.bsky.social · 4d

this is a comparison for instance between my maternal grandmother and me. some of the regions are cut off (like 3 of them 2 irish one brittany). theres like 30 regions total here

1 1

tewoh the cat @tewoh.bsky.social · 4d

if ur getting 100% then if you arent then you likely will be. someone like me wouldn't be because theres a ton of noise

2 1

tewoh the cat @tewoh.bsky.social · 4d

however by doing so you might obscure other groups that also existed in the same region, or other ancestry. its a tradeoff

1

tewoh the cat @tewoh.bsky.social · 4d

i suspect some of it is that regarding french canadians alongside France's legal stance regarding DNA tests, it was likely simpler to build a sample size for Canadian French groups vs France itself

2 1

tewoh the cat @tewoh.bsky.social · 4d

(my uncle for instance gets 15% quebecoise. 2% for my dad, 0% for me.)

another case is that it claims i have no acadian from my maternal line, but my maternal grandmother is 21% acadian according to ancestry. its things like that which still make me question the reliability of the exact results

1

tewoh the cat @tewoh.bsky.social · 4d

and to be honest, when looking at my relatives dna tests, its most likely just that theres so many different regions it gets a lot of things potentially off because of just Noise. one can see my relatives and there will be entire separate regions upwards to the double digits

1 1

tewoh the cat @tewoh.bsky.social · 4d

(i was not able to capture north africa on ancestry. they list the percentage as 0% anyways)

1 1

tewoh the cat @tewoh.bsky.social · 4d

as a point this is ancestry vs 23andme, it may just be because the number of regions but a lot of the percentages vary for me. i also have historically varied a lot when compared to relatives, such as my father

2 1

tewoh the cat @tewoh.bsky.social · 4d

ancestry's update definitely significantly improved their results, the only thing is regarding me im not as sure how reliable they are for me simply just because just how much trusted ancestry services vary with my DNA results