Admixtures are being extracted incorrectly on GEDmatch : 23andme

My understanding is that what’s supposed to happen is the script paints your chromosome, gives each population a unique color, and then calculates the proportions used for each population. The trouble is at least for my DNA kit and my father it’s only looking at the first 1 pixel column of the painted chromosome and omitting the remainder of the data.

Provided you have Photoshop your can open the Histogram panel, select each color, record the pixel count, sum the total pixel count, and then do the calculation yourself. Which takes a few hours and you get a similar but fundamentally different result. Because after all, a little bit of noise is all the separates me from being correctly ID’d as ‘northern European + eastern European + a quarter-ish Arab’ as opposed to just ‘northern European + eastern European’ or possibly ‘northern European + eastern European + west mediterranean’.

r/23andme - Bug Report: Admixtures are being extracted incorrectly on GEDmatch
r/23andme - Bug Report: Admixtures are being extracted incorrectly on GEDmatch

Actual RAW(I suspect they’re doing something to eliminate and redistribute “noise”) results:

north_atlantic_% 32.17
baltic_% 27.66
west_med_% 8.03
west_asia_% 11.4
east_med_% 2.96
red_sea_% 7.75
south_asia_% 2.63
east_asia_% 0.81
siberia_% 2.63
native_american_% 0.34
oceania_% 0.47
northeast_africa_% 0.34
sub_africa_% 3.48

Read more here: Source link