Genetic‑affinity plots (PCA) can produce different nearest‑neighbour ancient populations depending on whether modern or ancient samples define the axes; holding the marker set constant but changing the reference frame reorders apparent affinities. This is a practical demonstration that ancestry visualizations are reference‑dependent and not immutable facts about an individual.
— Calls attention to a weak point in public interpretations of ancestry claims: axis or reference choice can produce conflicting stories about who someone 'is related to', with social and political implications for identity and policy.
Davide Piffer
2026.09.26
82% relevant
Piffer reports that accounting for shared ancestry (i.e., population structure/reference relationships) reduces or reverses many climate–PGS associations — matching the existing point that reference panels, ancestry controls and portability drive whether ancient/modern PGS claims hold up.
Davide Piffer
2026.09.09
100% relevant
Davide Piffer’s two PCA comparisons that keep the same SNP panel but swap modern vs ancient reference samples show his genome's nearest ancient neighbours change depending on the frame.
Davide Piffer
2026.09.07
85% relevant
The article is an explicit, worked example of the existing idea that ancestry proportions can 'flip' when the reference panel or modelling choices change: the author’s 23andMe genotype projects as '100% Roman' using a 4‑person Roman/Italic AADR source but would likely produce different mixtures under alternative reference sets or rotated sources — exactly the phenomenon captured by the existing idea.
Davide Piffer
2026.09.03
80% relevant
The article is an explicit, small‑scale demonstration that ancestry attributions depend on which ancient sources are included and in what order: Piffer runs nested qpAdm models on northern and central Italian individuals and shows that adding or dropping particular reference populations (e.g., Iranian Neolithic, WHG/EHG, Yamnaya proxies) changes which components are retained — directly exemplifying the 'reference‑panel ancestry flips' problem.