Breed‑average GWAS misreads ancient behavior

Using breed‑average behavioral scores to find DNA variants captures between‑breed differences (ancestry, environment, breeder practices) rather than individual causal effects; porting those variant sets into ancient genomes can therefore overstate selection on behavior. The dog study (Ko et al.) reports lower genetic 'fear' scores in ancient dogs vs wolves, but much of the signal traces to population structure and a single site, so behavioral claims are fragile. — This warns researchers, reporters and the public not to take genetic 'behavior' differences in ancient samples at face value and highlights broader risks when polygenic scores developed in one social or breed context are used in another.