This is not correct reasoning about genetics. I've seen a lot of comments with this same incorrect claim on HN recently. In a complex trait like IQ, there is no reason to believe that stacking all of the alleles that each are individually associated with small gains in IQ will result in a large cumulative effect on IQ, let alone an additive effect. It could easily result in decreased IQ-- there are plenty of examples of negative epistatic interactions between loci. Even if you had all the individuals in the world in your GWAS, you couldn't estimate all of the possible pairwise interactions between loci, let alone fit the higher-order terms. The only way to really measure the effect of a combination of alleles is to find a subpopulation with that combination. That hasn't stopped people from trying to fit polygenic models for IQ based on all of these tiny effects, and unsurprisingly the best model only explains 10% of the heritability of the trait. I don't think that model has been carefully validated in a new cohort either, so I wouldn't be surprised if the true predictive power was lower.
dooglius|6 years ago
est31|6 years ago