Comment by Solstinox

These are the types of assumptions and blindspots that lead to unintended negative effects.

Your approach assumes that the distribution of those SNPs at the population level is immaterial.

What makes an allele attractive at one frequency can expose new liabilities at a different frequency.