>>10924995By multiple genes being needed for good intelligence, with each of them being necessary for a strong effect. (e.g a single one increasing your IQ by 0.1, two by 0.5, six by 10, seven by 12, eight by 20, nine by 30 and ten by 80) The phenotype of a recessive gene increases with the gene prevalence. 1% requires 10% of the resistant gene. 20% resistant gene will result in 4% of the resistant phenotype, 40% of resistant gene will result 16% percent resistant phenotype. This effect may get even more dramatic with multiple genes needed for strong resistance.
So full resistance is expected to appear rather rapidly after a long period of slow accumulation of resistant mutations.
Assortative mating can also increase it rather suddenly. (and assortative mating has been observed among autistic people)
With ten genes, five with a recessive resistant allele, five with a dominant one, the rarity of a fully resistant phenotype will be (x^2)^5*(1-x^2)^5, which gives us (assuming 100% random mating) only 2.476099 × 10^-14, with 10% of each resistant allele, exceedingly unlikely to happen even in one person in the world. With 20% of each allele, this rises to 6.19174×10^-10, less than one person in a generation. With 25% prevalence, it's 1.52859×10^-8, several people per billion. 30% gives us 2.03734×10^-7, about one in 50 million. 35% roughly 1.7 in a million. 40% one in 88818. 45% one in 17789. 50% one in 4315.
60% one in 359
65% one in 142
66% one in 117
67% one in 97
68% one in 81
69% one in 67
70% one in 57
This is how polygenic, highly advantageous traits may appear rather suddenly, seemingly out of nowhere, arising from a particular combinations of genes that were never particularly uncommon.
