If you trust that article you link, they’ve found 300 genes affecting height, together accounting for 15% of the variance. Oddly, they don’t talk about the variance of the putative IQ genes, but the past hundreds of putative IQ genes have had similar aggregate effects. One explanation of the similar statistics of IQ and height and their tiny effect sizes is that they’re both failures of the same statistical methods. “In fact, there have been no common IQ polymorphisms which have been replicated.” A few days after that is a thread on LW about failure to replicate. Plomin is cited as someone cleaning up the hype, so maybe he is more careful in his results of last month (that you cite).
If we know anything, it’s that there are no common alleles of large effect for IQ. Maybe most variance in height and IQ is due to mutational load. Large effects are quickly selected away, but small effects can’t can’t be corrected because there are so many other small effects competing to be corrected. Maybe the observed effects of 1/1000 of the variance are what we should be seeing.
But then maybe we don’t have to know anything. If there are no positive variants, if the relevant consideration is number of negative variants, then it may be possible to genetic engineering by always choosing the most common allele, not knowing what any of them due.
If you trust that article you link, they’ve found 300 genes affecting height, together accounting for 15% of the variance. Oddly, they don’t talk about the variance of the putative IQ genes, but the past hundreds of putative IQ genes have had similar aggregate effects. One explanation of the similar statistics of IQ and height and their tiny effect sizes is that they’re both failures of the same statistical methods. “In fact, there have been no common IQ polymorphisms which have been replicated.” A few days after that is a thread on LW about failure to replicate. Plomin is cited as someone cleaning up the hype, so maybe he is more careful in his results of last month (that you cite).
If we know anything, it’s that there are no common alleles of large effect for IQ. Maybe most variance in height and IQ is due to mutational load. Large effects are quickly selected away, but small effects can’t can’t be corrected because there are so many other small effects competing to be corrected. Maybe the observed effects of 1/1000 of the variance are what we should be seeing.
But then maybe we don’t have to know anything. If there are no positive variants, if the relevant consideration is number of negative variants, then it may be possible to genetic engineering by always choosing the most common allele, not knowing what any of them due.