Dalton Conley and Jason Fletcher in Nautilus:
[L]et us ask what is perhaps the most controversial question in the human sciences: Do genetic differences by ancestral population subgroup explain observed differences in achievement between self-identified race groups in the contemporary United States over and above all the environmental differences that we also know matter? In their best-selling 1994 book, The Bell Curve: Intelligence and Class Structure in American Life, Richard Herrnstein and Charles Murray indeed made the argument that blacks are genetically inferior to whites with respect to cognitive ability. Their “evidence,” however, contained no molecular genetic data, and was flawed as a result. But today we have molecular data that might potentially allow us to directly examine the question of race, genes, and IQ. We raise this pernicious question again only to demonstrate the impossibility of answering it scientifically.
If Herrnstein and Murray redux wanted to proceed, perhaps an obvious way would be to examine whether all the small differences across the genomes of the average black and average white person in a dataset “add up” in a way that suggests that one group has, on average, genetic signatures that predict higher levels of important phenotypes, such as educational attainment. There are at least two ways of “adding up” genomes. The first is to use polygenic scores. The second is the use of principal components. Both have serious drawbacks.
A polygenic score is a single number that captures the sum total of thousands of little effects in the genome on a given trait. It is constructed by running a million or more separate comparisons for each place along the 23 pairs of chromosomes where there is variation (i.e. you have an A-A and I have a G-A) measured in a dataset. When summed, these measures can predict—albeit noisily—the distribution of a given phenotype in the population. The best performing polygenic score to date is for height. A single number calculated from someone’s DNA can explain about 50 percent of the variation in actual height in the population. A score that has been developed for education (and cognitive ability) can explain about 7 percent of the variation in years of schooling, according to a 2016 Nature study, and that score has since been refined to improve its predictive power. So while these are not explaining all of the genetic variation (we think height is about 80 percent genetic and education is at least 25 percent genetic), they do predict. Someone at the upper end of the education distribution is likely to get more than two more years of schooling on average than someone at the bottom of the pack (lowest 10 percent) in terms of his or her polygenic score.
As it turns out, however, these scores when developed for one population—say, those of European descent—fail to predict for other populations.
More here.