Yvain's sarcastic summary: "Genetic differences explain 24% of between-national-populations differences in height and 8% of between-national-populations in BMI across Europe. Now that the only two massively polygenic traits that might vary among national populations have been successfully studied, I look forward to never having to read any further research of this sort ever again."
The paper: "Population genetic differentiation of height and body mass index across Europe", Robinson et al 2015 https://www.dropbox.com/s/azvli29ltd3crof/2015-robinson.pdf / http://moscow.sci-hub.bz/14c52b6fe833e179bcb170969145e32d/robinson2015.pdf ; excerpts:
"Across-nation differences in the mean values for complex traits are common1, 2, 3, 4, 5, 6, 7, 8, but the reasons for these differences are unknown. Here we find that many independent loci contribute to population genetic differences in height and body mass index (BMI) in 9,416 individuals across 14 European countries. Using discovery data on over 250,000 individuals and unbiased effect size estimates from 17,500 sibling pairs, we estimate that 24% (95% credible interval (CI) = 9%, 41%) and 8% (95% CI = 4%, 16%) of the captured additive genetic variance for height and BMI, respectively, reflect population genetic differences. Population genetic divergence differed significantly from that in a null model (height, P < 3.94 × 10−8; BMI, P < 5.95 × 10−4), and we find an among-population genetic correlation for tall and slender individuals (r = −0.80, 95% CI = −0.95, −0.60), consistent with correlated selection for both phenotypes. Observed differences in height among populations reflected the predicted genetic means (r = 0.51; P < 0.001), but environmental differences across Europe masked genetic differentiation for BMI (P < 0.58).
We reestimated the effects of each SNP in a within-family study design, which is unbiased by population stratification, and used these effect sizes to create a genetic predictor for both phenotypes (also termed a ‘profile’ or ‘polygenic score’) 36 .
There is no certainty that population stratification is completely controlled for in large-scale meta-analyses, and we thus repeated our analysis using (i) non-ascertained, unlinked (LD r 2 <0.1 and >1 Mb apart), common (minor allele frequency (MAF) >1%) HapMap 3 loci (~40,000 SNPs) and (ii) unlinked (LD r 2 <0.1 and >1 Mb apart), common (MAF >1%) HapMap 3 loci selected on the basis of their within-family association with each phenotype (~40,000 SNPs for both traits). This analysis provides an unbiased, genome-wide estimate, representing a lower limit of population genetic differentiation at common, unlinked loci. The maximum proportion of variance in a polygenic predictor attributable to population genetic differences was 24% (95% CI = 9%, 41%) and 8% (95% CI = 4%, 16%) for height and BMI, respectively, using 2,660 SNPs for height and 11,919 SNPs for BMI. For height, the largest proportion of population-level variance was captured by SNPs of low P value in the meta-analysis (Supplementary Fig. 6). For BMI, the continual addition of SNPs increased the proportion of populationlevel variance captured (Supplementary Fig. 6). For both traits, the among-population variation was greater in predictors that explained greater phenotypic variance (Supplementary Fig. 7). Our results were confirmed using the non-ascertained independent, genome-wide loci (height: 8.6%, 95% CI = 3%, 15.7%; BMI: 2.8%, 95% CI = 1.1%, 5.3%) and the set of independent, genome-wide loci selected on the basis of their within-family association (height: 11.9%, 95% CI = 4.5%, 21.8%; BMI: 8%, 95% CI = 3.4%, 14.7%). The lower among-population variance captured using non-ascertained loci reflects reduced prediction accuracy, likely due to the addition of a large number of loci with no detectable association. Subsequent results are presented using the predictor for each trait that captured the greatest amount of populationand individual-level variance (comprising 2,660 SNPs for height and 11,919 SNPs for BMI); however, the results remained the same irrespective of the SNPs selected (Supplementary Fig. 8). The predicted population genetic means for the traits are shown in Figure 1 alongside the observed values, estimated from an independent set of recently published data 25,37 accounting for trends over time.
Genetic differences among populations may occur by random, chance processes or through natural selection in the evolutionary past 19,38–46 . We thus compared our estimates to the values from a null quantitative genetic model of multivariate population differentiation 32,47 . We found strong evidence that the divergence of each trait was greater than expected under the neutral model (Fig. 2). The overall level of neutral genetic differentiation was small for both height (1.2%; 95% CI = 0.01%, 1.78%) and BMI (1.9%, 95% CI = 0.48%, 2.97), reflecting the average F ST (a measure of population differentiation due to genetic structure; Supplementary Note) of the SNP sets between the populations of 1% for height and 1.2% for BMI. Our results were confirmed using non-ascertained independent, genomewide loci (height, P = 3.29 × 10 −6 ; BMI, P = 0.018) and independent, genome-wide loci selected on the basis of their within-family association (height, P = 2.67 × 10 −6 ; BMI, P = 8.35 × 10 −5 ). We therefore reject the null model, and our results suggest that population genetic differentiation across these 14 European countries for height and BMI has been driven by selection on standing genetic variation across geographical regions in the evolutionary past.
[! Driven by selection. But there are few or no obvious drivers of differences in height or metabolism, suggesting any benefits are subtle, which further suggests that all sorts of other traits could be substantially geneticly affected even if we find that implausible... natural selection doesn't care if we can see pressures or not.]
We estimated the population genetic co-differentiation of height and BMI to ask whether selection has acted on both traits independently. We found a negative correlation between the population genetic means of −0.80 (95% CI = −0.95, −0.60; Fig. 2). This finding was consistent across predictors comprising non-ascertained genomewide loci (−0.77, 95% CI = −0.94, −0.55) and independent genomewide loci selected on the basis of their within-family association (−0.89, 95% CI = −0.97, −0.77). These results imply that selection has acted on common loci to increase height while reducing BMI and vice versa, and a genetic predisposition for tall stature at the population level was associated with a genetic predisposition for slenderness (low BMI). As height and BMI are nearly uncorrelated at the individual level (correlation among genetic profile scores within populations r = −0.016, 95% CI = −0.041, 0.001), selection for one trait should not elicit a response in the other. Our results suggest that selection has acted on both phenotypes, although, as some genes affect both phenotypes 48 , we cannot rule out differentiation in one trait having been mediated by selection for the other. The population genetic co-divergence shown here is inconsistent with random genetic drift because the expectation with drift is that the among-population genetic correlation will equal the within-population correlation 47,49"
The paper: "Population genetic differentiation of height and body mass index across Europe", Robinson et al 2015 https://www.dropbox.com/s/azvli29ltd3crof/2015-robinson.pdf / http://moscow.sci-hub.bz/14c52b6fe833e179bcb170969145e32d/robinson2015.pdf ; excerpts:
"Across-nation differences in the mean values for complex traits are common1, 2, 3, 4, 5, 6, 7, 8, but the reasons for these differences are unknown. Here we find that many independent loci contribute to population genetic differences in height and body mass index (BMI) in 9,416 individuals across 14 European countries. Using discovery data on over 250,000 individuals and unbiased effect size estimates from 17,500 sibling pairs, we estimate that 24% (95% credible interval (CI) = 9%, 41%) and 8% (95% CI = 4%, 16%) of the captured additive genetic variance for height and BMI, respectively, reflect population genetic differences. Population genetic divergence differed significantly from that in a null model (height, P < 3.94 × 10−8; BMI, P < 5.95 × 10−4), and we find an among-population genetic correlation for tall and slender individuals (r = −0.80, 95% CI = −0.95, −0.60), consistent with correlated selection for both phenotypes. Observed differences in height among populations reflected the predicted genetic means (r = 0.51; P < 0.001), but environmental differences across Europe masked genetic differentiation for BMI (P < 0.58).
We reestimated the effects of each SNP in a within-family study design, which is unbiased by population stratification, and used these effect sizes to create a genetic predictor for both phenotypes (also termed a ‘profile’ or ‘polygenic score’) 36 .
There is no certainty that population stratification is completely controlled for in large-scale meta-analyses, and we thus repeated our analysis using (i) non-ascertained, unlinked (LD r 2 <0.1 and >1 Mb apart), common (minor allele frequency (MAF) >1%) HapMap 3 loci (~40,000 SNPs) and (ii) unlinked (LD r 2 <0.1 and >1 Mb apart), common (MAF >1%) HapMap 3 loci selected on the basis of their within-family association with each phenotype (~40,000 SNPs for both traits). This analysis provides an unbiased, genome-wide estimate, representing a lower limit of population genetic differentiation at common, unlinked loci. The maximum proportion of variance in a polygenic predictor attributable to population genetic differences was 24% (95% CI = 9%, 41%) and 8% (95% CI = 4%, 16%) for height and BMI, respectively, using 2,660 SNPs for height and 11,919 SNPs for BMI. For height, the largest proportion of population-level variance was captured by SNPs of low P value in the meta-analysis (Supplementary Fig. 6). For BMI, the continual addition of SNPs increased the proportion of populationlevel variance captured (Supplementary Fig. 6). For both traits, the among-population variation was greater in predictors that explained greater phenotypic variance (Supplementary Fig. 7). Our results were confirmed using the non-ascertained independent, genome-wide loci (height: 8.6%, 95% CI = 3%, 15.7%; BMI: 2.8%, 95% CI = 1.1%, 5.3%) and the set of independent, genome-wide loci selected on the basis of their within-family association (height: 11.9%, 95% CI = 4.5%, 21.8%; BMI: 8%, 95% CI = 3.4%, 14.7%). The lower among-population variance captured using non-ascertained loci reflects reduced prediction accuracy, likely due to the addition of a large number of loci with no detectable association. Subsequent results are presented using the predictor for each trait that captured the greatest amount of populationand individual-level variance (comprising 2,660 SNPs for height and 11,919 SNPs for BMI); however, the results remained the same irrespective of the SNPs selected (Supplementary Fig. 8). The predicted population genetic means for the traits are shown in Figure 1 alongside the observed values, estimated from an independent set of recently published data 25,37 accounting for trends over time.
Genetic differences among populations may occur by random, chance processes or through natural selection in the evolutionary past 19,38–46 . We thus compared our estimates to the values from a null quantitative genetic model of multivariate population differentiation 32,47 . We found strong evidence that the divergence of each trait was greater than expected under the neutral model (Fig. 2). The overall level of neutral genetic differentiation was small for both height (1.2%; 95% CI = 0.01%, 1.78%) and BMI (1.9%, 95% CI = 0.48%, 2.97), reflecting the average F ST (a measure of population differentiation due to genetic structure; Supplementary Note) of the SNP sets between the populations of 1% for height and 1.2% for BMI. Our results were confirmed using non-ascertained independent, genomewide loci (height, P = 3.29 × 10 −6 ; BMI, P = 0.018) and independent, genome-wide loci selected on the basis of their within-family association (height, P = 2.67 × 10 −6 ; BMI, P = 8.35 × 10 −5 ). We therefore reject the null model, and our results suggest that population genetic differentiation across these 14 European countries for height and BMI has been driven by selection on standing genetic variation across geographical regions in the evolutionary past.
[! Driven by selection. But there are few or no obvious drivers of differences in height or metabolism, suggesting any benefits are subtle, which further suggests that all sorts of other traits could be substantially geneticly affected even if we find that implausible... natural selection doesn't care if we can see pressures or not.]
We estimated the population genetic co-differentiation of height and BMI to ask whether selection has acted on both traits independently. We found a negative correlation between the population genetic means of −0.80 (95% CI = −0.95, −0.60; Fig. 2). This finding was consistent across predictors comprising non-ascertained genomewide loci (−0.77, 95% CI = −0.94, −0.55) and independent genomewide loci selected on the basis of their within-family association (−0.89, 95% CI = −0.97, −0.77). These results imply that selection has acted on common loci to increase height while reducing BMI and vice versa, and a genetic predisposition for tall stature at the population level was associated with a genetic predisposition for slenderness (low BMI). As height and BMI are nearly uncorrelated at the individual level (correlation among genetic profile scores within populations r = −0.016, 95% CI = −0.041, 0.001), selection for one trait should not elicit a response in the other. Our results suggest that selection has acted on both phenotypes, although, as some genes affect both phenotypes 48 , we cannot rule out differentiation in one trait having been mediated by selection for the other. The population genetic co-divergence shown here is inconsistent with random genetic drift because the expectation with drift is that the among-population genetic correlation will equal the within-population correlation 47,49"
