Signals of polygenic adaptation on height have been overestimated due to uncorrected population structure in genome-wide association studies

Abstract

All of these studies have been based on SNP associations, in most cases with 81 effect sizes discovered by the GIANT Consortium, which most recently combined 79 82 individual GWAS through meta--analysis, encompassing a total of 253,288 83 individuals. [13,14] Here, we show that the selection effects described in these 84 studies are severely attenuated and in some cases no longer significant when using 85 summary statistics derived from the UK Biobank, an independent and larger single 86 study that includes 336,474 genetically unrelated individuals who derive their 87 ancestry almost entirely from British Isles (identified as "white British ancestry" by 88 the UK Biobank) (Supplementary Table S1). The UK Biobank analysis is based on a 89 single cohort drawn from a relatively homogeneous population enabling excellent 90 control of potential population stratification. Our analysis of the UK Biobank data 91 confirms that almost all genome--wide significant loci discovered by the GIANT 92 We began by estimating "polygenic height scores"-sums of allele 101 frequencies at independent SNPs from GIANT weighted by their effect sizes-to 102 study population level differences among ancient and present--day European 103 samples. We used a set of different significance thresholds and strategies to correct 104 for linkage disequilibrium as employed by previous studies, and replicated their 105 signals for significant differences in genetic height across populations. [4-11] 106 ( Figure 1a, Supplementary Figure S2). We then repeated the analysis using 107 summary statistics from a GWAS for height in the UK Biobank restricting to 108 individuals of British Isles ancestry and correcting for population stratification 109 based on the first ten principal components (UKB).ND 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint . http://dx.doi.org/10.1101/355057 doi: bioRxiv preprint first posted online Jun. 25, 2018[15] This analysis resulted in a 110 dramatic attenuation of differences in polygenic height scores (Figure 1a, 111Supplementary Figures S2--S4). The differences between ancient European 112 populations also greatly attenuated (Figure 1a, Supplementary Figure S5). 113 Strikingly, the ordering of the scores for populations also changed depending on 114 which GWAS was used to estimate genetic height both within Europe (Figure 1a, 115 Supplementary Figures S2--S5) and globally (Supplementary Figure S6), 116 consistent with reports from a recent simulation study.[16] The height scores were 117 qualitatively similar only when we restricted to independent genome--wide 118 significant SNPs in GIANT and the UK Biobank (P < 5x10 --8 ) (Supplementary Figure 119 S2b). This replicates the originally reported significant north--south difference in the 120 allele frequency of the height--increasing allele [4] Figure S2b), and that confounding due to stratificat...

An evolutionary compass for detecting signals of polygenic selection and mutational bias

“…However, one of our strongest signals (height) did not replicate the UK Biobank, while two other signals of selection suggested by earlier height studies also did not replicate in the UK Biobank (Turchin et al 2012;Yang et al 2015). The nonreplication of height is in concordance with other recent studies finding reduced evidence for selection on height in the UK Biobank cohort applying different methods Sohail et al 2018). The most conservative interpretation of the nonreplication of height is to suppose that some of the signals we and others observed in the GIANT cohort are driven by population stratification, and the UK Biobank analysis correctly removes this spurious contamination.…”

“…; Sohail et al. ). The most conservative interpretation of the nonreplication of height is to suppose that some of the signals we and others observed in the GIANT cohort are driven by population stratification, and the UK Biobank analysis correctly removes this spurious contamination.…”

Geographic variation and bias in polygenic scores of complex diseases and traits in Finland

“…Recently, it has been reported that PS differences between populations are prone to technical and confounding biases arising especially from population genetic differences (i.e. genetic divergence) or relatedness structure between the GWAS discovery and the target data Reisberg et al 2017;Berg et al 2018;Curtis 2018;Sohail et al 2018). To assess whether some of the results in Figure 2 and Table 1 might be affected by these problems, we next concentrate on evaluating our PS in several ways.…”

A method for genome-wide genealogy estimation for thousands of samples

“…, evidence for selection remains controversial40,[47][48][49][50][51]59,60,62 . It was recently reported that recent selection on increased height in Europeans has been overestimated and that estimates have been confounded by subtle population stratification40,50,51 . Our test finds an enrichment of selection evidence for both effect directions for height, across most populations except East Asians, using the large collection of UK Biobank associations.…”