Genome-wide analysis in over 1 million individuals of European ancestry yields improved polygenic risk scores for blood pressure traits
Jacob M. Keaton, Zoha Kamali, Tian Xie, Ahmad Vaez, Ariel Williams, Slavina B. Goleva, Alireza Ani, Εvangelos Εvangelou, Jacklyn N. Hellwege, Loic Yengo, William J. Young, Matthew Traylor, Ayush Giri, Zhili Zheng, Jian Zeng, Daniel I. Chasman, Andrew P. Morris, Mark J. Caulfield, Shih-Jen Hwang, Jaspal S. Kooner, David Conen, John Attia, Alanna C. Morrison, Ruth J. F. Loos, Kati Kristiansson, Reinhold Schmidt, Andrew A. Hicks, Peter P. Pramstaller, Christopher P. Nelson, Nilesh J. Samani, Lorenz Risch, Ulf Gyllensten, Olle Melander, Harriëtte Riese, James F. Wilson, Harry Campbell, Stephen S. Rich, Bruce M. Psaty, Yingchang Lu, Jerome I. Rotter, Xiuqing Guo, Kenneth Rice, Péter Vollenweider, Johan Sundström, Claudia Langenberg, Martin D. Tobin, Vilmantas Giedraitis, Jian’an Luan, Jaakko Tuomilehto, Zoltán Kutalik, Samuli Ripatti, Veikko Salomaa, Giorgia Girotto, Stella Trompet, J. Wouter Jukema, Pim van der Harst, Paul M. Ridker, Franco Giulianini, Véronique Vitart, Anuj Goel, Hugh Watkins, Sarah E. Harris, Ian J. Deary, Peter J. van der Most, Albertine J. Oldehinkel, Bernard Keavney, Caroline Hayward, Archie Campbell, Michael Boehnke, Laura J. Scott, Thibaud Boutin, Chrysovalanto Mamasoula, Marjo‐Riitta Järvelin, Annette Peters, Christian Gieger, Edward G. Lakatta, Francesco Cucca, Jennie Hui, Paul Knekt, Stefan Enroth, Martin H. de Borst, Ozren Polašek, Maria Pina Concas, Eulalia Catamo, Massimiliano Cocca, Ruifang Li-Gao, Edith Hofer, Helena Schmidt, Beatrice Spedicati, Melanie Waldenberger, David P. Strachan, Maris Laan, Alexander Teumer, Marcus Dörr, Vilmundur Guðnason, James P. Cook, Daniela Ruggiero, Ivana Kolčić, Eric Boerwinkle, Michela Traglia
Abstract
Abstract Hypertension affects more than one billion people worldwide. Here we identify 113 novel loci, reporting a total of 2,103 independent genetic signals ( P < 5 × 10 −8 ) from the largest single-stage blood pressure (BP) genome-wide association study to date ( n = 1,028,980 European individuals). These associations explain more than 60% of single nucleotide polymorphism-based BP heritability. Comparing top versus bottom deciles of polygenic risk scores (PRSs) reveals clinically meaningful differences in BP (16.9 mmHg systolic BP, 95% CI, 15.5–18.2 mmHg, P = 2.22 × 10 −126 ) and more than a sevenfold higher odds of hypertension risk (odds ratio, 7.33; 95% CI, 5.54–9.70; P = 4.13 × 10 −44 ) in an independent dataset. Adding PRS into hypertension-prediction models increased the area under the receiver operating characteristic curve (AUROC) from 0.791 (95% CI, 0.781–0.801) to 0.826 (95% CI, 0.817–0.836, ∆AUROC, 0.035, P = 1.98 × 10 −34 ). We compare the 2,103 loci results in non-European ancestries and show significant PRS associations in a large African-American sample. Secondary analyses implicate 500 genes previously unreported for BP. Our study highlights the role of increasingly large genomic studies for precision health research.