António Macedo, Patrick F. Sullivan, Pamela Sklar, Diana O. Perkins, David L. Braff, Eric D. Achtyes, Roman Kotov, Eli A. Stahl, Maria Helena Pinto de Azevedo, Colm O'Dushlaine, Elizabeth Bevilacqua, Célia Barreto Carvalho, Marquis P. Vawter, James Nemesh, Edward M. Scolnick, Jacquelyn L. Meyers, Jorge Valderrama, Shaun Purcell, Becky Kinkead, Douglas S. Lehrer, Peter F. Buckley, William Byerley, Humberto Nicolini, Fabio Macciardi, James L. Kennedy, Michael Escamilla, Ruben C. Gur, Dolores Malaspina, Ashley Dumont, Giulio Genovese, Helena Medeiros, Penelope Georgakopoulos, Colony Abbott, Diane Gage, Carlos N. Pato, Brooke M. Sklar, Roseann E. Peterson, Jordan W. Smoller, Steven A. McCarroll, Raquel E. Gur, Ayman H. Fanous, Laura J. Fochtmann, Stephen R. Marder, Sinéad B. Chapman, Mark Hyman Rapaport, James A. Knowles, Michele T. Pato, Janet L. Sobell, Evelyn J. Bromet, Conrad Iyegbe, Lynn E DeLisi, Jeffrey J. Rakofsky, Oleg V. Evgrafov, Jennifer L. Moran, Christopher P. Morley, Tim B. Bigdeli, Richard A. Belliveau, and Mantosh J. Dewan
Schizophrenia is a common, chronic and debilitating neuropsychiatric syndrome affecting tens of millions of individuals worldwide. While rare genetic variants play a role in the etiology of schizophrenia, most of the currently explained liability is within common variation, suggesting that variation predating the human diaspora out of Africa harbors a large fraction of the common variant attributable heritability. However, common variant association studies in schizophrenia have concentrated mainly on cohorts of European descent. We describe genome-wide association studies of 6152 cases and 3918 controls of admixed African ancestry, and of 1234 cases and 3090 controls of Latino ancestry, representing the largest such study in these populations to date. Combining results from the samples with African ancestry with summary statistics from the Psychiatric Genomics Consortium (PGC) study of schizophrenia yielded seven newly genome-wide significant loci, and we identified an additional eight loci by incorporating the results from samples with Latino ancestry. Leveraging population differences in patterns of linkage disequilibrium, we achieve improved fine-mapping resolution at 22 previously reported and 4 newly significant loci. Polygenic risk score profiling revealed improved prediction based on trans-ancestry meta-analysis results for admixed African (Nagelkerke’s R2 = 0.032; liability R2 = 0.017; P −52), Latino (Nagelkerke’s R2 = 0.089; liability R2 = 0.021; P −58), and European individuals (Nagelkerke’s R2 = 0.089; liability R2 = 0.037; P −113), further highlighting the advantages of incorporating data from diverse human populations.