Characterization of the fine-scale genetic structure of the Turkish population

Abstract

The construction of population-based genetic resources plays a pivotal role in the study of human biology and disease. In this study, the fine-scale genetic structure of the Turkish (TR) population was characterized using the whole-exome (WES, n =2,589)andwhole-genome(WGS, n =773)sequencesof3,362unrelatedin-dividuals from Turkey. Significant levels of admixture from Balkan, Caucasus, Middle East, and Europe were detected in the TR subregions, consistent with the history of Anatolia. Results of the population structure analyses showed that the TR and European populations have a closer genetic relationship than previously appreciated. Inbreeding coefficient calculations and runs of homozygosity analysis reflected the unique effects of the high rate of consanguineous marriage on the TR genome. A TR Variome comprising over 40 million variants was constructed using the data generated in this study. Derived allele frequency (DAF) calculations revealed that 28% of TR-WES and 49% of TR-WGS variants in the very rare frequency bins (DAF < 0.005) were not listed in the Genome Aggregation Database. The lists of clinically-relevant variants and human gene knockouts in the TR Variome were also listed in this study, presenting the potential of the TR Variome being an invaluable resource for future disease gene identification studies. Additionally, a reference panel for genotype imputation was generated using TR-WGS data. Since this panel significantly increased imputation accuracy in both TR and neighboring populations, it will probably facilitate genome-wide association studies in these populations. In the second part of the study, the sequencing data of a total of 3,599 unrelated TR individuals were assessed for previously reported pathogenic (RP) variants and predicted pathogenic (PP) variants in Online Inheritance in Men (OMIM) genes associated with a pheno-type. Analyses revealed that no less than 70% of TR people have at least 1 RP variant, and all individuals possess at least one RP and/or PP variant in their genome. Moreover, 25% of individuals carried at least one RP variant in the newborn screening genes. Each individual in the study also had at least a 1 in 17 chance of carrying an RP variant in one of the 73 American College of Medical Genetics recommended actionable genes. MEFV, ABCA4, CYP21A2, PAH,and CFTR displayed the highest cumulative carrier frequencies (CF), consistent with the high prevalence of the phenotypes they are responsible for. By estimating the CF and genetic prevalence in 3,251 OMIM genes using RP and PP variants, this study presents the most comprehensive data so far demonstrating the landscape of genetic disease in the TR population.