oa A Genome Wide Association Search for Type 2 Diabetes Genes in Arabic Populations

Type 2 diabetes (T2D) is a chronic condition that emerged as serious medical, social and economic problem worldwide. Qatar was ranked by the IDF among the top 10 countries in the world with the highest prevalence of T2D that exceeds 20%. The causes of T2D are multiple but the contribution of genetic is well recognized. So far, large-scale genome-wide association studies (GWAS) have identified more than 80 susceptibility loci. However, these investigations were carried out mainly on European populations. The main goal of the present study was to conduct a T2D-GWAS analysis in Arab population. A case-control study using 870 T2D cases versus 666 controls was performed to compare allele frequencies across the genome of Moroccan samples. The Illumina Human Core Beadchip was used to genotype 298,930 SNP in these samples. Genotype calls were assigned using the GenCall algorithm as implemented in Illumina GenomeStudio (version 2010.3; Illumina Inc.). Stringent Quality Control (QC) criteria for filtering SNPs and samples for analyses were applied. All statistical analyses were performed using PLINK version 1.07 (http://pngu.mgh.harvard.edu/ ∼ purcell/plink/). Associations of SNPs with T2D were tested using logistic regression (-logistic command in PLINK), assuming an additive genetic model, and with either no adjustment or adjusting on the five first principal components. Correcting for potential confounders was a necessity since the large inflation factor (Genomic Control; GC = 1.176) was detectable in the unadjusted analysis. Increasing the number of PC to adjust for 10 corrected and decreased the genomic control under 1.07 (GC = 1.06). Imputation of non-observed genotypes is possible using linkage disequilibrium at genetic loci and a database of haplotypes from diverse populations (reference panel). This approach was implemented in all Moroccan samples and led to impute up to 30,071,165 variants (SNP + short INDELS). We selected the best SNP candidates for replication using the following criteria: P-value of association below 10–5, Quality of imputation I² ≥ 0.7 and Minor Allele frequency ≥ 5%. These three criteria led to a shortlist of 154 variants. Some of these variants were redundant because of linkage disequilibrium (r²>0.6). When two variants were in LD we selected the one showing the strongest association. By focusing on independent signals, we selected a list of 26 SNPs for further replication genotyping study. Replication study was performed in additional 1500 Moroccans T2D cases and controls using the Fluidigm genotyping platform and following the vendor instructions protocol. In this replication analysis, only 2 of the 26 SNPs (7.7%) showed nominal evidence of replication with a P-value adjusted