Genome-Wide Association of Type 2 Diabetes in Africans: The AADM Study

The Africa America Diabetes Mellitus (AADM) study is a genetic epidemiological study of type 2 diabetes in Sub-Saharan Africa. Study participants were enrolled through university medical centers in Nigeria, Ghana, and Kenya. Ethical approval for the study was obtained from the Institutional Review Board (IRB) of each participating institution. All subjects provided written informed consent for the collection of samples and subsequent analysis. The case definition of type 2 diabetes was based on the American Diabetes Association (ADA) criteria. After providing informed consent, participants underwent the same enrollment procedures, which included collection of demographic information, medical history, clinical examination and a blood draw. Genome-wide SNP genotyping was done on either the Axiom™ PanAFR SNP array (n=1,808) or the Multi-Ethnic Global Array (MEGA) (n=3,423). After appropriate quality control, in silico imputation was done using the African Genome Resources Haplotype Reference Panel (at the Sanger Imputation Service). Imputed genotypes were filtered for variants with minor allele frequency (MAF)≥ 0.01 and information score (info) ≥ 0.3 for genetic association analysis. Genome-wide association analysis between type 2 diabetes and the imputed genotype dosages was done using a generalized linear mixed model, which adjusted for age, gender, body mass index, the genetic relatedness matrix and the first three principal components (PCs) of the genotypes.Metabolomics profiling of plasma samples of type 2 diabetes (T2D) cases and controls in Nigerians (West Africa) was done in the AADM Study. Plasma metabolites were measured in a total of 580 individuals (N=310 for the discovery phase and N=270 for the replication stage) using the global/untargeted approach on the Metabolon platform and following the manufacturer's standard operation protocols. The analytic methods are described in detail in Doumatey et al. [Genome Med 2024]. The measured metabolites level represented by peak areas are relative values. The peak area data were batch-normalized to remove the instrument batch effects (batch variability) and the batch-normalized data correspond to the median-scaled raw data. For each identified metabolite, the minimum value across all batches in the batch-normalized was imputed for the missing values. The batch-normalized and imputed data are natural log-transformed and consist of 1116 metabolites for the discovery cohort and 1071 metabolites for the replication. Welch's two-sample t-test on the log-transformed data was used to identify metabolites differentially expressed between T2D cases and controls . All other statistical analyses conducted on both the replication and discovery cohorts used the log-transformed data. To merge the discovery and replication, the same quality control samples (bridge samples) were run with each batch of the experimental samples in both cohorts and used to correct for additional variability and uniformize the procedures. The resulting merged data is the QC-normalized and imputed data that contains only metabolites that were common to both cohorts and successfully bridged for all batches (n= 891 metabolites).]]>
Chemical Annotation Discovery 5b1Chemical Annotation Replication 5b2Chemical Annotation Merged 5b3The purpose of the analysis was to identify genetic variants associated with type 2 diabetes in Africans. The phenotype trait is type 2 diabetes as a binary trait. Genome wide SNP genotyping was done on either the AxiomTM PanAFR SNP array (n=1808) or the Multi-Ethnic Global Array (MEGA) (n=3,423). After appropriate quality control, in silico imputation was done using the African Genome Resources Haplotype Reference Panel (at the Sanger Imputation Service.Inclusion criteria for type 2 diabetes: fasting plasma glucose concentration (FPG) ≥ 126 mg/dl (7.0 mmol/l) or a 2-hour post-glucose load of ≥ 200 mg/dl (11.1 mmol/l) on the oral glucose tolerance test (OGTT) or physician-prescribed pharmacological treatment for type 2 diabetes with the patient's clinical records indicating adequate justification for the therapy. Exclusion criteria for cases: probable type 1 diabetes based on the detection of autoantibodies to glutamic acid decarboxylase (GAD) and/or a fasting C-peptide ≤ 0.03 nmol/l. Inclusion criteria for controls: fasting plasma glucose of < 110 mg/dl or a two-hour post glucose load of < 140 mg/dl and no symptoms suggestive of diabetes (notably polyuria, polydipsia, and unexplained weight loss).]]>