Background: The latest trans-ethnic genome-wide association study on A1C (glycated marker for long-term glycemia/glycemic control) identified 22 loci that modify A1C independently of glycemia. Glucose-independent differences in A1C may impact its utility in diabetes/prediabetes diagnosis and management. Some of these loci overlapped genes implicated in erythrocyte phenotypes that vary in prevalence across populations. Thus, we sought to estimate ancestral differences in associations with A1C at these 22 loci and an additional 23 erythrocyte genes not previously known to be associated with A1C, using TOPMed WGS data.
Methods: We performed single-variant association analyses of A1C in 5224 non-diabetic individuals [2662 self-identified European-ancestry (EA): Framingham Heart Study & Amish; 2562 self-identified African-ancestry (AA): Jackson Heart Study] using age and sex-adjusted linear mixed-effect regression with an empirical kinship matrix. Cohort-specific results were combined by meta-analysis. Cochran heterogeneity (PQ) and Fisher’s exact (PF) tests were used to evaluate ancestral differences in effect sizes and minor allele frequencies (MAF).
Results: We identified signals in ten A1C genes/loci at P<0.001 (α=0.05/45) in ≥1 ancestry, including six erythrocyte genes (SPTA1, HK1, HFE, HBS1L/MYB, ANK1, G6PD). Two were genome-wide significant in the meta-analysis: HK1 mainly in EA (10q22, rs72805692, MAFEA=0.11, βEA=-0.07, MAFAA=0.02, βAA=-0.11, PMeta=3x10-9), G6PD mainly in AA (Xq28, rs5986991, MAFEA=0.003, βEA=0.07, MAFAA=0.48, βAA=-0.12, PMeta=8x10-26). While we identified ancestral differences in MAF (PF<0.01), effect sizes were similar across ancestry (PQ≥0.1) for all loci except G6PD (rs5986991-PQ=0.005). We identified ancestry-specific associations in five additional erythrocyte genes: PIEZO1 (16q24, rs61745086, MAFEA=0.01, βEA=-0.13, PEA=6x10-4), HBA2 (16p13, rs2541640, MAFEA=0.009, βEA=-0.15, PEA=5x10-4), SPTB (14q23, rs78964602, MAFEA=0.008, βEA=-0.16, PEA=5x10-4), HBB (11p15, rs334, MAFAA=0.04, βAA=-0.18, PAA=5x10-8), HBA1 (16p13, rs148228241, MAFAA=0.06, βAA=0.10, PAA=1x10-4).
Conclusion: Genetic variations at erythrocyte genes exhibit ancestral differences in A1C effect and MAF. These variants may differ in the extent to which they contribute to inter-individual A1C variation, which may impact glucose estimation and diabetes diagnosis by A1C, particularly in minority populations.

The latest transethnic genome wide association study on A1C identified 22 loci that modify A1C independently of glycemia. These loci overlap genes implicated in erythrocyte phenotypes that vary in prevalence across populations. We used TOPMed WGS data to estimate ancestral differences in associations with A1C at these loci and 23 additional erythrocyte genes not previously known to be related to A1C.
We conducted WGS association analyses of A1C in 5224 nondiabetic individuals [2662 European ancestry (EA): Framingham Heart Study & Amish; 2562 African ancestry (AA): Jackson Heart Study] using age and sex-adjusted linear mixed-effect regression, and meta-analyzed cohort-specific results. We used Cochran heterogeneity and Fisher’s exact tests to assess ancestral differences in effect and Minor Allele Frequency (MAF).
We detected single variant associations in 11 genes/loci (P < 0.001; α = 0.05/45; ≥ 1 cohort; Table). Variants had ancestral differences in MAF (P < 1%) and ancestry-specific signals (AA: G6PD, HBB, HBA1; EA: ANK1, PIEZO1, SPTA1, SPTB, HBA2).
Genetic variation at erythrocyte genes exhibit ancestral differences in A1C effect and MAF. These variants may differ in their contributions to inter-individual A1C variation which may impact glucose estimation and diabetes diagnosis by A1C particularly in minority populations.

While obesity and adiposity are widely accepted to be major risk factors for the development of type 2 diabetes (T2D), the mechanisms underlying these associations are poorly understood. Understanding how these traits affect the genetic architecture of T2D and related glycemic traits could implicate novel pathways and facilitate the development of targeted therapies. We hypothesize that genes mechanistically linked to dysfunctional insulin signaling might be uncovered by performing whole genome sequence (WGS) analyses of T2D and insulin-related traits stratified by obesity. Using a combination of clinical and genetic data from the National Heart Lung and Blood Institute’s Trans-Omics for Precision Medicine (TOPMed) project, we developed a bioinformatics pipeline to perform WGS analyses on T2D and related traits. The pipeline accounts for genetic similarities within families and genetic differences between populations that could confound a true association. Stratifying our analyses by a body mass index (BMI) of 30 kg/m2 revealed many loci with association values that differed greatly between the stratified groups. These results strongly suggest interaction between BMI and the genetic architecture of T2D in our dataset. In the future, studying the direction of effect of these variants and cross-referencing these results with tissue-specific expression data could lead to biological hypotheses that could be validated experimentally. Our studies may help facilitate mechanistic explanations of obesity-associated insulin resistance and ultimately increase our understanding of T2D pathophysiology.

Abstract
Background: Genome-wide association studies (GWAS) of glycemic traits identified more than 75 common variants (minor allele frequency (MAF) ≥ 5%) at around 60 genomic loci. However, most identified variants reside in the non-coding genome, requiring whole genome sequence (WGS) for efficient evaluation of rare variation, potential causal variants and functional hypotheses for the observed GWAS associations.
Aim: To identify and annotate rare variants associated with fasting glucose (FG) using WGS association analysis (WGSA) in the Trans-Omics for Precision Medicine (TOPMed) program.
Methods: We leveraged TOPMed WGS Phase 1 (6,452 individuals without diabetes) to perform a pooled WGSA of FG (range [2.90-6.96]). Four cohorts with deep (>30x) sequence coverage were included: three European-ancestry - the Framingham Heart Study (N=3,212), the Cleveland Family Study (CFS, N=198); two African American - the Jackson Heart Study (N=2,490), CFS (N=253); and one Samoan American - the Samoan Adiposity Study (N=299). We used a mixed-effect linear regression model adjusted for age, sex and age2 with an empirical kinship matrix to account for relatedness. Rank-normalized residuals were tested for association with individual single nucleotide variants adjusting for study and ancestry.
Results: Common variant associations were identified at known FG loci: MTNR1B (rs10830963, MAF=0.19, P=4.3×10-14), GCK (rs2971670, MAF=0.18, P=5.6×10-9) and G6PC2 (rs560887, MAF=0.18, P=1.1×10-6). Novel genome-wide associations (P≤5×10-8) were seen for rare variants (0.003≤MAF≤0.01) in 25 loci. One missense variant in 10q23 (rs79370279, MAF=0.004, P=4.9×10-8) lies within MMRN2 that inhibits endothelial cells motility and acts as a negative regulator of angiogenesis. Suggestive evidence for association of FG with MMRN2 was found in a 13K exome sequence analysis in the AMP-T2D portal (rs10887673, MAF=0.23, P=8.3×10-5).
Conclusion: TOPMed WGS-wide search refined known FG-associated GWAS loci and discovered novel rare associations. Future work includes increase of the sample size with the Phase 1 Old Order Amish Study and other TOPMed cohorts.

Background
A1C GWAS have identified 17 common genetic variants of which 11 act through erythrocyte pathways and may interfere with A1C diagnostic accuracy. Genetic effect size and minor allele frequency (MAF) may differ under differential erythrocytic selection pressures. We used WGS data to determine ancestral differences in associations with A1C across the allelic frequency spectrum in white and black populations.
Methods
We performed single nucleotide variant (SNV) WGS association analyses with A1C in 4,889 non-diabetic individuals (2,466 Framingham Heart Study (FHS) whites; 2,523 Jackson Heart Study (JHS) blacks) within the NHLBI Trans-Omics for Precision Medicine Program (TOPMed). We conducted mixed-effect linear regressions with an empirical kinship matrix adjusted for sex, age and principal components in the 11 GWAS loci (+/-100kb around the lead SNV). Within each locus, we performed conditional analyses on the lead SNV in FHS to identify distinct signals (P≤0.005). We performed fixed-effect meta-analysis across both ancestries using METAL and Fisher’s exact test to evaluate ancestral differences in MAF.
Results
The HK1 locus (10q22) was associated with A1C at genome-wide significance (rs72805692, MAFFHS=0.10, βFHS=-0.07; MAFJHS=0.02, βJHS=-0.11, PMeta=4x10-9). We identified signals in 5 of the 11 loci (SPTA1, HBS1L(MYB), HK1, CDT1, HIST1H4A/HFE) at P<0.001 in at least one cohort. We observed no ancestral differences in effect sizes for the strongest signal (Cochran test, PQ≥ 0.1); though MAFs were generally higher in whites compared to blacks (PFisher 0.4 to <2x10-16). At 4 loci, we identified 5 low frequency and rare distinct signals with marked ancestral differences in MAF (PFisher<0.01).
Conclusion
Genetic variation at erythrocytic A1C GWAS loci exhibits ancestral differences in MAF. Understanding the extent to which such ancestral differential genetic effects interfere with A1C accuracy is a next step to reduce ethnic health disparities in diabetes.