Abstract
Background: Genome-wide association studies (GWAS) conducted for the type 2 diabetes related quantitative traits fasting glucose (FG) and fasting insulin (FI) have found more than 75 common variants (minor allele frequency (MAF) ≥1%) at around 60 genomic loci. Most signals reside in non-coding regions, 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 FG and FI using WGS analyses.
Methods: We used Trans-Omics for Precision Medicine (TOPMed) Program Framingham Heart Study (FHS) freeze 2 WGS data from a total of 1,219 (FG) and 1,142 (FI) non-diabetic individuals. Mixed-effect linear regression models were adjusted for age, sex and age2, and for FI also for body mass index (BMI), and residuals (FG) or log-transformed residuals (FI(BMI)) from these models were tested for association with individual single nucleotide variants (SNVs). Gene-based tests using T1 test were conducted for groups of rare variants (all missense SNVs in gene exons or all SNVs within the gene regions).
Results: WGS analysis detected at the genome-wide level (P<5x10-8) two loci for FG: a new locus within glutamate dehydrogenase 1 (GLUD1, 10q23, MAF=0.003, P=3.0x10-8), and a known locus including melatonin receptor 1B (MTNR1B, 11q21-q22, MAF ∈ [0.26-0.43], P≤3.7x10-8), and one new FI(BMI) locus close to annexin A4 (ANXA4, 2p13, MAF=0.001, P=3.5x10-8). T1 tests performed with missense exonic SNVs revealed a novel significant FI(BMI) locus within nudix hydrolase 7 (NUDT7, 16q23, 11 SNVs, MAF ∈ [0.0004-0.003], PGene=4.7x10-7). NUDT7 is expressed in human and mice liver, and in mice is linked to insulin receptor substrate (Irs1 and/or Irs2) signaling pathways.
Conclusion: TOPMed FHS WGS-wide search for FG and FI identified rare variants in novel candidate loci. Increase of the sample size is planned to confirm this finding.

The majority of genetic variants significantly associated with type 2 diabetes (T2D) reside in the non-coding genome, with many causal variants still unknown. We leveraged Whole Genome Sequence (WGS) phase 1 data from NHLBI’s Trans-Omics for Precision Medicine (TOPMed) initiative to perform a T2D WGS association (WGSA) pooled analyses. WGSA analyses included samples with deep (>30x) sequence coverage in n=13,408 (n=2,607 cases and n=10,801 controls) from 12 studies (65% European and 34% African ancestry). We used mixed effects models adjusting for sex, age, ancestry, study, empirical kinship and 10 principal components to adjust for relatedness and population structure. In multi-ethnic analyses, common (minor allele frequency (MAF)>0.05) variant associations (P value<5x10-8) were identified at known locus with T2D: TCF7L2 (rs7903146, P value=2.0x10-12 and 8 additional variants). For variants that have not been previously described, we identified 8 rare non-coding variant (MAF<.01) associations in 3 known loci: GCKR (chr2:27987368, MAF=.03%, P value =2.0x10-9), ADAMTS9 (chr3:64184329, MAF=.06%, P value =2.6x10-9) and CDKN2A/B (chr3:64184329, MAF=.02%, P value =3.8x10-8and 5 additional variants). Three rare nonsynonymous variants are in candidate genes that are outside of known loci in NTN4 (A166T, MAF=.03%, P value =1.5x10-9) and DIABLO (A60V, MAF=.06%, P value =1.8x10-8) are between T2D loci TSPAN8 and HNF1A; and OIP5(F20I, MAF=.2%, P value =1.4x10-8) downstream of RASGRP1. Additional associations included non-coding rare variants in loci not previously described with T2D; including intergenic rs778917988 (MAF=0.03%, P value=2.0x10-8) near SESN3, a known glucose-homeostasis gene; and 7 variants in SEMA6D (chr15:47474740, MAF=.2%, P value =1.3x10-8), a gene previously associated with a common variant and BMI. In ancestry specific analyses, an intronic variant in the ncRNA RP11-99E15.2 (chr14:77210987, MAF=.12%, P value =1.8 x10-8) was significant in Europeans only. Preliminary results suggest multi-ancestry WGSA can discover novel loci for complex traits. Work is ongoing to refine annotation for non-coding gene-based tests, perform fine-mapping, and extend into phase 2 and 3 data (N~14,000 cases and ~55,000 controls).

The majority of genetic variants significantly associated with diabetes-­related glycemic traits reside in the non­coding genome, with many causal variants still unknown. Whole genome sequence association (WGSA) analysis allows us to fine­-map known and novel loci across the non­coding genome without depending on imputation. Here, we present an initial WGSA of fasting insulin (FI) and fasting glucose (FG) levels in phase 1 (released in the fall of 2016) TOPMed data (N=7,121) with deep (>30x) sequence coverage in five cohorts, three European­-ancestry (EA): Framingham Heart Study, N=3,209; Old Order Amish Studies, N=980, Cleveland Family Study, N=197, and two African­American (AA): Jackson Heart Study, N=2,487, Cleveland Family Study, N=248. For each cohort, we used linear mixed effects models (as implemented in EMMAX or MMAP software) adjusting for sex, age and BMI, with empirical kinship for relatedness and population structure. We restricted to variants with minor allele count > 5 in more than 1 cohort and meta­-analyzed within and across ancestry with METAL. In EA+AA analysis, common (minor allele frequency [MAF]>0.05) variant associations (P<5x10­8) were identified for FG at known loci: MTNR1B (rs10830963, P=2.5x10­16; rs12792753, P=1.4x10­8), GCK (rs4607517, P=1.16x10­10, and 13 additional variants), and G6PC2 (rs560887, P=5.4x10­10). At the MTNR1B locus, fine­mapping across ancestries reduced the significant FG associations from 46 (EA) to 2 (AA+EA). Novel FI associations were seen in rare (MAF<1%) variants: 18q12.1 (rs146884135, P=2.1x10­9), 4p12 (rs193168677, P=1.2x10­8), 1q25.3 (rs550837507, P=3.8x10­8), and DCLK3 (rs573417731, P=4.8x10­8). These results, which are being extended with phase 2 TOPMed cohorts (expected release is in the summer of 2017; expected total N=27,830), highlight the value of multi­-ancestry WGSA to refine known and discover novel associations for complex traits. Future analyses with combined phase 1 and phase 2 data will assess (1) enrichment of sub­significant associations using diabetes­specific functional annotations of the non­coding genome, and (2) the association of rare variants within these diabetes­specific functional annotations with fasting glucose and fasting insulin levels.

The majority of genetic variants significantly associated with type 2 diabetes (T2D) and glycemic traits reside in the non-coding genome, with many causal variants still unknown. We leveraged Whole Genome Sequence (WGS) phase 1 data from TOPMed to perform a (1) T2D WGS association (WGSA) pooled analysis and (2) meta- and pooled-analysis of fasting glucose (FG). WGSA analyses included samples with deep (>30x) sequence coverage in 5 cohorts, three European-ancestry: Framingham Heart Study (NFG=3209, NT2D=4007); Old Order Amish Studies, NFG=980, Cleveland Family Study (CFS, NFG=197, NT2D=357), and two African American: Jackson Heart Study (NFG=2487, NT2D=3343), CFS (NFG=248, NT2D=332). We used mixed effects models adjusting for sex, age, with empirical kinship and/or principal components to adjust for relatedness and population structure. In multi-ethnic analyses, common (minor allele frequency [MAF]>0.05) variant associations (P value<5E-8) were identified at known loci with T2D: TCF7L2 (rs7903146, P value=2.5E-11 and 7 additional variants); and with FG: MTNR1B (rs10830963, P value=2.5E-16; rs12792753, P value=1.4E-8), GCK (rs4607517, P value=1.16E-10, and 13 additional variants), and G6PC2 (rs560887, P value=5.4E-10). Additional associations with T2D included 12 rare variants (MAF<.01) in six loci not previously described; including intergenic rs778917988 (MAF=0.0003, P value=2.0E-8) near SESN3, a known glucose-homeostasis gene. Preliminary results suggest multi-ancestry WGSA can discover novel loci for complex traits. Work is ongoing to refine annotation for gene-based tests, perform fine-mapping, and extend into phase 2 and 3 data (N=87,724).