To identify rare or novel variants that affect risk for type 2 diabetes in Native Americans, we obtained exome sequence data on 177 Pima Indians. These unrelated individuals had been metabolically characterized for predictors of type 2 diabetes when they were nondiabetic, which included measures of insulin secretion, insulin action, and percent body fat (pfat). A subset (N50) had undergone abdominal fat and/or skeletal muscle biopsies for genome-wide gene expression profiling. Exome sequencing was performed by ShanghaiBio Corp. SNP validation showed that variants with a quality score of 100, that met HWE expectations, were unlikely to be artifacts. Among SNPs that met these criteria (N=408,487) only 88,005 mapped within exons;of these, 71,367 were coding (35,195 nonsynonymous), while 3,913 and 12,725 were in 5 and 3 UTRs, respectively. There were also 1,359 coding indels (698 predicted a frameshift). Variants occurring in 5 subjects were preliminarily analyzed for association with a metabolic trait in the 177 individuals, and those SNPS with the strongest associations are currently being genotyped in a sample of 3500 Pima Indians informative for type 2 diabetes and BMI. SNPs detected by sequencing are also being analyzed for association with cis-acting gene expression levels in individuals with expression data. In the preliminary analyses, 3 of the 8 top signals for pfat (P=10-5-10-6) resulted from novel variants. Five of these signals were in genes unknown to be associated with obesity (LYSMD4, ANKRD36, USP5, DHX32, AGPAT1) and require validation which is ongoing. Another top signal (CYB5A) had been detected in our prior GWAS for BMI, and the BMI association replicated in two independent samples (total N=7285, P=10-7). Another signal (PARP1) is an obesity gene in mice. We are also obtaining whole genome sequence data from Complete Genomics. To date we have received sequence data on 30 Pima Indians and we anticipate receiving 200 additional genomes in the coming year. The 30 genomes are currently being annotated for novel variants that are predicted to be damaging. These variants will be genotyped in large samples for assocation analysis with type 2 diabetes and obesity.
Showing the most recent 10 out of 14 publications
