Clinical and
biomarker phenotypic associations for carriers of
protein function-altering variants may help to elucidate gene function and health effects in populations. We genotyped 1127 Strong Heart Family Study participants for
protein function-altering single
nucleotide variants (SNV) and indels selected from a low coverage whole exome sequencing of American Indians. We tested the association of each SNV/indel with 35 cardiometabolic traits. Among 1206 variants (average minor allele count = 20, range of 1 to 1064), ~ 43% were not present in publicly available repositories. We identified seven SNV-trait significant associations including a missense SNV at ABCA10 (rs779392624, p = 8 × 10-9) associated with fasting
triglycerides, which gene product is involved in macrophage
lipid homeostasis. Among non-diabetic individuals, missense SNVs at four genes were associated with fasting
insulin adjusted for BMI (PHIL, chr6:79,650,711, p = 2.1 × 10-6; TRPM3, rs760461668, p = 5 × 10-8; SPTY2D1, rs756851199, p = 1.6 × 10-8; and TSPO, rs566547284, p = 2.4 × 10-6). PHIL encoded
protein is involved in pancreatic β-cell proliferation and survival, and TRPM3
protein mediates calcium signaling in pancreatic β-cells in response to
glucose. A
genetic risk score combining increasing
insulin risk alleles of these four genes was associated with 53% (95% confidence interval 1.09, 2.15) increased odds of incident diabetes and 83% (95% confidence interval 1.35, 2.48) increased odds of impaired fasting
glucose at follow-up. Our study uncovered novel gene-trait associations through the study of
protein-coding variants and demonstrates the advantages of association screenings targeting diverse and high-risk populations to study variants absent in publicly available repositories.