Genome-Wide Analysis of Dental Caries Variability Reveals Genotype-by-Environment Interactions

Abstract

Genotype-by-environment interactions (GEI) may influence dental caries, although their effects are difficult to detect. Variance quantitative trait loci (vQTL) may serve as an indicator of underlying GEI effects. The aim of this study was to investigate GEI effects on dental caries by prioritizing variants from genome-wide vQTL analysis. First, we identified vQTLs from ~4.3 M genome-wide variants in three cohorts of white children aged 3-5 (n = 396, n = 328, n = 773) using Levene's test. A total of 39 independent vQTLs with p < 1 × 10^-6 were identified, some of which were located in or near genes with plausible biological roles in dental caries (IGFBP7, SLC5A8, and SHH involved in tooth development and enamel mineralization). Next, we used linear regression to test GEI effects on dental caries with the 39 prioritized variants and self-reported environmental factors (demographic, socioeconomic, behavioral, and dietary factors) in the three cohorts separately. We identified eight significant GEIs indicating that children with vQTL risk genotypes had higher caries experience if they had less educated parents, lower household/parental income, brushed their teeth less frequently, consumed sugar-sweetened beverages more frequently, were not breastfed, and were female. We reported the first genome-wide vQTL analysis of dental caries in children nominating several novel genes and GEI for further investigations.

Keywords: dental caries; genome-wide vQTL; genotype-by-environment interactions.

Figure 1

Participant process workflow.

Figure 2

Manhattan plot for genome-wide vQTL scans of dfs in IFS, COHRA1, COHRA2, and meta-analysis; −log₁₀-transformed p-values (y-axis) for each SNP organized by physical position across the chromosomes (x-axis) are shown. Horizontal lines represent the genome-wide significant (5 × 10⁻⁸) and suggestive (1 × 10⁻⁶) p-value thresholds.

Figure 3

Interaction plots for significant GEI across three cohorts. In IFS, (a) rs1491071 with father’s educational attainment, CC genotype had higher dfs when father’s educational attainment < Four−year college; (b) rs1978471 with toothbrushing frequency, TT genotype had higher dfs when toothbrushing frequency < 2 times per day; (c) rs9830884 with household income, CC genotype had higher dfs when household income < 40,000 per year; In COHRA1, (d) rs7463853 with father’s income, GG genotype had higher dfs when father’s income < 25,000 per year; In COHRA2, (e) rs71508615 with sex, GG genotype had higher dfs when sex was female; (f) rs71508615 with mother’s income, GG genotype had higher dfs when mother’s income < 25,000 per year; (g) rs9685188 with breastfeeding status, TT genotype had higher dfs when breastfed; (h) rs73723358 with SSB, GG genotype had higher dfs when intaking SSB.

Figure 3

Interaction plots for significant GEI across three cohorts. In IFS, (a) rs1491071 with father’s educational attainment, CC genotype had higher dfs when father’s educational attainment < Four−year college; (b) rs1978471 with toothbrushing frequency, TT genotype had higher dfs when toothbrushing frequency < 2 times per day; (c) rs9830884 with household income, CC genotype had higher dfs when household income < 40,000 per year; In COHRA1, (d) rs7463853 with father’s income, GG genotype had higher dfs when father’s income < 25,000 per year; In COHRA2, (e) rs71508615 with sex, GG genotype had higher dfs when sex was female; (f) rs71508615 with mother’s income, GG genotype had higher dfs when mother’s income < 25,000 per year; (g) rs9685188 with breastfeeding status, TT genotype had higher dfs when breastfed; (h) rs73723358 with SSB, GG genotype had higher dfs when intaking SSB.