Exome Sequencing Data Analysis and a Case-Control Study in Mexican Population Reveals Lipid Trait Associations of New and Known Genetic Variants in Dyslipidemia-Associated Loci

Abstract

Background: Plasma lipid levels are a major risk factor for cardiovascular diseases. Although international efforts have identified a group of loci associated with the risk of dyslipidemia, Latin American populations have been underrepresented in these studies. Objective: To know the genetic variation occurring in lipid-related loci in the Mexican population and its association with dyslipidemia. Methods: We searched for single-nucleotide variants in 177 lipid candidate genes using previously published exome sequencing data from 2838 Mexican individuals belonging to three different cohorts. With the extracted variants, we performed a case-control study. Logistic regression and quantitative trait analyses were implemented in PLINK software. We used an LD pruning using a 50-kb sliding window size, a 5-kb window step size and a r² threshold of 0.1. Results: Among the 34251 biallelic variants identified in our sample population, 33% showed low frequency. For case-control study, we selected 2521 variants based on a minor allele frequency ≥1% in all datasets. We found 19 variants in 9 genes significantly associated with at least one lipid trait, with the most significant associations found in the APOA1/C3/A4/A5-ZPR1-BUD13 gene cluster on chromosome 11. Notably, all 11 variants associated with hypertriglyceridemia were within this cluster; whereas variants associated with hypercholesterolemia were located at chromosome 2 and 19, and for low high density lipoprotein cholesterol were in chromosomes 9, 11, and 19. No significant associated variants were found for low density lipoprotein. We found several novel variants associated with different lipemic traits: rs3825041 in BUD13 with hypertriglyceridemia, rs7252453 in CILP2 with decreased risk to hypercholesterolemia and rs11076176 in CETP with increased risk to low high density lipoprotein cholesterol. Conclusions: We identified novel variants in lipid-regulation candidate genes in the Mexican population, an underrepresented population in genomic studies, demonstrating the necessity of more genomic studies on multi-ethnic populations to gain a deeper understanding of the genetic structure of the lipemic traits.

Keywords: association study; dyslipidemia; exome analysis; genetic variants; mexican population.

FIGURE 1

Features of the analyzed gene variant set. (A) Cumulative variant distribution according to the minor allele frequency (MAF). Rare variants include Singletons, Doubletons, and variants with MAFs <0.01. Low Frequency variants include variants with MAFs = 0.01–0.05. Common variants are those with MAFs >0.05. (B) Distribution of coding and non-coding annotations, according to the Variant Effect Predictor tool (VEP) in the entire set of variants.

FIGURE 2

Manhattan plots from a meta-analysis based on the genome-wide association analysis. Panels show the −log p value for SNVs for each lipemic traits. Top hits for each trait are indicated in the figure. Red line indicates the significant threshold line: p = 6.4 × 10⁻⁵.

FIGURE 3

Forest and regional plots showing the top hits associated with lipemic traits. Forest plot showing odds ratio estimates and 95% confidence intervals (squared boxes) from the 3 cohorts DMS, MCDS, and MAIS, included in the study in HTG (A), HTC (C), and low HDL-C (E). Odds ratios for the meta-analyses are represented with a diamond. Regional plot of associated SNVs in HTG (B), HTC (D), and low HDL-C (F).