TIGP (BIO)—Detecting ongoing natural selection affecting allele frequencies across age groups to uncover genetic variants contributing to disease susceptibilities

Abstract

Genetic variants that affect a complex trait such as a common disease could also impact fitness and, consequently, are suppressed by purifying selection. Hence, genetic variance of a common disease could be largely contributed by mutations at low frequency in the population and are likely to be population specific. Here, we analyzed 509,817 whole-genome genotyped variants in 72,635 Han Taiwanese individuals to detect candidate variants experiencing ongoing selection by comparing differences in allele frequency across generations. We detected 168 variants significantly departing from the neutral expectation of allele frequencies in different age groups after controlling for the possible age-dependent genetic structure. Most candidate variants (160) appear to show decreases in allele frequency in younger generations which are consistent with the action of purifying selection, suggesting that these candidate variants could reduce fitness of their carriers and likely contribute to disease susceptibilities. Among them, 86 candidate variants (53.8%) are indeed reported previously to be associated with a wide spectrum of diseases including both early onset (e.g., Marfan syndrome, Matthew-Wood syndrome, etc.) and late onset diseases (many candidates (35) were reported to increase cancer susceptibilities). In particular, we identified a number of pathogenic variants that are in strong linkage disequilibrium (LD) in BRCA1 and BRCA2, separately. Analyses of the 1487 individuals whose whole-genome sequencing data are available further revealed strong signatures of positive selection favoring the alternative haplotype in BRCA1 and signatures of balancing selection on FADS2. We also performed phenome-wide association analyses across 30 physiological, hematological and cardiometabolic measurements to further detect any possible functional consequences for each of these candidate variants. We found that FADS2 rs2072114, an intronic variant, appears to be associated with multiple traits (i.e., total cholesterol, triglyceride, fasting glucose level in blood, hemoglobin level, red blood cell and platelet count, and heartbeat). FADS2 has been shown strongly linked to cardio-metabolic diseases. In addition, we also found several cancer-related pathogenic variants associated with the size of red blood cell and hemoglobin level. Our findings are expected to facilitate disease prevention and management.

附件下載