Fig. 5From: Genetic regulation of gene expression and splicing during a 10-year period of human agingPopulation-level age-specific splicing across the transcriptome. a Manhattan plot of the splicing-age discoveries. Each dot represents a cluster in a gene; the x-axis gives the position of the cluster in the genome, and the y-axis represents the strength of the association with age. We find 294 age-associated genes with 503 clusters of alternatively excised introns (3.4% of tested genes and clusters; FDR ≤ 5%). Three of the top ten genes that had the strongest association with age, i.e., SFPQ, PER1, and SETX, are related to the circadian rhythm, disruption of which accelerates aging [46]. PLK3, which was also in the top ten most associated genes, is implicated in stress responses and double-strand break repair. b Number of introns with at least one significant sQTL (sIntrons) at age 70 and 80 for uncorrected analysis (number of hidden factors = 0) and analysis corrected for one up to 14 hidden factors. We detected 550 and 509 sIntrons at age 70 and 80 (FDR ≤ 5%), respectively. The depletion of sIntrons at age 80, relative to age 70, is statistically significant (exact McNemar’s test; P=8.6×10−3). Dashed line indicates the number of hidden factors that maximizes discovery at each age. c Proportion of sIntrons discovered at age 70 (80) that validated (FDR ≤ 20%) at age 80 (70). The validation proportion of sIntrons at age 70 is significantly smaller than the proportion at age 80 (binomial proportion test; P=2.2×10−2), indicating a decrease in genetic regulation with age. d Scatter plot of sQTL effect sizes at each age for introns that are sIntrons in at least one age group. We observed a strong correlation of the fixed effect sizes between the two ages (Spearman’s ρβ=0.98). Blue line represents linear regression fitBack to article page