Progressive effects of single-nucleotide polymorphisms on 16 phenotypic traits based on longitudinal data

Background There are many research studies have estimated the heritability of phenotypic traits, but few have considered longitudinal changes in several phenotypic traits together. Objective To evaluate the progressive effect of single nucleotide polymorphisms (SNPs) on prominent health-related phenotypic traits by determining SNP-based heritability ($$h_{snp}^{2}$$hsnp2) using longitudinal data. Methods Sixteen phenotypic traits associated with major health indices were observed biennially for 6843 individuals with 10-year follow-up in a Korean community-based cohort. Average SNP heritability and longitudinal changes in the total period were estimated using a two-stage model. Average and periodic differences for each subject were considered responses to estimate SNP heritability. Furthermore, a genome-wide association study (GWAS) was performed for significant SNPs. Results Each SNP heritability for the phenotypic mean of all sixteen traits through 6 periods (baseline and five follow-ups) were significant. Gradually, the forced vital capacity in one second (FEV1) reflected the only significant SNP heritability among longitudinal changes at a false discovery rate (FDR)-adjusted 0.05 significance level ($$h_{snp}^{2} = 0.171$$hsnp2=0.171, FDR = 0.0012). On estimating chromosomal heritability, chromosome 2 displayed the highest heritability upon periodic changes in FEV1. SNPs including rs2272402 and rs7209788 displayed a genome-wide significant association with longitudinal changes in FEV1 (P = 1.22 × 10−8 for rs2272402 and P = 3.36 × 10−7 for rs7209788). De novo variants including rs4922117 (near LPL, P = 2.13 × 10−15) of log-transformed high-density lipoprotein (HDL) ratios and rs2335418 (near HMGCR, P = 3.2 $$\times$$× 10−9) of low-density lipoprotein were detected on GWAS. Conclusion Significant genetic effects on longitudinal changes in FEV1 among the middle-aged general population and chromosome 2 account for most of the genetic variance.