Heritability And Genome-wide Association Analyses For Fasting Plasma Glucose In Adult Twins

Objective:Fasting plasma glucose（FPG）,one of the criteria for diabetes diagnosis,is higher or even if below the diagnostic threshold for diabetes could make the mortality increased.However,there are not enough genome-wide association studies exploring the genetic influences on FPG level in Chinese population.Thus,we aim to explore the magnitude of genetic effect on FPG variation and further identify specific genetic variants associated with FPG level.Methods:The study subjects were adult twins from the Qingdao Twin Registration System,and each of the individuals received and signed a written informed consent.FPG level was measured on the Semi-automatic biochemical analyzer.Sex,ABO blood group and microsatellite DNA gene scanning and typing techniques were used to determine the zygosity.DNA samples were genotyped on the Illumina’s Infinium Omni2.5Exome-8v1.2BeadChip platform.The IMPUTE2 software was performed to impute untyped SNPs.Standard structural equation modeled by Mx software was used to calculate the heritability of FPG level,with adjusting for the age,sex and body mass index（BMI）.GEMMA was performed to evaluate the association between FPG level and SNPs genotypes in the SNPs-based analysis.Gene-based analysis was conducted on VEGAS2.The PASCAL was performed in pathway enrichment analysis.Results:There were 382 twin pairs in heritability analysis and 139 dizygotic（DZ）twin pairs were for GWAS analysis.The median FPG level was 5.14mmol/L（interquartile range:4.60-5.90mmol/L）with a range of 3.50-21.10 mmol/L in GWAS analysis.After adjusting for age,sex and BMI,monozygotic（MZ）twin correlation(r _MZ=0.68)was more than twice as large as that of DZ twins(r_DZ=0.20).So we first fitted the full ADE model and then compared the performances of it with the nested model（AE）.The best fitting model for FPG level was AE model.The heritability of FPG level was 67.66%（95%CI:60.50%-73.62%）.A total of 1,365,181 SNPs were included in this study after quality control.In SNPs-based analysis,the Q-Q plot indicated that there was no evidence of the bias from the possible population stratification.The Manhattan plot revealed no SNPs reached the level of genome-wide significance(P<5×10^-8).However,there were 28 SNPs exceeded the threshold for suggestive significance(P<1×10^-5),including rs10931893,rs295134,rs4516415,rs295114,rs1900706,rs159320,rs11691757,etc.The SNPs were on the chromosome 2,5,6,8,10,and 13 with 17,1,4,1,4 and 1 SNPs,respectively.The SNP with the smallest P value was rs10931893(P=1.53×10^-7)in SPATS2L gene on chromosome 2q33.1.After imputing untyped SNPs by using the 1,000 Genomes Project Phase 3 as reference panel,there were 7,405,822 SNPs being used to explore the association with FPG level.The results indicated that one SNP rs60106404(P=2.38×10^-8)in SPATS2L gene was reached the level of genome-wide significance.In addition,there were 216SNPs exceeded the threshold for suggestive significance,including rs60106404,rs295119,rs10931893,rs295134,rs1369842,rs4673814,rs10804097,etc.1,007 significant genes were found in the gene-based analysis including SPATS2L,ADCY5,KCNK5,PCSK1,TSPO,PTPRA,SETD7 etc.In addition,SPATS2L,KCNK5 and ARAP3 genes were consistent with the results in SNPs-based analysis.Pathway enrichment analysis discovered 719 biological pathways were associated with FPG level,includingβ-alanine metabolism,regulation of insulin secretion,glucagon signaling in metabolic regulation,IL-1R pathway,signaling by PDGF,taste transduction,the synthesis of PIPS at the plasma membrane,cysteine and methionine metabolism,signaling by FGFR,olfactory transduction,signaling by TGF beta receptor complex,etc.Conclusion:The heritability of FPG level was high in Qingdao,and some functional genes including SPATS2L,KCNK5,ADCY5,PCSK1 etc.,and biological pathways includingβ-alanine metabolism,regulation of insulin secretion,glucagon signaling in metabolic regulation,IL-1R pathway etc.were significantly associated with FPG level.Thus,our study provides significant clues for further exploring the extent of genetic factors effect on glucose.