Computational protocol: Variation in the UCP2 and UCP3 genes associates with abdominal obesity and serum lipids: The Finnish Diabetes Prevention Study

Similar protocols

Protocol publication

[…] The single nucleotide polymorphisms (SNP) for genotype analysis were selected from the region spanning the UCP2 and UCP3 genes (~34.4 kb) by using the International HapMap database and Tagger software []. Rs660339, rs659366, and rs1800849 were forced in the selection procedure. The SNPs covered 86.2% of the genetic information of the studied region (r2>0.8). It should be noted that the DelIns variant of the UCP2 gene is not included in the database, since it is not a SNP but a 45 bp insertion.The rs659366 and rs660339 variants of the UCP2 gene and the rs1800849 variant of the UCP3 gene were screened by the restriction fragment length polymorphism after digestion with MluI, HincII and HaeIII, respectively, with minor modifications to previously described methods [,,]. The DelIns variant of the UCP2 gene was analysed by gel electrophoresis of the PCR-product. The intergenic region variant rs653529 and four variants locating in the UCP3 gene (rs15763, rs1726745, rs3781907, rs11235972) were genotyped by using the custom Golden Gate genotyping reagents and consumables (Illumina Inc, San Diego, CA). Only 501 (rs1726745) or 502 (rs653529, rs15763, rs3781907, rs11235972) subjects were successfully genotyped by Illumina. For other variants, n = 507. [...] The data were analysed using the SPSS/WIN program version 14.0 (SPSS, Chigago, IL, USA). The normality of distributions of study variables was evaluated with the Kolmogorov-Smirnov test with Lilliefors' correction, and appropriate transformation was used when necessary. For variables with skewed distribution, Kruskal-Wallis test was used. Univariate analysis of variance was used to compare the effect of the gene variants on continuous variables. Adjustment for age, gender and BMI was done, when appropriate. In addition, serum lipoprotein and lipid concentrations were adjusted for the use of cholesterol-lowering medication as well. Chi square test was used in comparison of categorical variables. The relative changes in HOMA-IS from baseline to three years were calculated as follows: [(parameter 3-year - parameterbaseline)/parameterbaseline] × 100%. Longitudinal changes were examined using repeated measures of General Linear Model. Homogeneity of variances was tested using Levene's test. Cox regression analysis, adjusted for the study group, baseline weight, weight change and baseline fasting plasma glucose, was performed to evaluate whether the gene variants predicted the development of T2DM.Linkage disequilibrium (LD) statistics were calculated by Haploview software [] and haplotype analysis was done by THESIAS 3.1 [], which is based on the stochastic-EM algorithm. Haplotype analyses of the quantitative variables were adjusted for age, gender and BMI, when appropriate. The survival analysis for haplotypes was adjusted for the study group, baseline weight, weight change and baseline fasting plasma glucose.A p-value < 0.05 was considered statistically significant. Correction for multiple hypothesis testing was performed with false discovery rate (FDR) using Q-value 1.0 software. π0 was estimated with bootstrap method [] using λ range from 0 to 0.9 by 0.05. Due to the distribution of p-values, the λ was set to 0 for correcting the results of Cox regression. Essentially, this is a conservative way of calculating FDR and thereby produces the estimate implicit in the Benjamini and Hochberg methodology. In text, q stands for FDR, and is reported for each p < 0.05 and should be interpreted as minimum FDR that is incurred when calling that test significant. Data are given as means ± SD, unless otherwise indicated. […]

Pipeline specifications

Software tools Tagger, Haploview, THESIAS
Application GWAS
Diseases Diabetes Mellitus, Machado-Joseph Disease, Glucose Intolerance, Obesity, Abdominal
Chemicals Cholesterol, Glucose, Triglycerides