Computational protocol: Cardio-metabolic parameters are associated with genetic admixture estimates in a pediatric population from Colombia

Similar protocols

Protocol publication

[…] The assumption of normality of the quantitative variables was evaluated using the Kolmogorov-Smirnov test. The qualitative variables were presented with their frequency distributions; the quantitative variables were summarized according to their median and interquartile ranges. Food consumption data were analyzed based on the Program Evaluation of Dietary Intake (Evaluación de la Ingesta Dietética, EVINDI v4) []. Reports of nutrient intake were processed using the program PC SIDE (Personal Computer Version of Software for Intake Distribution Estimation) v 1.0. The allele and genotypic frequencies were calculated with the PLINK v. 1.07 program []. Ancestry proportions were calculated with the ADMIXMAP v 3.2 program [], which uses a frequentist-Bayesian method. The analysis of median differences was performed using Mann-Whitney U or Kruskal-Wallis tests. Individual cardio-metabolic parameters (body mass index, waist circumference, HDL cholesterol, systolic or diastolic blood pressure, fasting blood glucose, and insulin resistance) were coded as binary variables based on the cutoff values described above. Logistic regression analysis was performed using the dichotomous variables as an independent model to assess the effect of each of the genetic ancestries on each cardio-metabolic parameter. Because of the known effects of socioeconomic status, parental education, physical activity and diet on these components, each of these covariates were included in the model as control variables. Analyses were conducted with SPSS (Statistical Package for the Social Sciences) v. 19.0 statistical software and PLINK v. 1.07 []. Because all associations assessed were based on a priori hypotheses, both unadjusted and adjusted probability values are reported. The significance test were adjusted by the Bonferroni method, two threshold significance as a p-value <0.05/3 = 0.0167 (three ancestry), and p-value <0.05/27 = 0.0018 (global correction, three ancestries * nine cardio-metabolic parameters) was considered significant for avoiding the error of multiple testing. The threshold of significance for all other comparisons was p <0.05. […]

Pipeline specifications

Software tools PLINK, SPSS
Applications Miscellaneous, GWAS
Chemicals Glucose, Triglycerides