Analysis of gene expression to predict dynamics of future hypertension incidence in type 2 diabetic patients

BACKGROUND: The main focus of the Genetic Analysis Workshop 19 (GAW19) is identification of genes related to the occurrence of hypertension in the cohort of patients with type 2 diabetes mellitus (T2DM). The aim of our study was to predict dynamics of the future hypertension incidence, based on gene expression profiles, systolic and diastolic blood pressure changes in time, sex, baseline age, and cigarette smoking status. We analyzed data made available to GAW19 participants, which included gene expression profiles of peripheral blood mononuclear cells (PBMCs) from the diabetic members of 20 Mexican American families. METHODS: On the basis of mid blood pressure measurements at several time points, the coefficient of regression (slope) was calculated for each individual. We corrected the slope value in patients treated with antihypertensive medications. Feature preprocessing methods were used to remove highly correlated probes and linear dependencies between them. Subsequently, multiple linear regression model was used to associate gene expression with the regression coefficient calculated for each T2DM patient. Tenfold cross-validation was used to validate the model. We used linear mixed effects model and kinship coefficients to account for the family structure. All calculations were performed in R. RESULTS: This analysis allowed us to identify 6 well-annotated genes: RTP4, FXYD6, GDF11, IFNAR1, NOX3, and HLA-DQ2, associated with dynamics of future hypertension incidence. Two of them, IFNAR1 and NOX3 were previously implicated in pathogenesis of hypertension. CONCLUSIONS: There is no obvious mechanism that links all detected genes with dynamics of hypertension incidence. Identification of possible connection with hypertension needs further investigation.