Figure 1

Schematic diagram of the atherosclerosis prediction model construction and validation processes. Our prediction model was developed by four distinct steps - 1) the identification of significant genes by t-test analysis of the training set FH1, 2) selected biomarkers known to be directly related to atherosclerosis performed by Ingenuity Pathway Analysis, 3) (optional) selection of COXEN biomarkers among the atherosclerosis-related gene biomarkers obtained from the third step which led to the 48, 6, and 26 biomarkers for test set 1 (circulating T-cell), test set 2 (white blood cell) and test set 3 (macrophage), 4) the multivariate prediction model construction based on universal or COXEN predictors.