Fig. 2. Ewas Methodology Description and Output.
a For each exposure (i), we fit a Cox model to estimate CHD risk, while controlling for a set of adjusting variables (n exposures, in total). Through this process, we also collect the VIF and the P value regarding the proportionality assumption. After the fitting phase (yellow), we proceed with the multiple testing protocols (blue). We leverage confounding variables to estimate the likelihood of CHD development for each subject, and the disease status is then permuted accordingly, m times. For each exposure, the FDR is estimated, and associations with an FDR < 0.05 are considered statistically significant. b Isorhamnetin is an O-methylated flavonol from the class of flavonoids mainly found in green pepper, red onion, and dill. EWAS shows that isorhamnetin intake is negatively associated with CHD risk (HR: 0.91; 95% CI: 0.87–0.95; P value 1.59 × 10−5, from two-sided Wald test, with no adjustment for multiple comparisons). The P value regarding the proportionality assumption indicates the appropriateness of using the Cox model. The VIF of 1.27 is an indication of the absence of severe multicollinearity.