We next reasoned that if estrogen is important for PPD risk, we should be able to predict PPD status based on the degree to which E2 reprograms DNA methylation in the mouse. For each of the 1578 mouse E2 DMRs that overlapped with the human data set, we modeled the mean DNA methylation signature per DMR against the E2 treatment status. In a locus-specific manner, we inputted the human DNA methylation levels per individual in the discovery sample and attempted to predict PPD status using logistic regression. For each locus, the area under the curve (AUC) metric was used to measure prediction accuracy. We then attempted to combine biomarkers to increase predictability using the following algorithm (Supplementary Figure 3a). Linear discriminant analysis was used to combine loci in a forward step-wise manner such that model included loci were those that increased the AUC of the discovery sample until the value was maximized. This set of loci was then used to predict PPD status in the replication sample. The algorithm returned two loci at CpGs cg21326881 and cg00058938, corresponding to the
