Skip to main content
Fig. 3 | Genome Biology

Fig. 3

From: Leveraging biological and statistical covariates improves the detection power in epigenome-wide association testing

Fig. 3

Performance comparison of covariate-adaptive FDR control methods on real EWAS data. Comparison of the detection power across datasets, covariates, and methods (computed as log2 fold change (logFC) to the ST method, all numbers were added a pseudo-count of 1 to avoid 0s). The two left sidebars indicate the number of DMPs detected by ST with a target FDR level 5% and the estimated signal density (pi0), respectively. Rows are ordered by the signal density and columns are ordered by the method and covariate. Top panel shows the rank distribution of each covariate across datasets. The covariates are ranked based on the number of detected DMPs on each dataset. The mean value is used if ranks are the same, and a higher rank indicates higher detection power

Back to article page