Figure 9
From: Predictive screening for regulators of conserved functional gene modules (gene batteries) in mammals
False discovery rate estimation. A: Illustration of how false discovery rates (FDRs) were estimated by use of simulations. Values on the y axis represent the number of motifs below a certain p-score (x axis). Triangles show results for observed data, filled circles show results for permuted data (error bars show the 90% confidence interval from 100 simulations). The FDR was calculated as the ratio between the simulated expectation and the observation. Red dotted line: The number of clusters with at least one enriched motif. Note that several motifs where over-represented in the same cluster. B: Removal of paralogous genes from each cluster did not affect the number of detected motifs. Consequently, co-expressed paralogs is not an important source of false positives. C: The amount of DNA used per gene, and the phylogenetic footprinting stringency has a strong effect on the number of detected over-represented motifs. The sensitivity is higher when the amount of DNA is reduced. Error bars in B and C were obtained by using the 5th and 95th percentiles in the simulation to define the FDR.