Fig. 4From: Confidence bands and hypothesis tests for hit enrichment curvesComparisons of scoring methods surflex-dock, ICM, and their consensus, across 15 testing fractions using four testing procedures. The diagonal grid shows estimated score densities within activity classes for each scoring method; also shown are Kullback-Leibler divergences between estimated densities for the \(+\) and − classes. The upper grid shows p-values from testing for equality of hit enrichment curves from a pair of scoring methods, with colors corresponding to different testing procedures. The lower grid shows pointwise 95 percent confidence intervals that accompany results from the upper grid. Comparisons have not been adjusted for multiple testing as was the case in Table 2Back to article page