Predictive Models of Gene Regulation from High-Throughput Epigenomics Data
Figure 4
Distribution of -scores for up- (Up), down- (Dw), and non- (Nr) regulated genes for the optimal attributes for each experiment, calculated by maximizing the Information Gain and minimizing the absolute value of the median for the -score distribution of the Nr subset. The -axis shows the -score corresponding to the enrichment of the attribute. These distributions correspond to the set of LCG-IC loci of Pair1.