Integration of Residue Attributes for Sequence Diversity Characterization of Terpenoid Enzymes

Figure 3

Variable importance scores (1000 RF trials). Variable importance score distribution for 1000 runs of random forest classification on the amino acid index. Each boxplot represents distribution of each property (also called variable) represented on the horizontal axis. The properties have been ordered in decreasing order of the median score (red line in boxplot). For easier visualization, the set has been truncated to show the top 50 properties. The corresponding properties are shown in Table 1.