Figure 2.
A decision tree for the attribute “C-chromatin” learned from the SGD data. Starting from the top node, if a gene is annotated with the attribute listed in the node, then it travels down the edge labeled “+”; otherwise it travels down the edge labeled “−.” Leaf nodes are labeled with the number of genes in the training set that end up at the node, split into those that are annotated with “C-chromatin” in the SGD database (prefixed by “+”) and those that are not (prefixed by “−”). Ancestors and descendants of “C-chromatin” in the GO DAG were not allowed for making splits in this tree.
