Figure 9

Putative false positive functional associations. A density plot of gene pairs binned by semantic similarity and graph diffusion kernel (GDK) score indicates putative false positive functional associations. The semantic similarity for each gene pair was calculated as the maximum of the values for Biological Process, Cellular Component, and Molecular Function. The average of this value over all gene pairs is 2.52, and gene pairs with semantic similarity below this value have essentially no semantic similarity. Gene pairs with GDK scores of 50 and above have an average semantic similarity of 5.6, corresponding to 23 genes annotated to the parent category. Putative false positives were defined for exploratory purposes as having GDK scores of 50 and above but semantic similarity of 2 or below, corresponding to 850 genes annotated to the parent category. The putative false-positive region is indicated and contains 101 gene pairs.