Figure 2
From: Integrating phenotype and gene expression data for predicting gene function
Computing key similarities for predicting annotations. Three key similarity calculations are required to determine if annotation a should be assigned to gene g. Step 1: The lower threshold is the minimum of the total similarity of each gene with annotation a to all other genes with annotation a. Step 2: The upper threshold is the maximum of the total similarity of each gene without annotation a to every gene with annotation a. Step 3: The total similarity of gene g to all genes with annotation a is computed and the decision for assigning annotation a to gene g is made as illustrated in Figure 3.