Figure 1From: A Regression-based K nearest neighbor algorithm for gene function prediction from heterogeneous dataExpression Correlation Analysis. The expression correlations of gene pairs are divided into 40 bins within the range of (-1,1) with width 0.05. The y axis shows the fraction of pairs in the same function class in each bin. The red horizontal line shows the probability of a pair of genes being in the same class by chance (0.177).Back to article page