Fig. 4From: HPOAnnotator: improving large-scale prediction of HPO annotations by low-rank approximation with HPO semantic similarities and multiple PPI networksEach circle is a pair of two HPO terms in NHPO, with sharing the same numbers of proteins, say M. The y-axis is the average similarity score between two HPO terms over those proteins sharing the same M, and the x-axis is M, i.e., the number of shared proteins. The red line is fitted by a linear functionBack to article page