Supplementary Materials of “AptRank: an adaptive PageRank model for protein function prediction on bi-relational graphs”

1 Supplementary Experiments 1.1 Optimal Set of BirgRank Parameters To investigate how the four key parameters of BirgRank affect its prediction performance, we set each parameter as 0.1, 0.3, 0.5, 0.7 and 0.9, respectively, and test the prediction of protein functions using the yeast dataset with 50% of data as training set and the other 50% as testing set (missing function prediction, see the main text). As we can see in Supplementary Figure 1, different settings of these four parameters did not yield significant differences in performance. Therefore, we empirically set all the four parameters as 0.5 for all subsequent experiments in this study.