Supporting Information S1 - Relating Diseases by Integrating Gene Associations and Information Flow through Protein Interaction Network

All supporting information are given in this file, including a description of how cutoffs were calculated for MimMiner score and correlation, the results of the evaluation of the accuracy of the p-values, and also the results of clustering using Cfinder. Figure S1, Finding the optimum number of clusters. Figure shows (A) the number of clusters, and (B) R, as a function of number of iterations. R is minimized after 10 iterations. Figure S2, Empirical p-values vs p-value cutoffs. The empirical values were calculated by shuffling the gene list 672 times. Figure S3, The probability of finding shared KEGG pathways is plotted (in red) as a function of average MimMiner score (a) or average correlation (b). The blue line shows the fitted piecewise function. The separation points are considered the cutoffs above which the scores or correlations are significant.
(PDF)