Figure 1.
PD DDIs reflected at the molecular network level.
(A) A representative example of drug target distance: the minimum distance of the shortest path between two drugs' targets on PPI network. (B–C) Hit rate (B) and coverage(C) of known PD DDIs among the drug pairs grouped by their target distances on the PPI network.
Figure 2.
(A) Mapping drugs onto PPI network through drug-target associations. (B) Weighting PPI network by gene expression profile across human tissues. (C) Defining target-centered system, including known drug-targets and their first-step neighbor protein on PPI network. (D) Scoring systematic connectivity based on drugs' target-centered systems (S-score). (E) Scoring phenotypic similarity based on drug clinical side effects (P-score). (F) Cross-validating between S-score and P-score. (G) Integrating by a Bayesian probabilistic model. (H) Validating the prediction performance using GSP DDIs from two independent databases, DrugBank and STITCH. (I) Comparing with other methods. (J) Explaining the molecular mechanisms of the polypharmacy side effects recorded in TWOSIDES.
Figure 3.
(A) All possible drug pairs ranked and binned into groups by their S-score. (B) Fold-enrichment of hits of known PD-like DDIs in each bin (y-axis) plotted against average S-score (x-axis). (C) ROC curves to evaluate the performance of various scoring methods using DrugBank PD-like DDIs as GSPs. (D) ROC curves to evaluate the prediction performance of S-score using three types (PD, PK, pharmaceutical) or known DDIs in DrugBank as GSPs. (E) Associations of S-score with drug clinical side effect similarity (P-score). Average of P-score (y-axis) is plotted for top N% of drug pairs ranked by S-score (x-axis). P-value for top 5% drug pairs against all possible drug pairs was calculated from Wilcoxon rank sum test.
Figure 4.
(A) Desipramine and trimipramine. (B) Zonisamide and memantine. The drug target-centered systems (ellipse) colored for different drugs, includes drug targets (vee) and their first-step neighboring proteins (circle) in the PPI network. The edges in the network were weighted by the PCC based on cross-tissue expressions.