Prediction performance of the dual-layer network model for drugs targeting the PI3K pathway in CGP.

<p><b>(</b>A) Bar graph showing the prediction performance of five models using experimental data from the CGP study, which was quantified using the Pearson correlation between predicted and observed activity areas. RF: prediction result by random forest using combined features of genomic and chemical data; SVR: prediction result by support vector regression using combined features of genomic and chemical data; CSN: prediction results using only the cell line similarity network; DSN: prediction results using only the drug similarity network; Integrated: prediction results using the dual-layer integrated cell line-drug network model. (B) Pearson correlation distribution of the five different models using t-tests to determine differences between two groups. (C) Correlations between observed and predicted activity areas using the dual-layer integrated cell line-drug network model.