Figure 6

AKT gene module predicts sensitivity to AKT/mTOR signaling inhibitors.(A) Spearman rank correlation between the predicted AKT perturbation activity score and drug sensitivity (− log10GI50) values assessed over a panel of 45 breast cancer cell lines. Drugs have been ranked according to strength of correlation. Magenta: significant positive correlation, white: no significant correlation, cyan: significant negative correlation. Drugs that are PI3K/AKT/mTOR inhibitors are indicated in red. The Wilcoxon rank sum test P value testing for any skew of these inhibitors towards positive correlations is given. (B) Scatter plots of drug sensitivity against AKT activity scores for the two top ranked drugs in (A). Spearman rank correlation coefficients (SCCs) and associated P values are indicated. (C) Spearman rank correlation between the predicted AKT perturbation activity score and drug sensitivity (− log10GI50) values as assessed over a panel of 39 breast cancer cell lines from the Genomics and Drug Sensitivity in Cancer project. Note that for this set, the number of cell lines for each drug differs, hence individual P values are suppressed. Drugs that are PI3K/AKT/mTOR inhibitors are indicated in red. The Wilcoxon rank sum test P value testing for a positive skew of the PI3K/AKT/mTOR inhibitors towards positive correlations is given. (D) As (C), but for the data from Daemen et al. [34], which used − log10GI50 values. Correlations estimated over 42 breast cancer cell lines. Drugs that are PI3K/AKT/mTOR inhibitors are indicated in red. The Wilcoxon rank sum test P value testing for a positive skew of the PI3K/AKT/mTOR inhibitors towards positive correlations is given. +ve, positive; −ve, negative; Cor, correlation; GDSC, Genomics Drug Sensitivity in Cancer Project; RNA-seq, RNA sequencing; SCC, Spearman rank correlation coefficient; Sig., significant; DS, drug sensitivity.