Fig. 3: External validation on independent SMKI datasets: relationships between the model predicted AE probabilities and the observed AE incidences. | Nature Communications

Fig. 3: External validation on independent SMKI datasets: relationships between the model predicted AE probabilities and the observed AE incidences.

From: Decoding kinase-adverse event associations for small molecule kinase inhibitors

Fig. 3: External validation on independent SMKI datasets: relationships between the model predicted AE probabilities and the observed AE incidences.

Columns ac are results based on two neratinib and one imatinib studies, respectively. The top panels show the scatterplots for the corresponding predicted probabilities vs all the frequently reported TEAEs (in ≥10% of study subjects) and the bottom panels for the corresponding observed TEAEs vs all AEs predicted to occur in ≥10% of study subjects. On each plot, the blue line represents the fitted linear regression line, and the gray band represents the 95% confidence interval; Pearson’s r (two-tailed) with p value is displayed. Yellow points represent four frequently reported TEAEs with predicted probabilities of <0.1; purple points AEs with both predicted probabilities and observed incidences of ≥10%.

Back to article page