Skip to main content


Figure 10 | Journal of Cheminformatics

Figure 10

From: Predicting cytotoxicity from heterogeneous data sources with Bayesian learning

Figure 10

Correlation of Bayesian scores of a test set calculated from the IC50 and percent inhibition models. The test set consisted of 10% of the molecules from the percent inhibition cytotoxicity model training set and the IC50 cytotoxicity model training set. The Bayesian scores are binned to get 16 bins. The pies represent the number of molecules within those bins with size proportional to the number of molecules. Red segments represent the proportion of non-cytotoxic molecules and blue segments represent the proportion of cytotoxic molecules. A similar plot (not shown) was obtained using all of the data in the training set.

Back to article page