Skip to main content
Fig. 7 | Journal of Cheminformatics

Fig. 7

From: Deep learning-driven prediction of drug mechanism of action from large-scale chemical-genetic interaction profiles

Fig. 7

Results of ten examples of M. tuberculosis inhibitors predicted by the D-MPNN with RDKit descriptors. Applying the cutoffs derived from ROC curves of the validation set, the predictions were binarized. In the heatmap, light green represents a positive prediction and light yellow represents a negative prediction. If the positive result predicted by the model matches the ground truth, the color is shown as dark green (TP), whereas if the result does not coincide with the label, the color appears dark yellow (FN)

Back to article page