Fig. 2: Active learning approach, target-specific score and trypsin-domain-specific score. | Nature Communications

Fig. 2: Active learning approach, target-specific score and trypsin-domain-specific score.

From: Simulations and active learning enable efficient identification of an experimentally-validated broad coronavirus inhibitor

Fig. 2

a Schematic representation of our active learning approach. b–d Correlation between experimental drug efficacies (as measured by maximal response in the biochemical assay, x-axis) and different virtual screening scores (y-axis) - mean normalized vinardo docking score, h-score of the docking pose and dynamic h-score (averaged over MD trajectory), respectively. Green and pink denote correctly and incorrectly classified compounds, respectively. Squares mark potentially reactive compounds, circles non-reactive ones. e Correlation between experimental (true) binding affinities and predicted binding affinities for trypsin-domain proteins in PDBbind version 2020.

Back to article page