Skip to main content
Fig. 4 | BMC Bioinformatics

Fig. 4

From: Task-driven knowledge graph filtering improves prioritizing drugs for repurposing

Fig. 4

Predicted ranks of four antiviral Compounds and SARS-CoV2 entities compared to the ranks of all Compounds. The light green and blue bars show the ranks of predictions of the type (SARS CoV2 entity, treats\(^{-1}\), Compound) for all Compounds in the graph, using the original and the modified dataset for training. The predicted ranks of the four antiviral Compounds are highlighted. The red and dark green bars show the change in the predicted rank of a given antiviral compound when the model was trained on the original dataset versus when the model was trained on the modified dataset. In A, the SARS-CoV2 entity is the SARS-CoV2 Spike entity. B shows the mean ranks of all 27 SARS-CoV-entities in the graph with each of the four antiviral compounds (for all individual ranks see Additional file 1: Table S3). Expected median rank is 9816, lower ranks indicate higher-than-average probabilities of treatment

Back to article page