Skip to main content


Fig. 4 | Journal of Cheminformatics

Fig. 4

From: The polypharmacology browser: a web-based multi-fingerprint target prediction tool using ChEMBL bioactivity data

Fig. 4

 Recovery statistics of targets of 670 drugs by various fingerprints and combined method used in PPB. The bar plots shows an average, a fraction of known targets found, b number of targets predicted and c hit rate calculated for 670 drugs (see Additional file 1: Fig. S1). For each drug analysis was performed at three different levels considering all targets (grey), targets with p value of ≤0.01 (green) and targets with p value of >0.01 (red). d Success rate for finding at least 1 known target of drugs among top 5 predicted targets by each method. e Average Tanimoto coefficient of binary substructure fingerprint between the query and the most similar bioactive ligand associated with correctly predicted known targets in the results list. f–h Percentages of targets of one fingerprint found by another fingerprint and percentages of targets unique to this fingerprint, considering three different targets lists as mentioned before

Back to article page