Skip to main content
Fig. 2 | Journal of Cheminformatics

Fig. 2

From: Comparison of structure- and ligand-based scoring functions for deep generative models: a GPCR case study

Fig. 2

The measured sphere exclusion diversity (SEDiv) (a) and internal diversity (IntDiv) (b) of a randomly sampled 1000 (@1k) subset of a variety of virtual libraries and datasets of characterised molecules with activity against particular targets belonging to a target class, or single targets. Internal diversity shows counterintuitive behaviour such as, measuring GDB13 as more diverse than GDB17 and hERG active molecules as diverse as molecules active against any family A GPCR, any kinase or a virtual library designed towards achieving diversity. Conversely, sphere exclusion diversity measures diversity in line with chemical intuition

Back to article page