Synergy Maps: exploring compound combinations using network-based visualization

Table 1 Descriptors calculated for the compounds, which were used for later visualization in Synergy Maps

Represented space	Descriptor type	Implementation
Physicochemical	771 physicochemical descriptors	PaDEL [41]
Structural	2,048 bit folded Morgan fingerprints [49] of radius 2	RDKit [42]
Biological	1,080 bayes affinity fingerprint [44]	In house Naive Bayesian models

Three diverse spaces were selected for representation of compounds, to give maximum insight into the properties of synergy in these different spaces. Physicochemical space are likely to differentiate compounds according to their ADMET properties; structural space will differentiate compounds according to the structure of their chemical graph; finally biological space attempts to differentiate compounds according to their relative affinity for protein targets. Hence, with this selection of descriptors our software is able to highlight structure in multiple facets of a dataset.

ISSN: 1758-2946