Systematic assessment of analog relationships between bioactive compounds and promiscuity of analog sets
Abstract
A computational approach is introduced for the systematic detection of all synthetically accessible analogs of given active compounds. Notably, for more than 30% of all active compounds in ChEMBL, no analogs were detected. For the remaining compounds, sets of analogs were identified with varying R-groups at all possible substitution sites or only a subset of substitution sites. Active compounds without analogs were widely distributed over different target sets. They provide a large source of candidates for further chemical exploration and have been made available in an open access deposition. For the majority of compounds with analogs, chemical exploration had no detectable influence on promiscuity (i.e., the ability of a compound to interact with multiple targets). However, for a subset of ∼26% of active compounds with analog sets, notable increases in promiscuity were observed. In these cases, average promiscuity increased from ∼1.5 targets per active compound to approximately four targets for analogs. However, these increases were largely caused by small numbers of highly promiscuous analogs.