Mining promiscuous chemotypes in PubChem
Promiscuous compounds are associated with frequent hitters in HTS assays. We developed a method for rapid and automatic identification of chemotypes associated with frequent hitters based on matching molecular pairs concept. Using PubChem bioassay database we identified frequent hitters and “probable” chemotypes responsible for compound promiscuity. A scoring scheme was designed from distribution of biological activities across assays, substances, and molecular matching pairs, allowing for ranking of “most” promiscuous chemotypes. Identified promiscuous chemotypes can used as a filter for prioritization of HTS hits, and compound library design.