Polypharmacology directed compound data mining: identification of promiscuous chemotypes with different activity profiles and comparison to approved drugs.
Level 5 - mechanism / opinion, no new human data
In silico computational database mining and chemoinformatic analysis without clinical human data (graded by design analogy).
PubMed 21070069 · doi:10.1021/ci1003637
What was done
Public bioactivity databases were mined starting from ~35,000 compounds active against human targets at potencies of at least 1 µM. The authors built a structural hierarchy consisting of active compounds, atomic property-based scaffolds, and unique molecular topologies. Network representations were constructed to evaluate scaffold-target family relationships and assess activity profiles across target families.
What was found
The analysis identified 33 distinct topology chemotypes containing molecules active against at least 3 different target families. A subset of these promiscuous chemotypes was significantly enriched in approved drugs compared to general bioactive compounds. Among these, 190 approved drugs had an average of only 2 known target annotations despite belonging to the 7 most promiscuous chemotypes, which collectively exhibited activity across 8 to 15 target families.
Why it matters
This computational framework identifies structural scaffolds prone to multi-target binding. It highlights specific approved drugs that are strong candidates for polypharmacological profiling and potential drug repurposing.
Limits
The findings are purely computational and derived from curated database records; no new experimental or biochemical validation was performed. The analysis is susceptible to literature and testing biases in public repositories, where heavily studied drugs may appear more promiscuous due to broader testing.
Cited by
- contradicts Most approved drugs impact at least 40 different biological pathways and mechanisms across the human brain and body.