Determining the Degree of Promiscuity of Extensively Assayed Compounds.
Level 5 - mechanism / opinion, no new human data
Preclinical computational database analysis without human participants (level 5 by design analogy)
PubMed 27082988 · doi:10.1371/journal.pone.0153873
What was done
The authors conducted a computational cheminformatics analysis of public domain bioactive compounds that were extensively tested across hundreds of assays against hundreds of targets. Assay promiscuity was distinguished from target promiscuity and analyzed separately for primary and confirmatory assays, incorporating assay frequencies and inactivity records.
What was found
Differences between assay promiscuity and target promiscuity were small. Across the most extensively tested compounds, the average degree of target promiscuity was 2.6 to 3.4 and the median was 2.0. The abstract reports no specific count for total compounds analyzed.
Why it matters
This paper demonstrates that bioactive compounds retain a low baseline degree of target promiscuity even when subjected to extensive screening, contrasting with the higher promiscuity reported for approved drugs. This supports the hypothesis that promiscuous compounds may be enriched during clinical development or more intensively profiled once therapeutic utility is shown.
Limits
The abstract does not state the exact sample size (number of compounds evaluated). The study relies on public screening database records, which may have heterogeneous assay formats, varying active/inactive cutoffs, and potential reporting biases.
Cited by
- contradicts The average approved small molecule drug can bind between 20 and 30 different proteins in the human body.