Chemists are harnessing the wisdom of the crowd to help them improve molecular diversity in a library of potential pharmaceutical candidates. Researchers at Johnson & Johnson, San Diego, USA, and their colleagues are tapping into the collective experience of their global medicinal chemistry community to prioritize molecules to be targeted for acquisition which may have previously been overlooked or rejected.
The process involves four steps: first, they screen out obvious non-drug structures. Secondly, they fingerprint the collection and cluster them using a similarity algorithm. The third step involves filling any diversity holes. The final step then has the community vote on the validity of the various possible candidates based on properties to build up diverse but likely candidates.