CDD Vault for medicinal chemistry decision support in large collaborations

From the desk of CDD CSO Sean Ekins, M.Sc., Ph.D., D.Sc.

Last week I attended the More Medicines for Tuberculosis (MM4TB) consortium meeting in Bilbao Spain. The MM4TB consortium consists of over twenty partners from thirteen countries in Europe, America, Asia and Africa. This event happens twice a year and represents an opportunity to bring the various groups together to discuss scientific progress. The meetings are also a chance to socialize with colleagues and share TB experiences and insights over two full days of work and leisurely lunches and dinners.

bilbao

One aspect of the meeting stood out for me: Following a large phenotypic screen performed by Dr. Rita Szekely and Monica Rengifo at EPFL, there was a set of under one hundred hits that were identified and confirmed as low uM actives and not cytotoxic. The P.I. Dr. Stewart Cole suggested that all the chemists in the group go through the hits and flag any that might be problematic based on their experience. So at the end of the first day, chemists from England, Russia, Hungary, Switzerland and a few token biologists commented on the compounds as Rita projected them on the screen from her CDD Vault. I was in attendance as someone bridging the chemistry and biology. What was remarkable was how few compounds were “thrown away” by the chemists and how dynamic the discussion was. Each chemist declared their interest or not in particular scaffolds. It was also much easier to do this in person and using CDD Vault to select the compounds as a collection. The set of hits was also profiled through the PAINS filters after the meeting with the help of Dr. Alex Clark and these removed three compounds (quinones and rhodanine alerts) that were in agreement with the chemist’s decisions. The compounds will now progress to target identification and screening in target based screens within the consortium.

It was terrific to see how CDD Vault platform aided this collaborative decision making in real time. This combination of the independent multi-chemist analysis aided by this collaborative platform and PAINS filters again highlighted the potential that relatively simple filters might have for non-chemists that use CDD Vault. While rules and filters cannot replace good medicinal chemists, they can point out potential compounds that may be problematic. While there are several external websites and mobile apps that calculate PAINS it might be a useful tool to incorporate into systems that house screening data and are used for decision making.

It will be exciting to see where this latest batch of hits end up, and I look forward to the next meeting in Paris in June.