Finding needles in molecular haystacks: what computation really delivers

Developing new bioactive molecules is an expensive work, but computational toolsincluding docking (see Figure 1) can make the process faster and more focused.However, no prediction matters until it survives in vitro or in vivo testing.In this seminar, I will walk through the key steps involved in screening large compoundlibraries, but also in persuading biopharma partners to test the most promisingcandidates. I will highlight successful antiviral projects and how such virtual screeningscan be applied beyond drug discovery. For example, in a recent collaboration withindustry we helped design safer, more effective insecticide formulations [3].Of course, not every computational project ends with a success story. I will also sharesome failures to give a realistic picture of the strengths and limitations of currentcomputational strategies.

[1] Cyclodextrins for use in Coronavirus infection therapy. Patent num.PCT/EP2023/057735

[2] Zika Virus Protease Inhibitors and Methods of Use Thereof. Patent num.PCT/CN2017/088420

[3] Biodegradable iminosugars and aminocyclitol glycosides as alpha-glucosidaseinhibitors with synergistic insecticidal, acaricidal, or both activities. Patent num.P202230916.