What is it?
Hit identification is the last step in the early stages of drug discovery. Our docking and virtual screening protocol, which takes protein flexibility and waters into account, can boost the enrichment and increase your chances of finding a hit compound.
Our docking algorithm has been validated in a blind competition, ranking first in two affinity prediction tasks and top 10 in pose prediction. Furthermore, its different options allows us to cover different scenarios, from knowledge rich -where a structure with a bound compound might be available- to cases where only an homology model of the target is at one’s disposal. In terms of throughput, with our cluster, for a typical sized pocket, we can screen up to 1M of compounds in a couple of days, or 100K compounds in a week if we include protein flexibility.
How does it work?
These are the steps we typically follow:
- Protein preparation: The most suitable protonation states will be determined and hydrogens will be added.
- Mixed solvent molecular dynamics: The protonated target is simulated in a mixture of water and a fragment-size molecule, like benzene.
- Pocket profiling: Pharmacophores, structural waters and pocket ensembles are obtained from the simulations.
- Homolog search: Exhaustive search among public databases to find similar homologs with bound small molecules to further refine the pharmacophore.
- Molecular docking: The library of choice will be docked against the pocket ensemble and the obtained poses will be rescored after taking the pharmacophore and the waters into account.
- Ranking: The best compounds will be inspected and a short list of the most promising ones will be provided.
What will you obtain?
- A PyMol scene with binding mode predictions for the most promising compounds.
- A ranking of compounds.
- An extensive report summarizing the structural insights obtained.
- Conference calls with the team to discuss the results.