There is not much to validate. If there are known actives for your protein of interest, then you can dock them and see if known actives have, on average, better scores than just random decoys. Of course, you can plot a Reciever Operating Characteristic (ROC) curve and calculate multiple measure-of-success metrics. This would be up to you to decide if the stats are good enough to start a large-scale prospective screen.
If no known actives exist, then you'd just screen a large library hoping that there would be a few true actives among the top scored hits (in fact, one tractable hit would be enough).
For docking result validation, you can repeat the docking process with another software or use the web servers such as BINDSURF (https://bio-hpc.ucam.edu/achilles/) and check the similarity of the results. In addition, you can redock crystal structure of target protein with inhibitor, then compare the interactions.
It is always depend on the types of compounds. If there is room for biological testing it is always the best solution. But again, it depends on many factors and the type of study.
Several techniques could be applied - re-docking, cross-docking, database enrichment, consensus docking etc. However, the most reliable docking protocol validation procedure is the experimental (in vitro; in vivo;) with further correlation calculations.
The best way to validate the docking results is to get expermintal data. Redocking and using of decoy ligands are the alternative insilico methods to validate docking.
- Badges
- Science topic
- Similar topics
- Filing