I have 55 compounds that have certain values of IC 50 (In vitro), converted to log(1 /C). I tried to correlate these data with the binding energies obtained using some docking programs. However, a satisfactory correlation is not obtained.
Hi.
I recommend the MM/PB(GB)SA binding free energy analysis of docking poses after energy minimization.
http://ambermd.org/tutorials/#implicit
You can then plot -logC vs the lowest binding free energy.
it is a known flaw that simple docking scores fail to correlate with actual real life data. Plus, is the IC50 of all the compounds from a single assay or did you collected from various articles? If from various articles than IC50 can only be used to compare them if all has the same assay condition. Ki values are more amenable for comparison across diff bioassay condition but sadly many people simply report IC50. But you can convert IC50 into ki by using cheng-prusoff eqn. But that too might be waste of time.
which program did you use? simply redocking the bound substrate and and other known inhibitors and comparing their docking scores with those of ur own compounds should be suffice. These predictions u r trying to generate cannot replace bioassay. Aslo u can do some tricks like docking some decoy compounds for that protein which will give very poor scores compared to known active compounds which makes ur docking prediction more believable.
Dear Sir,
All the compounds from a single assay. I use AutoDock Vina. Can you give clarification about cheng-prusoff eqn.?
- Badges
- Science topic
- Similar topics
- Engineering
- Petroleum Engineering
- Oil