I want to combine many PDB complex to generate shared pharmacophore. But the problem is when I generate shared pharmacophore of all PDB complex, I can't generate any shared pharmacophore. So what I did is I create pharmacophore separately based on similarity of the ligands. What the other method or approach that I can used to get shared pharmacophore?
Use Autodock Vina for separated dockings first. Use the flexible residues approximation. Now the question of a shared pharmacophore is rather complex. In a recent study of the docking of hallucinogens with the 5-HT2A receptor, we found that not all the molecules interact in the same way with the binding site.Therefore, dock each molecule for separate, analyze their interactions and observe what they have in common. regards, Juan
If you have MOE there is an option to do that. Check PharmDock as well.
I usually align the proteins/enzymes (backbone) disregarding the complexed ligands as the active sites are usually conserved. It is amazing to see how the ligands that interact through the same binding mode end up well aligned with each other. You can then build composite pharmacophore models based on the aligned ligands or structure-based models using the aligned proteins.
Hello,
Some additional information would be perfect - Which software do you use? how many structures do you want to superpose? Do you want to create a model manually or authomatically?
Hi,
If you have a lot of PDB structures, it is better divide them to groups by comparing
the binding pocket (the comparing parameters could be rmsd and sequence simularity). Then in every group you can selected the main hot spots of protein-ligand interactions (Basicli it will be your overall pharmacophore model of the group). Then you must compare obtained pharmacophore models, if they are similar you can create normal shared pharmacophore models, otherwise it does not make any sense (this may be a case of your fail). - structure based pharmacophore models.
Also you can selected whole active compounds toward to your target, clustering them. After that base on every cluster create its pharmacophore model. - ligand based pharmacophore models.
The good tools for such work is LigandScoup (Here is an opportunity to take a free license for 1 month) (http://www.inteligand.com/ligandscout/),
Moe (http://www.chemcomp.com/MOE-Molecular_Operating_Environment.htm), Discovery Studio (http://accelrys.com/products/discovery-studio/pharmacophore.html). Futhermore, these programs can create shared pharmacophare models automatically.
Dear Dr Ewgenij Proschak,
I use Ligand Scout and the complex is about 32. I prefer to do it automatically. But if it can't, doing it manually should be fine.
- Badges
- Science topic
- Similar topics
- Filing