Dear all,
I'm performing virtual screening on a library of 150 compounds. I used Vina, I selected the best 20 compounds and I docked them with AutoDock. I also used another program (surflex) in order to compare my results. Since the scoring functions are differents in the two programs (Autodock and surflex), I can't select the "real" best compounds. So I'm looking for free, online tools that can rescore my docking results. Any suggestions please?
Kind regards.
No single-mouse-button-click tool exists for calculating accurate protein-ligand binding free energies. But one can try calculating protein-ligand binding free energies using molecular dynamics simulations, such as free energy perturbation, linear interaction energy . . . to find out why obtaining quantitatively accurate protein-ligand binding free energies with a single-mouse-button-click tool is not possible . . .
Tu dois avoir alors bcp de faux positifs. Essayes de scorer les poses obtenues à l'aide de fonctions d'évaluations différentes de celle utilisée pendant le docking. Chaque fonction ayant des inconvénients, l'analyse par plusieurs fonctions te permettra de détecter les faux positifs. bon courage
Given such a small number of compounds (150), I would recommend to screen them all in your experimental assay. At least you would avoid false negatives that the docking filter would produce.
Dear Researcher
To rely on the virtual screening protocol. You have to include some false positives or decoy molecules to the dataset and dock them. Sfter docking You need to calculate percentage of true positive out of total molecules. In other words this is also known as enrichment factor (EF). If enrichment factor or percentage of rue positives is high you can rely on that docking protocol for virtual screening. There is no need of comparison in such cases. False positives or decoys are the compounds which have structural similarity to your molecules but are inactive for the target of interests. Hope this helps...
Dear Sichaib Zeineb,
You can find the following link of screening scorer useful. It will also give you validation statistics through ROC analysis (http://stats.drugdesign.fr/).
If you want consensus for protein protein or protein nucleic acid docking then you can use Consrank.
https://www.molnac.unisa.it/BioTools/consrank/help/help.htm
http://stats.drugdesign.fr/
By rescoring, you mean that you want to validate your docking protocol, virtual screening utility or scoring accuracy.
You can go for RMSD (root-mean-square deviation), which required careful interpretation as you have to consider conserved interactions while looking for lower RMSD value.
Two more methods can be applied; enrichment factor (EF) and receiver operating characteristic curves (ROC). They are used for checking whether a docking program is able to select active ligands with respect to inactive ligands (decoys) and whether it is able to select these active ligands in the top % of a ranked database.
In my opinion, the enzymatic assay would be a logical choice. Since you have the limited number of molecules in the database, the possibility to encounter insignificant enrichment values is more.
you can select best compound on the basis of hydrogen bond and binding energy or minimum inhibition constant
any one you can select if common in both docking tools
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication