Short answer: take the pose with the most negative binding energy

Long answer:

There are two aspects you should consider when analyzing your docking results:

1) binding geometry (docking problem)

2) binding affinity (scoring problem)

I start with the 2nd. You need a function which predicts a binding affinity. Luckily, there are some computational ways to get it, more luckily these ways can be systematically improved. If you wish, you can choose a better theory for binding thermodynamics... Then you select the pose with the highest affinity (e.g. most negative Gibbs free energy). No matter, what quantity the "affinity" in your model/theory is. In our lab, we try to develop a quantum-chemistry-based scoring functions. Honestly, we are approaching, but still are not there... Here's a recent review: http://dx.doi.org/10.1002/cplu.201300199

For binding geometries it is more tricky. There is no systematic way for improvement your theories. There exist plenty of (quite diverse: http://dx.doi.org/10.1002/prot.10115) algorithms and the must be carefully tested on known structural data. Usually, they are not fully transferable between various ligand scaffolds. Based on geometric criteria only, it is in my opinion far more difficult to select the best ligand.

I think the best Z-value is the best docking result. some tools give other values or scores of docking. you can refer to the help menu to know about the significance of these scores.

The result that best explains your experimental findings is the best docking result. For example, if your system allows you the experimental determination of FRET distance between ligand and the intrinsic donor/acceptor probe, you may use that to filter your docking result.

Thanks but what about table of energy or binding energy or inhibit constant ? Can i choose the best model from these parameters?

After completion of docking one could open the rank by energy values. From the distribution bar plot find out the maximum free energy change conformation (definitely it should be negative). The configuration of maximum negative free energy change should be taken as most stable configuration.In other terms one can finalize the minimum inhibition constant conformation as final one as reverse of it gives the maximum binding constant.

Hi,

This is the most complicated step in docking for me. The molecule you did have some natural ligand docking? If the answer is yes you can compare them visually with its natural ligand binding and thus determine the best conformations. Otherwise you must use other criteria such as RMSD or the binding energy, taking into account the most stable..

Thanks all, from your view, i think my result should has the most negative binding energy and lowest inhibit constant, is that true?!

Short answer: take the pose with the most negative binding energy

Long answer:

There are two aspects you should consider when analyzing your docking results:

1) binding geometry (docking problem)

2) binding affinity (scoring problem)

I start with the 2nd. You need a function which predicts a binding affinity. Luckily, there are some computational ways to get it, more luckily these ways can be systematically improved. If you wish, you can choose a better theory for binding thermodynamics... Then you select the pose with the highest affinity (e.g. most negative Gibbs free energy). No matter, what quantity the "affinity" in your model/theory is. In our lab, we try to develop a quantum-chemistry-based scoring functions. Honestly, we are approaching, but still are not there... Here's a recent review: http://dx.doi.org/10.1002/cplu.201300199

For binding geometries it is more tricky. There is no systematic way for improvement your theories. There exist plenty of (quite diverse: http://dx.doi.org/10.1002/prot.10115) algorithms and the must be carefully tested on known structural data. Usually, they are not fully transferable between various ligand scaffolds. Based on geometric criteria only, it is in my opinion far more difficult to select the best ligand.

yes most negative free energy change and minimum inhibition constant is the desirable conformation of binding. If very close condition appears for many configuration (which is rare) then choose one of them, based on probable geometry, orientation, number of hydrogen bonds, possible staking interaction etc.

Thanks to all

In addition to the predicted dG you should also pay attention to the size of the cluster. It has been shown that the more populous a cluster is, the more likely it is that that cluster's conformation is correct. The dilemma is when the most populous cluster is not the lowest energy cluster ...

@douglas how you solve the issue of most populous cluster not being the lowest energy cluster?

Does anyone tell me How to get most Populous cluster from DLG file ?

For Autodock we must consider the binding energy of the ligand. Low Binding energy has better docking score.

So, if you want best result of docking I concur with colleaguees, but if you search the better pose, you need to compare all results with substrate and/or analysis the interactions between your ligand and side chains, because not always the low binding energy is the best pose.

same additional questions:

1. is the low binding energy is the best pose for (Conformational isomerisms)

2. why we don't use mean of binding energies?

3. in comparing between different legends, is comparing the (low binding energy) from unknown commonness isomerism are logic?

thank you.