I assume you meant how to decrease the binding energy. Mind the negative sign. If the binding energy increases, that will mean less favourable binding. To decrease binding energy you simply need a better ligand.

@uttampal yes you are right i mean how to decrease binding energy.

If you want to decrease the binding energy, you could simply try out other docking programs (e.g. AutoDock Vina, ZDOCK, HDOCK, HawkDock). Different programs utilize different (although sometimes similar) docking algorithms to add a ligand to your protein of interest. This may give you a more "ideal" binding energy. Some docking programs are better suited for specific proteins/ligands. You should try others and then compare the results. From here, you can use your best judgement as to which output best reflects the "actual" binding of your ligand. Hopefully, you have some previous experimental work to go off on.

Do you happen to have an existing co-crystallized structure of your protein-ligand complex? If so, you could attempt redocking to see if that gives you the same binding energy as well. You didn't provide a whole lot of background in regards to this, so I'm not sure if this approach is even applicable in your case.

It's also possible that you did exactly what you should have and that's just simply the binding energy you get. It may not matter if you adjust your docking parameters again. Although many strides have been made to improve the accuracy of these docking methods, they are still by no means perfect.

Alternatively, and as first mentioned by Uttam Pal, you could try using a better ligand that allows for a more favorable binding interaction with your protein of interest. Although, I'm not sure you should do this if you're working with very specific ligands for your research.

Hope this helps...even a little bit.

Good Luck!

-Noah

Noah Schuster gave some excellent suggestions. Here are some more:

1) Repeating the docking several times may give you a better result. It often does. Docking is a stochastic method and every time you run it, you’ll get a different result. See the attached file and ref1. All the more reason to run it several times and do a clustering (ref2).

ref1

Article Evaluation of therapeutic effect of Premna herbacea in diabe...

ref2

Article Novel anti-inflammatory activity of epoxyazadiradione agains...

2) You have used the exhaustiveness of 8, which is the default value. So, you can increase it and that might give you a better solution.

3) Use several different crystal structures of the target protein. Generally multiple entries are available in the PDB. Try to obtain as many structures as you can and dock your ligand with all of those crystal structures. You’ll be amazed how much the binding energy differs from one target crystal structure to the other. See Table 5 and Table S1 of the following paper (and the attached file).

Article Probing the Hydrogen Bond Involving Acridone Trapped in a Hy...

4) Vina does not use user supplied partial charges (http://vina.scripps.edu/manual.html#faq Why don't the results change when I change the partial charges?). So adding partial charges is not necessary. Besides, you can only add one type of charge. When you added Kollman charges and then again added Gasteiger charges, the Gasteiger charges replaced the previously applied Kollman charges. You should be aware of that especially when using AutoDock 4.

5) Binding energies obtained by different scoring functions should not be compared. Compare the binding energy of ligand with that of the positive controls (self docking experiments with known ligands) obtained by same scoring function or compare it with the experimental results as suggested by Noah Schuster. And use the docking method that best describe your experimental observations. In my experience, AutoDock 4 underestimates binding energies as compared to that of obtained by fluorescence spectroscopy, whereas Vina results are more close to the fluorescence spectroscopy results. PatchDock with FireDock refinement gives slight overestimation (I can share you the papers if you want). You need to try different methods to find out which method gives you better results. Lower binding energy given by a scoring function could be a overestimation. So, lower prediction is not always a better prediction.

Use good software having good search algorithm..it will generate number of possible conformation per ligand and will give you the best ligand with highest binding affinity towards receptor.

I will suggest Glide and FlexX for this purpose.

And if you want to do it by autodock or vina you need to focus on lead optimization techniques to get better binding affinity.

I agree with Kiran Lokhande

Try different docking softwares and do multiple simulations. You can implement MM/GBSA or MM/PBSA calculations.