Dear Kapil,

If you know the binding site of the ligand to the protein, so, you should just pick the best pose (the lowest energy) in the binding site as the binding pose. If you does not know the binding site, you should take all the protein surface as the potential binding site, do the blind docking and then use further analysis tools (such as MD simulation or experimental evidances) to identify the main binding site. In this case, the binding energy is not enough evidance to specify the binding site.

You should know that the calculated binding energies by docking tools consist of the different contributting forces to the binding process such as H-bonds, electrostatic interactions, pi-pi interactions and ... . However for the publication, it depends what you want. You could report the binding energy, contributing and driving forces, H-bonds, involved residues in the binding process and also design some mutation experiments...

Best

The short answer is: Docking does not work.

The long answer is: Docking sometimes works, but the docking score does not tell me which is the "docking works" case and which is the "docking does not work" case.

---

Use all relevant experimental information and intuition to select the best candidates and evaluate them by estimating the protein-ligand binding free energies using molecular dynamics simulations coupled with free energy evaluation. The free energies are directly comparable to experimental binding affinities; docking scores are not.

But you will probably notice large uncertainty in the calculated binding free energies. So, in the end, you will realize that the free energy calculations are not very reliable . . . but when one says "free energy" it must be right, right? Wrong.

Thank you so much for your guidance.

So the conclusion as per my knowledge is

1) Predict the binding site of the protein( from Expt evidence).

2) then look for energy score from docking tool in the binding region.

3) Do the simulation to confirm the reliability of the autodock vina..

4) Then publish it.

Is it correct ??

or i made a gamble in understanding.

Plz help..

Thanking You.

Kapil: In the second step, one really needs to explore the complexes with one's eyes and employ one's intuition in selecting the poses for further evaluation. The third step needs to be followed by experimental validation; of course one can publish blind predictions but without the experimental validation any structural information that one obtains from the calculations/simulations is meaningless.

Of course, a lot also depends on "what you are looking for", "why are you doing the docking in the first place", and "what is the that you want to publish".

Docking is supposed to help to reduce the chemical space in search for ligands that bind to the target protein but for practical purposes (e.g., drugs) the binding affinity is just one of many aspects and often not really that much important . . . which is why "virtual screening", which is just a fancy word for a simple loop in the code, has not resulted in an increase in approved drugs. Most "docking-based" drugs are obtained via multi-pronged approach in which docking is usually small-scale and targeted and constitutes only a small part in a largely empirical approach that is not transferable to other cases. And then there is the serendipity element, which is responsible for most success stories. But in the end, most (bio)pharma companies sell only one drug: HOPE. And they do it quite openly (just check their websites: smiling people and "HOPE" in capital letters).

Thank you @ Martin Klvana ...

The words are so true and helped me in understanding the actual scenario

Thank you..