Hi, the purpose of MD after docking calculation is to confirm the poses found by docking. The MD calculation allows to establish if the most populated cluster from docking is also found by MD. In fact the static view provided by docking should be verify by using MD. In your case a MD simulation could be useful to confirm if the major contacts found will be maintained during the MD in order to present more reliable results. In general, the combination of two in silico techniques (docking, MD) could improve the reliability of the results. I suggest to perform MD after docking studies.

It really depends on what you intend to do with the result, and why you ran the docking in the first place:

Previously unknown substrate:

If you didn't know the substrate and you used docking to screen then the best thing is to run a real experiment to validate, this is a must -- in this case the MD may help justify the expense of the assay if that is an issue-- but the experimental data is what matters everything else is about how to get there. (MD is fine, but if you are working with large numbers of compounds I personally favor cross docking (using multiple distinct docking methods as orthogonal data )

Know substrate:

A docking location is suggestive of interactions, a well done MD study-setup will get you more specific details with moderate reliablity. A poorly done MD is just noise collected around you point of interest (probably the docked site).

The real question is what do you want to gain from the result?

Not needed: http://blog.eyesopen.com/ant/2013/11/26/cd-and-md/

Hi Dr Martin Indarte,

So you say that MD is not needed after docking. Can you suggest the tools that I can use to validate the accuracy of my docking results?

For your information, my docking result is comparable with the crystal structure of the same protein-ligand from other organism.

There is no simple answer.  One thing is to do RMSD calcualtions of self dockink ligands on a particular database of protein-ligand complexes, the other way is to calculate enrichment factors.

Hi dear

With out any ...

I would recommend you read bellow paper,I’m sure that you find your solution.

http://www.sciencedirect.com/science/article/pii/S0165614714002144

At the end of the day, Computer Simulations do NOT create new data. They extrapolate meaning from existing data and this can be use full --- For explaining what is going on, or decision making for experiments.

Validity is not the same as real. It's valid in this sense if it is reasonable, comprehensible, 'done correctly', and does not disagree with real data. One way to check this is to do a different computation with the same data and see if the results are highly similar. (different can be an alternate version of the same thing, docking/MD etc. or not-- important to understand how they work before you start picking at them).

For real -- you have to do the experiment, there is no way around it.

Thank you all!

Your opinions and suggestions are greatly appreciated. You are such a great help!

Its an interesting discussion. If I have to run MD simulations, after molecular docking (autodock), which free software is available? I believe, I might have to do for nano seconds. How long it would take for a normal machine? Are there any directions available for that?

Thanks for all your helpful comments that help me a lot when I getting to be familiar to docking studies. I am new to this research field, anyone can help me get information related to ligand-protein docking validation. What are experimental methods that validate the inhibition potential of some ligands to a specific protein target ?

I found above discussion nice. Although it is late but I want to give my view for future readers. As I understand molecular docking is to study interaction of a ligand with a receptor in vacuum which is far away from the real situation. Whereas MD simulation is understanding the dynamics between interaction of ligand and receptor under solution with ions which is much closer to real condition. So if objective is to screen large library one can do molecular docking and do MD with the selected best ligand, but if you want to study an specific ligand interaction with some receptor one should use both molecular docking and MD simulation.

No, but it's a good option.