ALL Docking is conformation dependent - however different programs handle conformational variability differently and as the user you are reponsible for knowing what these are and that they are implemented sanely for your context.

However, having just perused the hex manual (http://hex.loria.fr/manual800/hex_manual.pdf chapters 4 and 5) I suspect your difference may be in part from site selection differences . Vina requires you to manually define a search space (in Angstroms). Whereas by default hex searches a sphere around the Receptors center of mass.

HEX will also do light sampling for induced backbone changes.

Vina doesn't do anything like this for a given run its will allow you to select side chains to be allowed to freely rotate but thats it.

To Do Something comparable in Vina you would have to generate a number of Receptor poses first then run them.

Since you have the compute capacity to run both, you might want to consider consensus docking with both programs:

something like ...

Run all the conformations you can find as inputs to both programs (and new conformations in the output of HEX should be used as inputs for all ligand in both HEX and Vina until they (the conformations) stop changing (creating new) significantly.

then use the best of each set of Ligand-receptor-program (ignoring conformation; for example ligands as smiles or smart, receptors a primary sequence only) for comparing.

Go ahead am message me if you'd like to discuss workflows -

Thank you dear Jacob.

In detail, I did generate 7 different conformations of my receptor and kept them in two PDBQT and PDB formats for docking in Vian and Hex, respectively. Vina results of docked inhibitors were stable in all 7 poses, in contrast of my expectation. I performed Vina blindly with a grid box encompasses all around each receptors. Conformational changes were also observed in Chimera Energy Minimization Tool after substitutions. As you mentioned all docking programs scoring functions are based on conformations. So there is no chance to find out a relationship of experimentally proved affinity reduction with binding energy results here?

I actually didn’t understand your last part of answer. You mean, docked receptor in Hex be used as a receptor for new docking with both Vina and Hex?

first things first ...

1) that is an odd way to run Vina - typical usage assumes you know something about the binding region on the receptor and the size of the ligands then dock to the relevant area. (see the Vina manual http://vina.scripps.edu/manual.html for Search Space,  Exhaustiveness, & the FAQ as well)

If you are NOT doing a standard method, and are instead surveying the surface.

    First be aware that (as far as I know) know one has run thorough checks on calibration or methodology.

    Second Be sure you have room around the box for the ligand to freely rotate or be separate (Check the position at the end of every run).

    You will need to increase the exhaustiveness setting, and may (should) run replicates with different seeds

2)if you are using the MM function in Hex the resulting receptor conformation will likely be different from the input. 

you can also try allowing side chain flexibility in Vina.

These are non- comparable means of accounting for small-moderate conformational variation.

You can extract the receptor position from either and feed in to the other.

3)Also don't for get to check that you are using the same hydration states of the protein every program handles it differently, and that your ligand pdb&pdbqt are valid representations of the same thing.

4) But a the end of the day Docking won't pick up everything. Each is a Docking program is a distinct conceptualization, optimized for speed first, low alpha (false positives) second, often determined by low RMSD by pose verse crystal structures, and everything else after; energy, beta (false negatives),  with consensus docking (combine results) its easy to nudge (marginally improve) either Alpha or Beta - but to do both or a scoring function you need a good background comparison (see MAOB, PDBbind etc) more often you people will just spend more effort tweaking the conformation set up.

    For more resolution you need to move to something like MD, or MD-QM (FEP etc.) but realistically for most cases you'll get a better ROI (value for effort) from a true orthogonal assay, especially, if you care about real conformation or energy and not comparison.

in your case I'd suggest adding a NMR if viable 2D, or fluorinated probably but talk an NMR person (your looking for basic contacts (shifts) you can map to existing structure or model -- if you don't want to deal with labeled ligand you can them move to something like HADDOCK

Thank you. It was solved. I have used Java based Vina performing PaDEL software before. The data were the same. I do not know why. So I used Vina in cmd command prompt. Outputs were quite relative to what I expected. I think PaDEL is well in a job with a same receptor. Here I had to change the configuration file for a few other receptors. I think after changing the configuration file, PaDEL needs to be rerun.

LOL,

Better yet just use some shell scripts.

Wisualization is important but beyond that know your inputs and config files ,

trust the keyboard over the mouse