You may also generate an ensemble of starting structures by molecular dynamics, cluster the conformers generated and use representatives of the various clusters for independent docking runs.

Hello Abdo

Choosing the right residues as Flexible, makes the job easy for AutoDock to mimick the nature. For Choosing the residues, for flexibility, consider two foremost things :

1. Are there any loop structures in close vicinity of Active site.

2. Does the Active site work by Open and closed conformation or not.

For the first case, If there are loop regions in close vicinity of active site, then try to visually inspect how many amino acids residues may be directly or indirectly affecting the conformation of Active site. Try to use all those residues as Flexible. Even a region of 15-20 A, may affect the active site conformation.

For the Second case, if the active site work by open or closed mechanism. Then be sure that the conformation which is important for activity of Receptor protein, those residue which control open and closed gating mechanism, try not to set them to be flexible.

regards

Tanuj

You may also generate an ensemble of starting structures by molecular dynamics, cluster the conformers generated and use representatives of the various clusters for independent docking runs.

It depends on the size of the ligand . . .

Is there any true advantage of flexible over rigid docking?

Dear Annemarie Honegger

That's mean I do MDS then use rigid docking?

Hello Abdo

As Annemarie suggested, MDS can assist you in gaining insights of how the conformation of your protein vary in solvent system.

Use the ensemble of protein structure, derived from overlap of protein conformation, to visualize how the residues around Active Site vary in Conformation. Using that information, select out various conformer which you observe to have variation in vicinity to Active Site. Individually dock the ligand with different conformer and see which conformation has best binding energy with the ligand.

Alternatively, you can perform Protein- Ligand rigid docking and do MDS of lowest energy pose of the Protein-Ligand complex. And observe the conformational changes in protein-ligand system.

Also, gmx-PBSA can be performed on the trajectory file generated using MDS, to gain insights of interaction of Protein-Ligand in solvent system.

regards

Tanuj

I'd think very carefully before using flexible residues in docking. The most likely outcome is to add noise to what is already quite a noisy technique, in terms of the error and noise in the scoring function.

At most I would allow only the most flexible and most central residues of the active site to move, i.e. the ones that might get in the way of a good binding mode for the ligand. In such cases it may be good to use a program that allows you to select a small number of reasonable conformers for the flexible side chains, such as GOLD. However if your aim is simply to allow awkward residues to get out of the way of the ligand, practically speaking you might do just as well to chop the relevant residue(s) off at the beta carbon (i.e. mutate to alanine) .

I would prefer to allow flexible residues and loops at the active site to only take up conformers/positions I know they can access. For this reason I would favour, instead, rigid ensemble docking using a number of different crystal structures for the target taken from the PDB, to create the ensemble. However you can only do this if multiple crystal structures exist.

A third alternative technique to consider is rigid docking followed by molecular mechanics relaxation, allowing selected sidechains to take part in the minimisation.

It is important to assess the potential flexibility of your potential binding site before docking - if you dock into a pocket from which a previous ligand has been deleted (structure of a complex), and the backbone around your putative binding pocket is mostly involved in regular secondary structure, you may get away with just allowing for side chain mobility. If you are docking to an Apo structure, or if flexible loops make up part of the binding pocket, MDS can help you to assess which parts are most likely to be mobile, and to assess whether the movements are so large that you have to use multiple conformers or which parts of your molecule should be allowed to adapt to what degree. Also, it you have multiple structures of your protein, compare them carefully for indications of mobility - differences between structures, B-factors, crystal contacts that may influence local conformation.

I really work on protein model (there is no solved structure yet). I performed MDS for the model and I'm trying to test small molecule binding.