On 19 Nov 2009, at 09:34, Anastassis Perrakis wrote: > Dear all, > > I am risking life and limb here, and possibly wining a first row seat to > 'Morten's Inferno' for modelers! > > I want a software that will do some docking of a flexible ligand to a rigid > protein model. Consensus google search suggested me to > use AutoDock 4/Vina. (I am looking for free software). > > However, what I want to do is 'non-standard': I know from a crystallographic > structure (brrrr....) the position of a Vanadate, which has to be (give or > take) > the position of my phosphate atom in my substrate (I have a > phosphodiesterase). Thus this point should get fixed in space (no > translation, or ideally > restrained translation of some sort), while rotation around it should be > allowed, plus that point should be the 'root' of the torsion tree of the > ligand > to allow it to change conformation around it. AutoDock (but not Vina) seems > to be able to do it in simulated annealing mode, by setting the translation > to 0.
Hi Anastassis, I suggest you to stick with Autodock4 for a quick and dirty trick. Define your ligand as a "flexible residue", choose the phosphorus atom as the root and put it in the rigid part of the docking (i.e. with the receptor). Then perform a run using the "do_local_only" option in the dpf file (you have to comment out ga_run in the dpf). It is equivalent to a minimization. That should do the trick. Florian