Hi Fernandez, 

Not sure if Autodock Vina uses LGA, probably uses Broyden-Fletcher-Goldfarb-Shanno algorithm (http://vina.scripps.edu/manual.html#faq).

Also look into their publication

Article Software News and Update AutoDock Vina: Improving the speed ...

Ligands bind into pockets / cavities . . . 

The ADVina molecular docking engine uses a so-called gradient optimization algorithm. Please check our publication for more information.

Article Multidrug resistance protein P-Gp interaction with nanoparti...

 Thank you for your answers. Also, I would like to confirm this process that I have done. First, I did a blind docking to my protein since I do not know its binding site. Then I noticed certain clusters of ligand conformations. The lowest energy confirmation binded to a different site while the rest binded to other sites. One of the sites have 5 conformations in it and I took this site for another calculation. Is this process valid?

A more efficient approach would consist of these two steps:

1. Find (potential) binding sites: pockets, cavities, conserved amino acid residues, etc. 

2. Dock into the (potential) binding sites.

Ren Vien Anthony Fernandez

"Eventually, we settled on the Iterated Local Search global optimizer, similar to that by Abagyan et al. In this algorithm, a succession of steps consisting of a mutation and a local optimization are taken, with each step being accepted according to the Metropolis criterion. In our implementation, we use Broyden-Fletcher-Goldfarb-Shanno (BFGS) method for the local optimization, which is an efficient quasi-Newton method."

Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem. 2010;31:455–461.