I have done docking studies using AUTODOCK VINA, where some ligand binds to allosteric sites but some at unknown sites. Will there be any other druggable pockets in the enzyme? Can I select those ligands for next step?
In autodock vina if you maximize the grid you will run a complete blind docking. If you want dock in a particular pocket then you have to do it in a software which allows you to predict binding site.
Softwares like schrodinger, discovery studio, MOE, etc have algorithms to predict binding site and you can do docking in that specific site. If you have a protein with ligand then you can select that space as a binding pocket also.
To identify which binding is good and which binding site is most efficient, you have to do simulation study.
Even if (and it is a big "if") the docking hit would actually bind into some cavity on the enzyme's surface, this would not necessarily mean the compound is an inhibitor. Very likely, it would just bind without interfering with the protein function.