I working with a ligand that has shown an affinity towards a certain protein. I am interested in knowing its natural binding site using MD. Any suggestions or advice?
Try to make a docking first, using autodock(free) or gold (buying the licence) , and with the generated structures ( protein +ligand), start the MD simulation. Other way, is eprform a QM calculation or QMMM calculation putting QM the ligand and the binding site of the protein, and leaving the rest as MM.
For MD simulations you will need first to have a 3D structure of the complex (for example using docking as mentioned by Melchor Sanchez-Martinez). Then, using MD simulations you can study structure changes of the protein due to the ligand binding (by doing MD simulations of complex and protein separately), binding modes of different ligands to the same (or different) specific binding sites of protein, and also energetics of the ligand-protein interactions using, for example, binding energy using methods such as MM_PBSA or MM_GBSA.