Good afternoon!

I am new to molecular docking. I have watched lessons on youtube, read forums, but still have questions about the algorithm of the work. I hope to find answers here. Sequence of work in AutoDock4: 1. protein preparation 2. ligand preparation 3. gridbox construction 4. map calculation 5. docking calculation 6. analysis In tutorial of molecular docking on known systems is considered. We know in advance where ligand binds to the protein. But what is the step-by-step algorithm of non-directed docking? And I have no prior knowledge. I think I need to create a gridbox as large as possible (126x126x126, 1 A). If the protein does not fit entirely in the gridbox, then move the center of the gridbox so as to capture the entire structure. That is, at the first step of molecular docking I will have several gridboxes created and calculated. Then I will gradually refine the gridbox values. For example, the next step will be for 126x126x126, 0.375 A maps. And so repeat until the map reaches a size, for example, 50x50x50 0.375A. And if my thoughts are correct, this is where my questions arise. 1. Should the maps overlap with each other in order to cover the whole space? 2. How should I decide which regions of my protein is less favorable? So I can further forget about them and focus only on the most favorable? I look forward to your answers and recommendations. Thank you in advance!