Speed & miscellaneous

[jaketanderson]

Hi Yash, I suggest you avoid trying to stop a docking iteration if it takes too long and instead capping the maximum number of evaluations. See the vina documentation where it shows how you can define max_evals. I don't know what a good number will be, you will need to try different numbers until you find one that prevents protein-ligand pairs from going over your threshold time.

Is there a reason for using this specific box center?

YuGilsonLab-DataAnalysis/AutodockPDB.py

Line 23 in 90ec615

v.compute_vina_maps(center=[15.190, 53.903, 16.917], box_size=[20, 20, 20])

It looks like here you dock until you hit 20 poses but are only interested in taking the top 5, so maybe make those values match. And then I'm also wondering if you really need more than one pose. If you're comparing accuracy to experiment then you can probably just take the one very best pose and use that. Unless maybe you're planning on doing a more complicated analysis over the top n number of poses.

YuGilsonLab-DataAnalysis/AutodockPDB.py

Lines 20 to 25 in 90ec615

	def dock_vina(receptor_file, ligand_file):
	v.set_receptor(receptor_file)
	v.set_ligand_from_file(ligand_file)
	v.compute_vina_maps(center=[15.190, 53.903, 16.917], box_size=[20, 20, 20])
	v.dock(exhaustiveness=32, n_poses=20)
	v.write_poses(ligand_file[:-5] + "_docked.pdbqt", n_poses=5, overwrite=True)

It might be worth recording to file somewhere the scores and energies of all the docked poses you are given. It should be straightforward to do so using the documentation. It would be interesting seeing how different the energies are for the same compound in DOCK6 and in Vina, regardless of whether they docked it close to its experimental location. But that's probably a secondary priority for this analysis.