I know some computational chemists who generate many, like a hundred, different conformations of the ligand to dock it the protein. I use Glide, and it considers ligand flexibility.
So, it is still required to generate so many conformations?
Thanks.
I am not familiar with the Schrödinger package, but in general, having a flexible ligand does not guarantee that you are doing a good sampling of the conformational space. The flexible ligand has limited time to go through different conformations. You would get a better sampling by starting with different conformations in the first place.
Also, I found this in the Glide manual:
"Glide can dock ligands with up to 300 atoms and 50 rotatable bonds. Since many peptide ligands may exceed these limits, this effectively limits docking application to peptides of around 11 residues or less. In some cases, it is necessary to freeze the amide bonds in their trans configuration to dock peptides of longer lengths."
So, it may also be due to limited capability of the software for including all the rotational degrees of freedom. If you freeze amide bonds, you are eliminating the chance of visiting certain conformations. So, for long ligands, it makes sense to use different constrained conformations in the beginning.
Dear Mr. Lakkaniga,
In Glide docking, "exhaustive enumeration of ligand torsions generates a collection of ligand conformations that are examined during the docking process".
https://www.schrodinger.com/science-articles/docking-and-scoring
https://www.schrodinger.com/kb/1034
Thanks.
Best wishes,
Sk. Abdul Amin
Docking algorithms aim to find the conformation with global minimum on the potential energy surface. But, generally, they can't, especially for large sized ligand. Starting from many different conformations will increase the possibility to reach the lower potential energy points, which correspond to the better conformations.
No, it is not required to pre-generate conformations for flexible docking. It does not make much sense and is time consuming. However, I am one of those computational chemist who usually do that, because I had cases when I could not dock ligands correctly (even quite rigid sometimes, also in Glide) unless I pre-generated conformations. So my advice, try different methods yourself and find what works best in your case.
- Badges
- Science topic