ccsb-scripps / AutoDock-Vina

AutoDock Vina
http://vina.scripps.edu
Apache License 2.0
627 stars 212 forks source link

keep key amino acid residues when docking #280

Open TVect opened 9 months ago

TVect commented 9 months ago

Hello, may I ask if it's possible to ensure the interaction of certain key amino acid residues during the docking process?

In fact, we need to dock a series of analogs to further determine which compound might be more effective. During the docking process, we've observed significant conformational differences among these analogs after docking. We would like certain parts of these analogs' conformations to remain relatively consistent post-docking, and to interact with specific key amino acids. I'm not sure if this is feasible or if this is meaningful. :joy: :joy: :joy:

diogomart commented 9 months ago

Hello, In same cases, defining a smaller search space that restricts molecules to part of the binding pocket can help, even though this is not a general solution.

A more general solution is what we call "anchored docking" in autodock-gpu (not possible in vina unfortunately). You would define a SMARTS pattern and pin one or more atoms to some coordinates: https://github.com/ccsb-scripps/AutoDock-GPU/wiki/Anchored-docking

rwxayheee commented 9 months ago

Hi @TVect In addition to what Diogo suggested, if you already have a reference protein-ligand structure, you could do restrained conformer generation for the analogs to ensure their core structures align with your reference ligand (well enough for FEP calculations). The confromers can be ranked in Vina by a single-point score calculation (--score_only) or perform a local search (--local_only) if you don't mind them move a little bit.