Hello everyone!
I have an anti-viral drug target structure with 1087 amino acid residues which are predicted by alphafold. I wanted to know how can I prepare this alphafold predicted protein structure for performing VS and Molecular docking studies. Like should I check for any missing residues or energy minimized the structure before creating its pdbqt file for docking? or use the structure as it is?
Please guide me in solving this problem, as I am new to using an alphafold predicted protein structure.
Thank you,
Alvea
Check the confidence score. Ignore/remove parts predicted with low confidence. See what remains. If the pocket of interest is complete and predicted with high confidence, it can be used directly for docking. Optimize individual protein-ligand complexes. Run molecular dynamics simulations of selected complexes to assess the stability of predicted binding modes and to further optimize the protein-ligand shape and charge complementarity.
Firstly, validate the structure by checking for Ramchandran plot. If required, edit the phi and psi angles using coot and then revalidate. After validation, go for domain analysis (Interpro) and finding the binding pocket (CASTp). Visualize the the binding pocket in pymol and give a check on the residues surrounding the pocket. Now proceed for site specific docking by building the grid box around the pocket.
Hope this helps
Pavan K Madasu Martin Klvana First of all, thank you for your suggestions. I have tried to validate the structure using SAVES program and ProSA web server. The Procheck program is showing me a warning sign for the Ramachandran plot, however, there is 92% of the residues in the allowed region. And z-score in ProSA is -16.63. The overall quality score using ERRAT is 92.9719.
Should I try to run an all-atom molecular dynamics simulation for 10 ns to better conform the structure before performing docking?
Make sure that residues in the disallowed regions doesn't surround the active site cavity or flank near it...If it so, energy minimize the structure and proceed further
Pavan K Madasu My alpha folded protein has three domains, where Domain 3 is catalytically active and has 3 aspartic acid residues in the active site binding pocket. Therefore, I have only selected Domain 3 and run SAVES program again. It is giving me significantly better results compared to before (when I consider entire protein) with 0.0% of residues of Domain 3 being in the disallowed regions.
Should I start docking right away while using only the domain 3 structure?
Sounds great....Proceed further with only Domain 3 as well as use complete protein and go for site specific docking
Geometry is one thing, conformation another: a geometrically fine structure can be conformationally completely wrong. I do not expect serious geometrical problems with AlphaFold structures; I would rather focus on the binding site conformation -- whether it is ready to bind a particular ligand or not.
Martin Klvana Then in that case, should I perform energy minimization for like 1000 steepest descent at the ligand binding pocket of a single domain only?
Energy minimization has to be performed till you reach stabilized values
Alvea Tasneem
Energy minimization might be helpful, after docking. If the ligand cannot be placed in the binding site due to severe steric clashes, run MD on the protein alone and dock into representative energy-minimized MD snapshots.
Pavan K Madasu and Martin Klvana
So finally, should I only take my domain 3 as a receptor file, perform multiple-ligand docking, and then run a 100 ns molecular dynamic simulation after the complexes are formed?
- Badges
- Science topic
- Similar topics
- Filing