Hi,
I'm a beginner in molecular docking. I want to know what the purpose of chosen just one chain in docking process? how to determine the most influential binding site of amino acid residue of protein/enzim?
Thanks in advance.
thanks for suggestions. I used blank docking, there are two variables e.g. Scoring and RMSD. what are the RMSD and scoring roles so we can analyse the binding site?
Having multiple chains in one PDB file may have different reasons
a) You have multiple copies of the same protein within the asymmetric unit of the crystal - that's just the way the protein packs to form a crystal, in this case you can use any of the individual chains for docking.
b) The protein is active as a multimer, several identical chains make up the biological unit. Here the active site may reside entirely with one chain, or may be located at the interface between two chains - in the first case, you may get away with docking to the isolated subunit, although the exposed hydrophobic surface of what in the complex would be the interdomain interface may lead to false-positive results. In the second case, using only one chain would not give you an accurate model of your putative binding site.
c)The active protein consists of multiple different protein chains. Again, the binding site you are looking for may either lie entirely within one chain or at the interface of chains, or the additional chains present in a given pdb may represent modifiers of activity, such as proteinacious inhibitors that actually obstruct the active site.
Before you attempt any docking, learn as much as you can about the protein you are investigating - look at the structures, compare asymmetric unit to the biological unit offered by the pdb (although this annotation is not always reliable), read the papers associated with the structures (follow the NLM link) and the UniProt record, look for known binding site motifs in the sequence or active site residues identifies by biochemical methods
If there are multiple entries in the pdb for a given protein, don't just arbitrarily use the first you find, but compare all there are - they may represent a non-liganded structure or may contain different ligands or inhibitors, and by comparing them you can learn a lot about potential conformational changes upon ligand binding, which may critically influence any docking results.
Scoring is a way how your docking programs rank the different results they produce. You will normally get many different results, which you can cluster to simplify matters (Put results which are very similar to each other into one category). An rmsd (root-mean-square deviation) is a measure of how much two different solutions differ from each other or from a pre-defined reference complex (basically, clustering consists of putting all solutions within a low rmsd of each other into a common category):
To test docking algorithms, you start from an experimental complex, take it apart, and allow the docking algorithm to dock the component, the plot the score of each solution against rmsd from the original (experimental solution): in an ideal world, the docking solution with the best score should have the lowest rmsd, and any solution with high rmsd should have a significantly poorer score.
In blind docking (if you do not know the correct answer ahead of time), plotting scores as a function of rmsd from the best-scoring solution will help you to to see whether your docking solutions converge towards one solution (more reliable), or whether very different solutions have close-to-the-best scores (docking results very unreliable.
how could we find the binding site in protein/enzym so we could find out which amino acid residues need to be docked?
Check out the UniProt annotations for your protein and the publication for the newest available structures. If there is experimental evidence for the location of the binding site, it should be cited in those records. There is also a wide variety of servers for binding site prediction available: https://www.google.com/search?rls=en&q=protein+binding+site+prediction+tool
Try out several of those and see whether you can get convergent predictions.
Use CASTp server which allows you to predict the active sites of different proteins. http://sts.bioe.uic.edu/castp/index.html?1bxw
Open this link and click on calculation, then upload the pdb file. Active sites of the protein will be generated within few seconds.
1. The A or B chains in the protein structure are sometimes dimers, so that the A and B chains are the same, one of which can be chosen as a docking target.
2. You can try the web server to find the protein binding site area, for example: http://www.scfbio-iitd.res.in/dock/ActiveSite.jsp
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology