I docked 2 proteins and analyzed the docked structures using PDBsum. There is no literature stating any residues involved in the interaction. How do I draw a conclusion from the results I have obtained from PDBSum regarding, which of the structures to choose as most probable?
I might comment that predicted docking based on predicted structures is pretty risky! Depending on the I-TASSER scores there may be significant errors in the models. I would be inclined to consider the results as quite tentative, particularly where sidechain-sidechain interactions are concerned.
I would suggest comparing the docking results and looking for any consensus between the predictions - even though the precise residue-residue interactions might not overlap that much it would be encouraging to see that the interaction was between the same pair of structural regions. Many interactions are determined by shape complimentarity after all.
Something to consider: are the predicted structures similar to those already known? If by some lucky chance you can find two similar structures with a known interaction that would help a lot.
Finally something like the methods used by Weigt's group in finding residue-residue contacts might be a useful compliment to this: http://www.pnas.org/content/106/1/67.long
you may predict the most probable binding pocket of two proteins by pocket finder or Pdb template of both protein are known than align your query proteins structure with their respective template structure and extract the conserved structural fold and important residues of query proteins.
Your question is bit confusing, if you give little more information i may help you.
1. Docking of what (Pr+Lig, Pr+Pr etc) and which software?
2. Have you used PDBSUM for interaction study (LigPlot or DimPlot)?
3. Is your protein is modelled one or taken from PDB?
4. and what exactly you expect from your analysis?
iF you provide this information i will be helpful to understand.
1. docking of protein with protein
2. I have used cluspro software to dock the 2 proteins
3. pdbsum generate gives the residues of interaction.i have not used dimplot or ligplot i think!
4. neither of my proteins had their structures in pdb.. i used I-tasser to generate the structures
5.i want to determine the most probable binding site and residues involved (their is no literature for guidance)
When i docked with cluspro i got 10 most probable structures (which i analyzed using pdbsum generate).. I wish to determine which of them is most probably...
You may check the results of the docking program and how it is scored it your results ! You can check this server http://compdiag.uni-regensburg.de/procos/
hope it helps !
You can use cons-PPISP online server for protein-protein binding site prediction. It will give you some list of binding site. these binding site you can compare with your results and based on that you can choose the most probable structure.
I would also suggest you to choose HADDOCK for protein-protein interaction. This tool will give you some docking score. based on that also you can screened best binding conformation.
I might comment that predicted docking based on predicted structures is pretty risky! Depending on the I-TASSER scores there may be significant errors in the models. I would be inclined to consider the results as quite tentative, particularly where sidechain-sidechain interactions are concerned.
I would suggest comparing the docking results and looking for any consensus between the predictions - even though the precise residue-residue interactions might not overlap that much it would be encouraging to see that the interaction was between the same pair of structural regions. Many interactions are determined by shape complimentarity after all.
Something to consider: are the predicted structures similar to those already known? If by some lucky chance you can find two similar structures with a known interaction that would help a lot.
Finally something like the methods used by Weigt's group in finding residue-residue contacts might be a useful compliment to this: http://www.pnas.org/content/106/1/67.long
when i used i-tasser, i used already known structures of similar proteins as template. I hope that increases he chances of getting more dependable structures for the unknown proteins..?
thank u all for ur help.. I will try as suggested
@thomas: yes similar as in similar in sequence and belonging to the same family (having somewhat similar function). Thank u...i will check them with validation tools...
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools