Dear Satyajeet,

This is a very common question people alwalys used to ask. 

The validation is the most important steps in homology modeling, crude models obtained has to be validated. For this purpose, 

1) SAVES server (http://nihserver.mbi.ucla.edu/SAVES/) from NIH MBI Laboratory is mostly useful for analysis. 

The quality can be evaluate using PROSA, VERIFY3D, ERRAT, and AMOEBA.From this server, you can calculate ERRAT plot from which you came to know which residue is creating problem in your model and You can also do loop modeling using ModLoop module available in Modeller depending on your model  based on these residues.

2) Again calculate the RMSD between template PDB and homology model to check how your model is deviated from the original PDB stucture.

3) The most important issue is the active site so please go through the literature and find out the amino acid residues involved in the active site of the same class of protein. In each class there are few conserved amino acid residues are present which are mainly involved in the catalysis. The orientation and the conformations of these residues plays an important role in catalysis. Yo can compare the active site with other enzyme form the same family. If everything goes well then you are on the safe side otherwise you have to refine it again and again to achieve the goal.

4) Do the energy minimization using any one of the available software but be careful that overall structure should not be altered. Again check the quality of the minimised model you will get better results than what you get before.

Futher repeat the same procedure till you get better quality  model in all respect. if you have any query just let me know.

best regards,

Gaurao Dhoke

Use Errat Server, and  Rampage (Ramachandran Plot Assessment)

http://nihserver.mbi.ucla.edu/ERRAT/

http://mordred.bioc.cam.ac.uk/~rapper/rampage.php

Dear Satyajeet,

This is a very common question people alwalys used to ask. 

The validation is the most important steps in homology modeling, crude models obtained has to be validated. For this purpose, 

1) SAVES server (http://nihserver.mbi.ucla.edu/SAVES/) from NIH MBI Laboratory is mostly useful for analysis. 

The quality can be evaluate using PROSA, VERIFY3D, ERRAT, and AMOEBA.From this server, you can calculate ERRAT plot from which you came to know which residue is creating problem in your model and You can also do loop modeling using ModLoop module available in Modeller depending on your model  based on these residues.

2) Again calculate the RMSD between template PDB and homology model to check how your model is deviated from the original PDB stucture.

3) The most important issue is the active site so please go through the literature and find out the amino acid residues involved in the active site of the same class of protein. In each class there are few conserved amino acid residues are present which are mainly involved in the catalysis. The orientation and the conformations of these residues plays an important role in catalysis. Yo can compare the active site with other enzyme form the same family. If everything goes well then you are on the safe side otherwise you have to refine it again and again to achieve the goal.

4) Do the energy minimization using any one of the available software but be careful that overall structure should not be altered. Again check the quality of the minimised model you will get better results than what you get before.

Futher repeat the same procedure till you get better quality  model in all respect. if you have any query just let me know.

best regards,

Gaurao Dhoke

Very interesting question. I would add another tool PROSA2 which really helps in validating the protein structure by comparing the global energy profile of model to energy profiles of a non redundant set of good quality models.

https://prosa.services.came.sbg.ac.at/prosa.php

You can use PBDsum for model validation by ramachandran plot and for this you have to upload your model pdb in PDBsum here

http://www.ebi.ac.uk/thornton-srv/databases/pdbsum/Generate.html

hello, you may refer procheck and errat for your model valiation

First do ramachandran map analysis and then do superimpose your model with your experimental structure using pymol.

Yes as Govidarajan said, do superimposition with your template and try to draw ramachandran plot. If your modeled structure shows more than 95% of similarity with template in both the case, then you can predict structure is good.

You can use the link below to validate your Homology Model.

http://swissmodel.expasy.org/docs/structure_assessment

The program WHAT_CHECK is also an elaborate validation program for structure models.

You can use:

PROSA, ANOLEA, VERIFY3D, RAMPAGE, ERRAT, SWISSPROT Tools.

If you want to check/know whether the model predicted is is secondary structure you can use GOR V, JPRED, SOPMA.

Use Saves SAVES server and PROSA...

you can use PROCHECK or rampage web-server. The generated output are very detailed. 

Please find these link where PROSA, Procheck, ERRAT plot, Verify 3D

Link Evaluation Webserver  link:

http://nihserver.mbi.ucla.edu/SAVES/

https://prosa.services.came.sbg.ac.at/prosa.php

In this paper, you will get their implication:

Paper Link

https://www.researchgate.net/publication/248382055_Cystathionine__Lyase_like_Protein_with_Pyridoxal_Binding_Domain_Characterized_in_Leishmania_Major_by_Comparative_Sequence_Analysis_and_Homology_Modelling

Article Cystathionine β Lyase like Protein with Pyridoxal Binding Do...

You can use PROSA, PROCHECK, SAVES, RAMPHAGE, etc servers to validate your model.

Is it a model of Glogular or Membrane Protein ?

You should check:

Structural consistency, i.e., the values of bond distances, angles, and dihedrals are within the range of canonical values for each type of bond. That can be don with WHATIF, PROCHECK or simple using a forcefield and searching for high energy sites.

Native-like neighbouring contacts, i.e., side chains in contact are not chemically aberrant (too many negatively charges side chains, for instance), and the distributions of  non-local contacts of amino acids follows trends observed in PDB. These can be checked with statistical potential such as ANOLEA, PROSA or ERRAT.

Finally what I named as "Biological appropriateness", i.e., simply the folding pattern selected has high probability of hosting your sequence, this in comparison  to what we can find in the PDB. This can be checked with VERIFY_3D or Rd.HMM. VERIFY-3D can give both false positives or false negatives, but Rd.HMM does seem to be strongly biassed to false negatives. In other words, a marginal, but acceptable VERIFY-3D score may indicate that your model is right, but in reality is an incorrect folding, carefully refined in sillico to "look good". Rd.HMM instead is likely to rate as inappropriate a prediction with a correct fording, but a poor refinement, but when approved, you can be almost certain to have a useful native-like prediction, even is there are no close homologes deposited in the PDB. However, Rd.HMM is laborious and does not work well for small proteins or intrinsically disordered proteins.

I suggest you read this papers:

http://www.cicy.mx/Sitios/Journal/pdf/2011/October/HighDefinition/IBCJ_2011_1_1_26S.pdf

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0012483

Best wishes

Rogelio

thanks to all

http://www.scfbio-iitd.res.in/software/proteomics/protsav.jsp. That is Protsav, it is a metaserver that uses a combination of different tool including ProsA-web, P

You can use GalaxyRefine for refinement before validating, ProsA is good too

Hi,

Validation of developed structures can be done using tools such as-

i) ERRAT

ii) ProSA

iii) ProQ

iv) RAMPAGE