My protein sequence has been modeled by swiss-model server. How can I validate the result?
For validation you should go for Ramachandran plot. Read Ramachandran plot.
there is lots of server which can do R-plot for your protein 3D structure.
Link: http://services.mbi.ucla.edu/SAVES/
Good luck
Dear Bakr,
You may validate it through SAVES server. There are 4-5 servers in it like Errat, Verify3D, Ramachandran plot etc. You may run all of those to validate the structure. Also in addition, you may validate it through ProSA web server which analyze the protein structure and matches with X-Ray and NMR calculated structures. Also, you may also analyze your structure through TM-align. I hope this helps you.
Best Regards,
I like the SAVES server too. It returns results very quickly.
http://services.mbi.ucla.edu/SAVES
Here you can also find (with links) Procheck (http://www.ebi.ac.uk/thornton-srv/software/PROCHECK/) and What_Check (http://swift.cmbi.ru.nl/gv/whatcheck/) - I guess, they are the most commonly used for the 3D structure validation.
Procheck returns different structural and geometry characteristics in two files - .sum and .sco - they includes Ramachndran, bonds, length, angle, residue properties, side-chain properties etc. Also Procheck provides its own overall factors, called GFactors (see in *.sum file from results). The GFactors give the average assessment over all other structural characteristics. That common value can be very useful if you want to obtain the common view on the structure or to compare different models.
Procheck you can also download and run from the command line in Linux. Just write some simple script on cshell or bath and you can get very quick validation for hundreds possible protein structures.
Good validation tools also MolProbity (http://molprobity.biochem.duke.edu/) and ProSA (https://prosa.services.came.sbg.ac.at/prosa.php)
Procheck and What_Check both of them are included in SAVES server. So you do not need to run them separately.
All the computational method only tells you how your modelled structure not violating some kind of metric. You cannot say the structure is valid because they look good in the validation servers. As structure modelling bases on the sets of criteria, they will look good those validation methods anyhow. You need experimental evidences.
I would suggest using Model Quality Assessment Methods (MQAPs) such as ModFOLD - http://www.reading.ac.uk/bioinf/ModFOLD/ModFOLD_form_4_0.html , Pcons, ProQ, QMean.
Also have a look at the Protein Modelling portal from the PSI - gives you instructions on how to use model quality estimation - http://www.proteinmodelportal.org/?pid=quality_estimation
Have a look at the CAMEO contineous assessment of these MQAP tools http://www.cameo3d.org/quality-estimation/
Validation does not tell you how much the model is close to reality, but the validation algorithms take into account only certain geometrical parameters, or topological, or even of energy to meet their criteria. The model, only if the protein is globular, must also possess the criteria of globularity and should be tested also for that. However, it is always a prediction in silico, that is, without water. The structure of proteins is totally water dependent. A very accepted test for its good results is to check the stability of the model using molecular dynamics in explicit solvent at neutral pH. It takes more time, but you have a huge number of structural details, even at the atomic level, which allow you to validate the goodness of what you have modeled both in a reasoned way that through experiments in solution. I understand that it takes a lot of time but I don't see alternative solutions.
I would suggest using Model Quality Assessment Programs (MQAPs). Have a look at the protein modelling portal http://www.proteinmodelportal.org/?pid=quality_estimation.
I suggest using the best methods from the CASP assessment such as ModFOLD4
http://www.reading.ac.uk/bioinf/ModFOLD/ModFOLD_form_4_0.html, ProQ, Qmean etc.
These methods will give you an overall estimate of model quality as well as a per-residue error estimation, i.e. which areas of the model are trustworthy and which areas are not.
https://www.ebi.ac.uk/thornton-srv/databases/pdbsum/Generate.html
http://services.mbi.ucla.edu/SAVES/T/?job=170859
Molprobity web server is an excellent validation server for 3D structures
- Badges
- Science topic
- Similar topics
- Correction