I would like to validate a homology model. After peroforming MD simulation, one way is to check the stereochemical quality, Ramachandran plot, dssp, B-factor, etc and map the model with the existing crystal/ solution structure. Could someone please suggest any other way of theoretically or experimentally validating the model? Especially, can CD spectra or HDX-mass Spectrometry predict the percentage of loops?
PS - It would be great if people won't suggest me any online server for validation. I am already aware of those servers.
hi
The following servers ma be useful
http://sysbio.rnet.missouri.edu/apollo/
http://d.pcons.net/
http://services.mbi.ucla.edu/ERRAT/
Did you do molecular dynamics simulation with Gromacs?
If you want to check the structure of the model after performing the molecular dynamics simulation, you can use it to analyze the gmx gyrate and analyze the second structure.( gmx do_dssp)
Please follow the link below
http://manual.gromacs.org/programs/gmx-do_dssp.html
http://manual.gromacs.org/programs/gmx-gyrate.html
Do you want to select the best model from the multiple models?
@Balachandra, I have got only one model, however, I am keen to know your answer even for an ensemble of structures.
@Milad, Thanks for your answer. I am aware of DSSP but that is not my question. I should rewrite my question in detail. My question was regarding validating my model. I feel your answer would suit only if I was aware about the secondary structure of missing parts (homology model predicted loops), however, that is missing in the crystal structure. I would like to know what additionally can be done apart from what I mentioned in my question.
If you have ensemble models, use proq3 or SVMQA to select the best model. These two are the best methods in CASP12 blind prediction.
You can also analyse the energy of your model protein here (Qualitative Model Energy ANalysis or QMEAN). I think having a low energy output would be a good verification for what you are searching for.
https://swissmodel.expasy.org/qmean/
Hello
In my opinion, you can employ bioinformatics tools such as- ProSA, ProQ, ERRAT & RAMPAGE for evaluating your homology models. These tools will provide you with C scores, Overall Quality, LG scores etc... and thus differentiating the best model from the rest.
Best of luck.
Regards
@Sahar Qazi: all the programs that you have mentioned that was developed long back. Moreover, ProQ has an updated version ProQ3, which was the second best method in model selection according to the CASP12 experiment.
- Badges
- Science topic