I am a beginner in the field of structural bioinformatics. I have done some modeling projects however I am always confused in minimizing the modeled structure. One of my teachers told me to minimize the modeled structure after generating the final model. On the other hand, another instructor with whom I did modeling project did not recommend me to minimize the modeled structure. He told that it is worthless to minimize the modeled structure as the structure has already suffered a lot of constraints as it was built using the template structure. In that case, the conformation of the minimized structure will be worse than before. Both fellows tried to convince me in their own respect but still I am confused myself. Could anyone please clarify this so that I will decide whether or not to apply this step in modeling my proteins.
Some applications will automatically perform energy minimization while modeling the protein. some applications we have to do manually energy minimization. But Its a mandatory to perform energy minimization as its included in the protocols of Comparative modeling.But let me know which application you used for modeling. If you are using SPDBV for modeling you have to perform energy minimization. If you are using ICM Molsoft which is a commercial tool it will automatically perform Energy minimization while modeling.
There's no wrong and right here. Although I side with those that consider minimizing the energy heavily subsequently to do more harm than good, it certainly depends on the condition, model and structure and your actual question. It'll take quite some experience to learn those conditions. Be merry and take it on your list of open questions.
It depends what you need your model for. If you build your model to interpret some experimental data, or to check which residues are exposed to solvent, etc. you don't need to run any minimization - it will not change the model much: even if you get improvement of the model quality, the improvement would be marginal, and will not change your answers.
But if you plan to use a model for docking, minimization may be good idea. But try to do constraint minimization, so that residues conserved between model and template do not move.
If your model comes from SWISS-MODEL or Modeller, and does not contain severe errors (like knots or wrongly placed indels that cause backbone passing through the protein core), it will not have clashes and the minimization would not harm. If you see backbone clashes, try to remove them by changing the target-template alignment. If you see side chain clashes, try to remove them using a program like SCWRL (again, in a mode that preserves the conformation of conserved side chains).