I performed a homology modeling using modeller after which on evaluation the ramchandran plot showd residues in favored region 91%. After this I performed MD simulation of protein for 10ns for minimization using gromacs. Further, extraction of last frame from completed MD , the ramchandran plot showd residues in favored region 82%. I also checked complete MD ramachandran plot but it showed same results. So why did the model generated showed low results after MD and how should I resolve it. (Results attached)
Before submitting your model to an MD simulation you may refine it using specialized platforms.
I used to obtaing good results with this: https://zhanggroup.org/ModRefiner/
I recommend you submit the model with 91% and check the quality after the refinement if it had not improved enough, submit the generated model until you obtained the desired quality. Please pay attention to the quality values because in determined moment it could decrease, so that means that you shall use the last model with the best quality.
You have to keep in mind that the conformation of an experimental structure represents the equilibrium structure of that molecule under the conditions of the experiment, averaged over a huge number of individual molecules and a large time. A homology model largely mirrors the quality of the template used.
In MD, you add energy to get the conformational fluctuations that represent the potential thermal motions of the molecule. In any snapshot of this process, the conformational energy of the structure usually is higher than that of the equilibrium state. The thermal energy you pumped into the MD simulation is reflected by the additional non-ideal torsion angles. To get to the equilibrium state best represented by the particular snapshot, you would have to bleed off the thermal energy by simulated annealing.
- Badges
- Science method