Molecular Dynamics Simulations HEXA

Run the MD simulation

Here we want to give a receipt for how to analyse the MD simulation result as we did it in our section.

• check the trajectory

First of all we checked the trajectory, to see if our simulation finished successfully and the file is not corrupted.

gmxcheck -f traj.xtc

• Visualistation

Next we want to visualise our results:

trjconv -s topol.tpr -f traj.xtc -o protein.pdb -pbc nojump -dt 10 
pymol protein.pdb 

Still TODO

• create a movie and skip the g_filter step

Still TODO

• energy calculations for pressure, temperature, potential and total energy

In the next analysis step we calculated the energy values for pressure, temperature, potential and total energy with following commands:

echo 13 0 | g_energy -f ener.edr -o pressure.xvg 
echo 12 0 | g_energy -f ener.edr -o temperature.xvg 
echo 9 0 | g_energy -f ener.edr -o potential.xvg 
echo 11 0 | g_energy -f ener.edr -o total_energy.xvg 

We visualised the results of the different runs with the xmgrace program:

xmgrace pressure.xvg
xmgrace temperature.xvg
xmgrace potential.xvg
xmgrace total_energy.xvg

• minimum distance between periodic boundary cells

Next, we calculated the minimum distance between periodic boundary cells. A low distance means, that the part of the protein which is in this boundary cell have contacts with itself. This should not be the case, because one part of the protein should not have contacts with the completely equal part of the protein. Therefore, a low periodic boundary cell shows that the quality of the model is bad and the simulation my be wrong. To calculate the minimum distance we used following command:

g_mindist -f traj.xtc -s topol.tpr -od minimal-periodic-distance.xvg -pi 

We visualised the results with xmgrace:

xmgrace minimal-periodic-distance.xvg

• RMSF for protein and C-alpha

In the next step, we analysed the root mean square fluctuations for the complete protein and also for the C-alpha atoms. With the RMSF you can calculate the differences between two nearly identical structures. In our case, we have a lot of very similar structures. In general we use the same structure but over the simulation time, the structure moves and therefore we got a lot of very similar, but not equal structures during the simulation. We calculate the RMSF between the start structure and the average structure, which is the average of all structures calculated during the simulation. Furthermore, we also calculated the B-factors of the different residues of the structures. Therefore, we can get a good insight in the flexibility of the protein structure. Furthermore, we calculate this for the complete protein and the C-alpha atoms, to get the possibility to see how flexible the backbone and the residues are. Therefore, we used following commands:

echo 1 0 | g_rmsf -f traj.xtc -s topol.tpr -o rmsf-per-residue.xvg -ox average.pdb -oq bfactors.pdb -res 
echo 3 0 | g_rmsf -f traj.xtc -s topol.tpr -o rmsf-per-residue_c.xvg -ox average_c.pdb -oq bfactors_c.pdb -res 

We visualised the rmsf-per-residue file with xmgrace. The pdb files were visualised with pymol. Furthermore, we aligned the calculated structures with the start structure with pymol to get a RMSD value. Additionally, we looked at the parts of the protein which are really flexible to see how the structure change over time.

Detailed results

• [Wildtype]
• [Mutation 436]
• [Mutation 485]