Difference between revisions of "Task 10 - Molecular Dynamics Simulations"

From Bioinformatikpedia
(Created page with "By now you should all have finished your MD simulations. Two files that were produced are the most important ones, the XTC file (trajectory) and the EDR file (energy). Here are …")
 
(Talk)
 
(One intermediate revision by one other user not shown)
Line 1: Line 1:
  +
==Talk==
  +
  +
[[File:MD_analysis.pdf]]
  +
  +
 
By now you should all have finished your MD simulations. Two files that were produced are the most important ones, the XTC file (trajectory) and the EDR file (energy).
 
By now you should all have finished your MD simulations. Two files that were produced are the most important ones, the XTC file (trajectory) and the EDR file (energy).
   
Line 54: Line 59:
   
 
One more interesting thing should be done. You have to compare the RMSFs for statistically significant difference.
 
One more interesting thing should be done. You have to compare the RMSFs for statistically significant difference.
  +
For this we look at students t-distribution.
   
 
In order to do this follow the steps:
 
In order to do this follow the steps:
   
* Calculate the average
+
* Calculate the average of the delta RMSF between WT and mutant for each residue
   
* Calculate the standard deviation
+
* Calculate the standard deviation for all delta values
   
 
* From the standard deviation you can calculate the standard error dividing it by the square root of the nr of points
 
* From the standard deviation you can calculate the standard error dividing it by the square root of the nr of points

Latest revision as of 13:48, 31 July 2012

Talk

File:MD analysis.pdf


By now you should all have finished your MD simulations. Two files that were produced are the most important ones, the XTC file (trajectory) and the EDR file (energy).

Here are the commands to get a fixed trajectory and one pdb trajectory to create a movie:

The reference is the file with the name *_solv.pdb.

  • `trjconv -f $filename/traj.xtc -s $filename/ref.pdb -pbc nojump -o $filename/traj_nojump.xtc <<EOF\n0\nEOF`;
  • `trjconv -f $filename/traj_nojump.xtc -s $filename/ref.pdb -center -o $filename/traj_center.xtc <<EOF\n1\n0\nEOF`;
  • `trjconv -f $filename/traj_center.xtc -s $filename/ref.pdb -fit rot+trans -o $filename/traj_norot.xtc <<EOF\n1\n0\nEOF`;

The output is now the correct trajectory.

  • `trjconv -f $filename/traj_norot.xtc -s $filename/ref.pdb -o $filename/traj.pdb <<EOF\n1\nEOF`;
  • `g_filter -f $filename/traj.pdb -s $filename/ref.pdb -ol $filename/filtered.pdb -fit -nf 5 <<EOF\n1\nEOF`;

The last two steps were done to get a reasonable PDB file to create a movie with.


As there is a very good tutorial already available you will focus on following the instructions found here: http://md.chem.rug.nl/~mdcourse/analysis1.html[1] Mind, there is a second page of the analysis tutorial.

In this tutorial you will first find steps to assure quality of the simulations itself and secondly analysis of the molecules. In the second part the tutorial addresses a specific protein - skip the specific analysis and questions.

The steps should be applied for WT and the mutants simulated. The results should be compared. Most XVG files used for the graphs contain the raw data and could be further analysed (stdev, average...). Whenever an ECHO is preceding the command this means that these are the selections chosen in the interactive part of the command.


The following steps are done:

  • check the trajectory
  • Visualize in pymol only: the intrafit command is correct this way: intra_fit [filename without filetype ending]
  • create a movie and skip the g_filter step
  • energy calculations for pressure, temperature, potential and total energy
  • minimum distance between periodic boundary cells
  • RMSF for protein and C-alpha
  • Pymol analysis of average and bfactor
  • Radius of gyration
  • solvent accesible surface area
  • hydrogen-bonds between protein and protein / protein and water
  • Ramachandran plot: compare it to the standard ramachandran plots. also compare WT and mutations
  • RMSD matrix
  • cluster analysis
  • internal RMSD

One more interesting thing should be done. You have to compare the RMSFs for statistically significant difference. For this we look at students t-distribution.

In order to do this follow the steps:

  • Calculate the average of the delta RMSF between WT and mutant for each residue
  • Calculate the standard deviation for all delta values
  • From the standard deviation you can calculate the standard error dividing it by the square root of the nr of points
  • Now calculate the ratio: average/stderr
  • For this value check the t-distr (two tails and the degree of freedom is the nr of points)

Is the RMSF significantly different?

Have fun!