SI2011 track3 Setting up an MD simulation

From NBCRwiki

Jump to: navigation, search

Back to Track 3



In this tutorial, we show step by step how to set up a protein system and submit it to an Molecular Dynamics Simulation in the CADD workflow. As a test system, we will use the structure of H5N1 avian influenza neuraminidase, a drug target for influenza infection.

In order to run an MD simulation with NAMD, we need to:

  • Obtain the initial coordinates of your protein system.
  • Generated a Protein Structure File (psf), which stores all structural information of the protein.
  • Specify the force field used in the simulation.
  • Generate a configuration configuration file, in which the user specifies all the options and conditions that NAMD should apply in the simulation.

For this tutorial, the users will need

Loading Neuraminidase system in VMD

The initial coordinates of the neuraminidase protein has been prepared for you here chainA.pdb. It is taken from the neuraminidase structure in complex with the antivirus drug molecule oseltamivir (pdb code 2HU0). Since we will only perform the simulation of the protein, the drug molecule has been removed from the structure.

Although the neuraminidase structure has four identical monomers, in this tutorial, we will only work with one of the monomers in order to save computational time. However, the user should keep in mind that, since the tetramer is the biological unit of the Neuraminidase, it may be desirable to simulate the entire tetramer in your simulation.

To load the structure, start the VMD program, and go to Main menu -> New Molecule. In the Molecule File Browser choose the file chainA.pdb and click on the load button to load the structure. Your VMD display should look like the Figure below:


Adding Hydrogens and Solvating the Neuraminidase system in VMD

Since crystal structures from the Protein Data Bank does not contain the hydrogen atoms (X-ray crystallography can not resolve hydrogen atoms), we need to add them to our structure before we continue. In order to do that, we will use the Automatic PSF Builder module inside VMD to generate the PSF file of the Neuraminidase system. The final PDB and PSF will contain guessed coordinates for hydrogen atoms. Here, we should emphasize that although the Automatic PSF Builder is very convenient for novice users and, therefore, is used in this tutorial, one should always be careful with the results from the automatic procedure and examine the structures carefully before starting a simulation.

In the Main menu, go to Extensions -> Modeling -> Automatic PSF Builder. See Figure below. A window called AutoPSF will open up. A force field has been already selected, which we will use for our system. By default, the CHARMM force field is used to set up the Neuraminidase system. If you're interested in learning how to use other force fields, such as the Amber FF, please let an instructor know.

In step2, select Protein and click on "I'm felling lucky" at the bottom. If a window opens up asking you to specify the location of the original PDB, click OK and select, in your local directory, the file chainA.pdb.


A new entry called chainA_autopsf.psf will be created in the Main menu. Additonally, you will also find in your local directory the files chainA_autopsf.psf and chainA_autopsf.pdb. If you experienced problems generating these files, please let the instructors know. These files can also be download here chainA_autopsf.psf and chainA_autopsf.pdb.

We should emphasize again that the above procedure, although very straightforward, is only meant for novice users to quickly set up a simulation system. Before running any simulations for your research, you should carefully examine the results from the Autopsf procedure. Meanwhile, we recommend intermediate users to consult more advanced tutorials

As the next step, we will now solvate and neutralize the system. In the Main menu, go to Extensions -> Modeling -> adding Solvation Box. Make sure the PSF and PDB options are set to chainA_autopsf.psf and chainA_autopsf.pdb respectively. In the Box Padding sexton, change all values (Min and Max for x,y and z) to 8. See Figure below.


An additional entry will appear in the Main menu and two additional files will be created your local directory: solvate.psf and solvate.pdb. These files can also be download (solvate.psf and solvate.pdb

Finally, to neutralize the system, go to Main menu -> Extensions -> Modeling -> Add ions

A window called Autoionize will open. Make sure the PSF and PDB options are set to solvate.psf and solvate.pdb respectively. In this example, we will only neutralize the system by adding counter-ions. Select the option that indicates to only neutralize the system. Then, click on the Autoionize bottom, see figure below. Two additional files will be created in your local directory, ionized.psf and ionized.pdb). These files can also be downloaded ionized.psf and ionized.pdb


Running NAMD through the CADD workflow

Now we are ready to run the MD simulation using the CADD pipeline. If you haven't already, start the pipeline program, and go to 'MD analysis'-->'Load workflow'-->'AMDNBCR', as shown in the figure below. This workflow will allow us to run the NAMD program (version 2.8) on the cluster ROCCE at NBCR. We will use this workflow today to run regular MD simulation, and again on Friday to run accelerated MD (aMD) simulation.


In the CADD pipeline window, you will see four input options: 'Config File Browser' allows you to specify the NAMD configuration file (In our case , Neuraminidase.conf), and 'Input Directory Browser' lets you select the directory where the rest of the input files are stored. 'NumProcessors' is the number of CPUs that you will use to run the simulation--given the number of workshop participants, we suggest that you use no more than 16 processors for your job.

The Neuraminidase.conf specifies which type of calculations you want to perform (energy minimization or/and MD simulation) and in which conditions. For example, in this tutorial, our configuration file inform NAMD to carry out first a short energy minimization run (in order to remove any steric clashes existent in our initial structure) followed by an MD simulation of 5000 steps in which the ensemble NVT is applied. In this tutorial, we have provided the configuration file for you, and you should feel free to open it in any text editor to read the configuration file. If you have any questions with the file, please let the instructors know.

Now, please copy the files (ionized.pdb, ionized.psf and Neuraminidase.conf )in the AMDNBCR_0.2_input/ directory. Once you have the files ready, click on the green thunder label on the CADD pipeline menu bar:

These files can also be download here ionized.psf , ionized.pdb and Neuraminidase.conf )


You should find that the 'AMDws' label in the middle of the workflow window turns red:


This indicates that the workflow is running. The simulation should finish within a few minutes.

Examining the output files

Once you finished running all the simulations, you may examine the output files in the directory specified in the previous section. If you used the default settings, the output files should be found at your home directory, under CADDworkflows-->MDAnlaysis-->AMDNBCR_0.2_output.


The Neuraminidase.dcd file is the trajectory of the NAMD simulation, which can be loaded in VMD for visual examination. From your VMD main menu, choose 'File'->'New Molecule' -> 'Browse' and load the file ionized.psf first and then the trajectory file Neuraminidase.dcd. You should be able to animate your trajectory.

Back to Track 3

Personal tools