Qwikmd Tutorial

University of Illinois at Urbana-Champaign
Beckman Institute for Advanced Science and Technology

Theoretical and Computational Biophysics Group
Computational Biophysics Workshop
QwikMD - Easy Molecular

Dynamics with NAMD and
VMD
Tutorial by Rafael C. Bernardi, Till Rudack, Joao V. Ribeiro,

Angela Barragan, Muyun Lihan, Rezvan Shahoei and Yi Zhang
July 12 2017
QwikMD Developers:
Joao V. Ribeiro
Rafael C. Bernardi
Till Rudack
A current version of this tutorial is available at
http://www.ks.uiuc.edu/Training/Tutorials/
CONTENTS 2
Contents
1 Introduction 3
1.1 NAMD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 QwikMD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Required programs 4
2.1 For Linux/Mac Users: . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 For Windows Users: . . . . . . . . . . . . . . . . . . . . . . . . . 5
3 Getting Started 5
4 Installing the Required Programs 6

4.1 VMD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4.2 QwikMD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4.3 NAMD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
5 Running my First Molecular Dynamics Simulation 7

5.1 Ubiquitin in implicit solvent . . . . . . . . . . . . . . . . . . . . . 8
5.1.1 Preparing structures and starting simulations . . . . . . . 9
5.1.2 Analyzing during a Live Simulation . . . . . . . . . . . . 11
5.2 Ubiquitin in a Water Box . . . . . . . . . . . . . . . . . . . . . . 12
5.2.1 Starting a New Simulation . . . . . . . . . . . . . . . . . 12
5.2.2 Creating a Salt Solution . . . . . . . . . . . . . . . . . . . 13
5.3 Running your Simulation outside of QwikMD . . . . . . . . . . . 14
6 Tackling common scientific problems 15

6.1 Cancer mutation in Ras . . . . . . . . . . . . . . . . . . . . . . . 16
6.2 HIV protease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
6.3 Proton transport through a membrane by bactheriorhodopsin . . 25
7 Steered Molecular Dynamics 31

7.1 Biomolecular interactions during protein unfolding . . . . . . . . 31
7.1.1 Preparing a SMD system . . . . . . . . . . . . . . . . . . 32
7.1.2 Analyzing during a Live Simulation . . . . . . . . . . . . 33
7.2 Setting-up steered molecular dynamics to study protein complex
interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
8 Structure refinement with MDFF 37
9 Closing Remarks 40
1 INTRODUCTION 3
1 Introduction
The relationship between structure and function of proteins is directly connected
to atomistic aspects of protein dynamics. As a major methodology in structural
biology, molecular dynamics (MD) simulations permit the exploration of the
physical mechanism underlying the function of proteins by examining their dy-
namical behavior. Although the advances in MD simulations are leading to a
new level of knowledge of macromolecular complexes, reaching the million-to-
billion atom regime, the vast majority of MD users are interested in straightfor-
ward MD simulations of relatively simple proteins. To assist both experimental-
ists and any novice to MD to overcome the initial learning curve barrier of MD
simulation software, we developed a VMD-NAMD plugin, QwikMD, that con-
nects the widely employed and user-friendly molecular graphics program VMD
to the widely used MD program NAMD. By employing QwikMD, one is able to
setup an MD simulation in just a few minutes, allowing quick studies of point
mutations, partial deletions and even atomic force microscopy experiments. The
plugin makes it easy for a new user to perform MD simulations, while it also
servers as a learning tool. Many info buttons provide the theoretical background
underlying the MD procedures carried out in modern MD simulations. The info
button windows also provide links to more complete explanations in our website
or in some of our publications. Dont forget to use these buttons to learn more
about QwikMD and MD simulations.
QwikMD in an automatic tool! QwikMD is an automatic tool
that aims to prepare as many different biological system as possible,
however covering all possible systems that are studied employing
MD simulations is a hard quest. Please inspect the system prepared
by QwikMD carefully before starting the MD simulations. We are
working hard to develop new features and extend QwikMD applica-
bility and if you have suggestions to improve QwikMD please contact
us.
1.1 NAMD
NAMD, recipient of a 2002 Gordon Bell Award and a 2012 Sidney Fernbach
Award, is a parallel molecular dynamics code designed for high-performance
simulation of large biomolecular systems. Based on Charm++ parallel objects,
NAMD scales to hundreds of cores for typical simulations and beyond 500,000
cores for the largest simulations. NAMD uses the popular molecular graphics
program VMD for simulation setup and trajectory analysis, but is also file-
compatible with AMBER, CHARMM, and X-PLOR. NAMD is distributed free
of charge with source code.
In order to run any MD simulation, NAMD requires at least four files:
a Protein Data Bank (pdb) file which stores atomic coordinates and/or
velocities for the system. PDB files are available for many proteins at
http://www.pdb.org, or can be generated using various structure mod-
eling programs.
2 REQUIRED PROGRAMS 4
a Protein Structure File (psf) which stores structural information of the

protein, such as various types of bonding interactions.
a force field parameter file. A force field is a mathematical expression of
the potential which atoms in the system experience. CHARMM, X-PLOR,
AMBER, and GROMOS are four types of force fields, and NAMD is able
to use all of them. The parameter file defines bond strengths, equilibrium
lengths, etc.
a configuration file, in which the user specifies all the options that NAMD
should adopt in running a simulation. The configuration file tells NAMD
how the simulation is to be run. For more details check NAMD Tutorial.
Warning! The goal of this tutorial is to introduce QwikMD and

NAMD by performing short molecular dynamics simulations. There-
fore, the examples provided are optimized so simulations can be
done in a reasonable period of time in a common computing facil-
ity. This means that some parameters and conditions under which
simulations are done in this tutorial are not suitable for scientific
studies. Whenever this happens it will be pointed out and alter-
natives or more appropriate parameters/conditions will be provided
in case you want to improve the simulations and/or you have more
computer power available.
1.2 QwikMD
QwikMD is a VMD plugin that aims on helping the user to start and analyze
MD simulations. The plugin helps, in particular scientists that are starting to
perform MD simulations to prepare the necessary files to run these simulations
from desktop machines all the way to large supercomputers. All the necessary
steps - from loading the PDB to generating the NAMD configuration files -
are performed automatically but giving the user all the information necessary
to reproduce the system created and the simulations performed. The Live
View option allows for the visualization and analysis of the simulation on the
fly, helping new users to learn more about MD simulations and expert users
to test their simulations before submitting them to run in a supercomputer.
QwikMD integrates VMD and NAMD, two widely used software developed by
the Theoretical and Computational Biophysics Group at University of Illinois
at Urbana-Champaign.
2 Required programs
The following programs are required for this tutorial:
NAMD: Available at http://www.ks.uiuc.edu/Research/namd/ (for

all platforms). Note that versions prior to NAMD 2.6 do not support the
CMAP correction of the CHARMM forcefield.
3 GETTING STARTED 5
VMD: Available at http://www.ks.uiuc.edu/Research/vmd/ (for all

platforms)
QwikMD: Included in VMD since the version 1.9.3
2.1 For Linux/Mac Users:

Text Editor: Nedit is a text editor which will be used throughout this
tutorial to view and edit some of the files associated with the simulations.
There are others such as pico, emacs, jot, and vi. Feel free to use whichever
text editor you are most comfortable with.
Plotting Program: We will use the free program xmgrace, available at

http://plasma-gate.weizmann.ac.il/Grace/, to view and analyze out-
put data from NAMD simulations. VMD also has an internal plotting pro-
gram which may be used to examine output directly from NAMD log files.
Other useful graphing programs are Mathematica, http://www.wolfram.
com/ (Purchase required), Matlab, http://www.mathworks.com/ (Pur-
chase required), and gnuplot, http://www.gnuplot.info/ (Free down-
load).
2.2 For Windows Users:

Text Editor: We will use WordPad to view and edit some of the text files
associated with the NAMD simulations. You may prefer to use Notepad or
another text editor of your choosing. Microsoft Word is a word processing
program and should not be used as a text editor.
Plotting Program: We will use Microsoft Excel, available at
http://office.microsoft.com/en-us/FX010858001033.aspx, to view
and analyze output data from NAMD simulations. VMD also has an
internal plotting program which we will use to examine output directly
from NAMD log files. Other graphing programs which you may find
useful are Mathematica, http://www.wolfram.com/, Matlab, http://
www.mathworks.com/, and scilab, available at http://www.scilab.org/.
Note that each of these programs, with the exception of scilab, requires
you to purchase software. Scilab is a free program with capabilities nearly
identical to Matlab. If Excel or any of the proprietary programs are un-
available to you, we encourage you to use scilab.
3 Getting Started
Tutoral Files: The simulations produced during the following sections
and all the files necessary to conclude this tutorial are available for down-
load at http://www.ks.uiuc.edu/Training/Tutorials/#qwikmd.
4 INSTALLING THE REQUIRED PROGRAMS 6
4 Installing the Required Programs

4.1 VMD
The VMD source code and binary distributions can be obtained after registering
at the VMD web page. Download the appropriate distribution file with your web
browser. Windows binary distributions are self extracting, so once the distribu-
tion file is downloaded, proceed to the installation directions. For source distri-
butions and Linux binary distributions, uncompress and untar the file. This will
produce a subdirectory named vmd-1.9.3. Unless otherwise specified, all refer-
ences to VMD code will be from this subdirector. To install the pre-compiled
VMD binaries, cd to the extracted folder, open the configure file and set the
installation directory location ($install bin dir and $install library dir)
to the desired location. After the edition and save of the configure file, cd
src and excute the command make install. One can add the installation lo-
cation ($install bin dir) to the PATH environment variable for convenience.
For more information regarding VMD installation and source code compilation
please visit http://www.ks.uiuc.edu/Research/vmd/current/ig/ig.html.
4.2 QwikMD
QwikMD: QwikMD is included in VMD in version 1.9.3 and later.
1 Open VMD by typing vmd in the Terminal window.
Launch VMD. Open VMD graphically by double-clicking on the VMD

icon under Applications in the Finder.
2 Open QwikMD by clicking Extensions Simulation QwikMD menu

item in the VMD Main window.
4.3 NAMD
You can build NAMD yourself or download binaries for a wide variety of plat-
forms. To run QwikMD you need NAMD installed in your machine and available
in your path.
If you start your first MD simulation with QwikMD you will be asked to
specify the path to your NAMD installation. The automatic path assignment
might fail on some operating systems due to individual permission settings. In
case the automatic path assignment fails please set the path manually using
the following instructions.
For LINUX/MAC users:

5 RUNNING MY FIRST MOLECULAR DYNAMICS SIMULATION 7
Setting the Path: To start to use QwikMD you will need to add the namd2
directory to your path in order for the operating system to locate it. To perform
that, add to your .bashrc (Linux) or .Profile (Mac) in your home folder the
following line:
export PATH=complete.path.for.namd:$PATH
Where (complete.path.for.namd) is the complete path to the actual folder
where the namd2 executable is available.
Example:
export PATH=/usr/local/NAMD 2.10:$PATH
If you are new to Linux, visit our guide at:
http://www.ks.uiuc.edu/Training/Tutorials/Reference/unixprimer.html
For Windows users:
Setting the Path: To start to use QwikMD you will need to add the namd2
directory to your path in order for Windows to locate it. This can be accom-
plished by right-clicking Computer or This PC on the Desktop and selecting
Properties Advanced system settings Advanced Environment Variables
(the precise procedure may vary depending on your version of Windows). Un-
der System variables, scroll down and select Path and then Edit. At the end
of the long line in Variable Value, add a semi-colon ; then the full path to
the directory containing namd2 (but do NOT add the executable namd2 at the
end). Click OK. Also make sure that you (and not only root) have permission to
execute namd2. Now open a new command prompt. Regardless of the directory
you are in, you should be able to type namd2 and run it.
5 Running my First Molecular Dynamics Sim-

ulation
Molecular dynamics (MD) is a computer simulation of physical movements of
atoms and molecules, widely used to study biological systems. The atoms and
molecules are allowed to interact, giving a view of the motion of the atoms and
consequently of protein domains in the case of protein simulations. MD has
emerged as an important research methodology covering systems to the level of
millions of atoms.
Histidine Residues. Of the 20 amino acids, histidine is the only

one which ionizes within the physiological pH range (7.4). This
effect is characterized by the pKa of the amino acid side chain. For
histidine, the value is 6.04. This leads to the possibility of different
protonation states for histidine residues in a protein, and makes the
consideration of the proper state important in MD simulations. The
viable states are one in which the nitrogen of histidine is protonated
(listed with residue name HSD in the topology file), one in which
the nitrogen of histidine is protonated (HSE), and one in which
both nitrogens are protonated (HSP).
5.1 Ubiquitin in implicit solvent

To start setting up a MD simulation with QwikMD and NAMD you will first
load a PDB file. If this is your first time working with MD simulations of proteins
you can use, as an example, the PDB code 1UBQ. Ubiquitin (PDB code 1UBQ)
is a small regulatory protein that is present in almost all tissues (ubiquitously)
of eukaryotic organisms. Here we are going to simulate Ubiquitin with implicit
solvent, where water molecules are represented by a dieletric constant.


3 Load the PDB by typing the PDB code in the blank space and clicking
Load. For this tutorial we are going to work with Ubiquitin, PDB code:
1ubq
Representations. To help the user to recognize the patterns of

the system, QwikMD loads the structures using the more common
representions. You can change the representation in the selection
window of the QwikMD by clicking on the current representation
and selecting a different option. Changing the representation of the
protein to Licorice and the color to Name you will note that the
X-ray structure from the Protein Data Bank does not contain the
hydrogen atoms of ubiquitin. This is because X-ray crystallography
usually cannot resolve hydrogen atoms. The pdb file you will gener-
ate with psfgen along with the psf will contain guessed coordinates
for hydrogen atoms of the structure. Later, energy minimization of
the protein will ensure their positions are reasonable.
To perform MD simulations one has to mimic the environment of the protein,

or any other molecule of interest. The most common solvent is water and there
are two main ways to mimic the solvent effect. Either simulating all the atoms
of the solvent - explicit solvent model - or by adding dielectric constant to the
electrostatic calculation - implicit solvent model. Next you will find a description
of these models as well as a description on how to add salt to the water solution
in order to make a more realistic solvent model.
Implicit Solvent
An implicit solvent model is a simulation technique that eliminates the need

for explicit water atoms by including many of the effects of solvent in the inter-
atomic force calculation. For example, polar solvent acts as a dielectric and
screens (lessens) electrostatic interactions. The elimination of explicit water
accelerates conformational explorations and increases simulation speed at the
cost of not modeling the solvent as accurately as explicit models.
But be careful, because implicit solvent models eliminate explicit water
molecules and represent water in an averaged manner, implicit solvent models
are considered less accurate than explicit solvent models. Always use caution
when employing implicit solvent for molecular dynamics research.
Generalized Born Implicit Solvent

Generalized Born implicit solvent (GBIS) models are one particular class of
implicit solvent models. There are two parts to a GBIS calculation. First, the
Born radius of each atom is calculated. An atoms Born radius represents the
degree of exposure of an atom to solvent. Atoms on the surface of a protein are
highly exposed to solvent, their electrostatic interactions will be highly screened
and their Born radii will be small. Atoms buried in the center of a protein will
not be very exposed to solvent, their electrostatics wont be screened much
and their Born radii will be large. Second, inter-atomic electrostatic forces are
calculated based on atom separation as well as the geometric mean of the two
atoms Born radii.
QwikMD uses Generalized Born Implicit Solvent when the Implicit Solvent
option is selected. You can learn more about Generalized Born method in the
manuscript Tanner, D.E. et al., J. Chem. Theory Comput., 7:3635-3642, 2011
http : //www.ks.uiuc.edu/Publications/Papers/paper.cgi?tbcode = TANN2011A.
X-ray Crystallography. X-ray crystallography methods utilize

the optical rule that electromagnetic radiation will interact most
strongly with matter the dimensions of which are close to the wave-
length of that radiation. X-rays are diffracted by the electron clouds
in molecules, since both the wavelength of the X-rays and the di-
ameter of the cloud are on the order of Angstroms. The diffraction
patterns formed when a group of molecules is arranged in a regular,
crystalline array, may be used to reconstruct a 3-D image of the
molecule. Hydrogen atoms, however, are not typically detected by
X-ray crystallography since their sizes are too small to interact with
the radiation and since they contain only a single electron. The
best X-ray crystallography resolutions currently available are around
0.9A.
5.1.1 Preparing structures and starting simulations

4 To perform simulations in implicit solvent we need to remove the oxygen
atoms from water molecules that were present in the crystal structure. To
do that, above the Selection window, click on Chain/Type Selection
and deselect the water molecules A and water. This will make the oxygen
atoms (in red) disappear.
5 Lets first run a Live Simulation with standard protocols. Mark the
Live View box under the Simulation Setup option and click Prepare.
You will have to select a folder where the input file (.qwikmd) will be
saved. During this step the autopsf function of VMD is called to create
the necessary files to run the MD simulation. QwikMD will also create
the necessary NAMD configuration files.
When the Prepare button is pressed, a series of operations is performed

to prepare the system and the simulation configuration files with the settings
selected by user. Two folders will be created in the working directory: files
created during the preparation step are stored in the setup folder, while files
needed to run the MD simulations are stored in the run folder.
NOTE:! If you want to run Live Simulation, make sure you have
the Live View box checked before you click Prepare.
QwikMD helps the user to prepare NAMD configuration files to run in a

range of computers, from the largest supercomputers with high-end parallel
platforms to the smallest laptop computers. QwikMD also allows, through
the Interactive Molecular Dynamics interface of NAMD, to run live simulations
where the user can look and analyze the trajectories while they are being created.
6 To run the simulation click in the Start Equilibration Simulation

Button and select the number of CPU cores to be used to perform the
simulation. Interactive Molecular Dynamics (IMD) used in the Live Sim-
ulation may take a few seconds to start. In the console window is possible
that the message message ERROR) Error connecting to localhost
on port 3000 shows up until the connection between VMD and NAMD
is established.
NOTE:! The connection between VMD and NAMD during the
initiation process of an IMD simulation may take from a few sec-
onds to minutes, particularly in explicit solvent simulations. If the
connection is never established, you may want to make sure the
network port 3000 is not blocked by the firewall definitions or if any
errors were introduced during the preparation phase by checking the
simulation log file in the run folder.
Warning!! Always click Finish, Reset or close QwikMD win-

dow before leaving VMD. NAMD activities started from Live
Simulations will continue to run in background for all the steps
requested unless you make sure the simulation was aborted before
leaving VMD.
5.1.2 Analyzing during a Live Simulation

VMD is a powerful tool for analysis of structures and trajectories and should
be used as a tool to think. Numerous tools for analysis are available under the
VMD Main menu item Extensions Analysis. In addition to these built-in
tools, VMD users often use custom-written scripts to analyze desired properties
of the simulated systems. VMD Tcl scripting capabilities are very extensive,
and provide boundless opportunities for analysis. QwikMD provides the user
with some of the most commonly employed analysis tools, also allowing the
analysis while performing live NAMD sections.
During the simulation, QwikMD allows the user to perform common analysis
in two analysis tabs: Basic Analysis and Advanced Analysis. Click on the
Basic Analysis tab on the top and there you can perform a few analyses during
the Live Simulation.
7 First, lets analyze how the Kinetic and Potential energies evolve during
the initial phase of the simulation. Expand the Energies section by
clicking on the I symbol, check the Kinetic and Potential boxes and
then press Calculate. The plot is displayed in the Plots section of
the tab. Each plot can be exported and saved using one of the options
available under the File menu of each individual plot.
8 Change the representation of the protein to licorice and the color to

name and then start the simulation by pressing Start Equilibration
Simulation. First the system is minimized. The atoms are rearranged
in order to optimize the energy. You can observe this in an increase of
the RMSD and a decrease of the potential energy. After the minimization
the system is slowly heated up. This means the system is coupled to an
external temperature bath, which can be observed by an increase in the
temperature plot. By this the system is getting an external energy, which
leads to an increase in the kinetic energy.
9 Follow the evolution of the Temperature while the system undergoes a

temperature ramp. The option to plot temperature is under the thermo-
dynamics tab
10 Calculate the Root-Mean-Square Deviation (RMSD). In the RMSD sec-

tion you can select the region to be evaluated and the superimposition
options. Select Backbone as RMSD evaluation region, check theAlign
Structure box and Backbone as the alignment region to eliminate sys-
tem rotational and translational motions from the calculation and then
click on Calculate.
Root-Mean-Square Deviation. The root-mean-square deviation

(RMSD) is the measure of the average distance between the atoms
(usually the backbone atoms) of superimposed proteins. In the study
of globular protein conformations, one customarily measures the
similarity in three-dimensional structure by the RMSD of the C
atomic coordinates after optimal rigid body superposition. When a
dynamical system fluctuates about some well-defined average posi-
tion, the RMSD from the average over time can be referred to as
the RMSF or root mean square fluctuation. The size of this fluctu-
ation can be measured, for example using Mossbauer spectroscopy
or nuclear magnetic resonance, and can provide important physical
information. QwikMD allows the user to perform RMSD analysis
during live NAMD simulations. More advanced options for RMSD
analysis can be done with VMD plugins available in VMD Main
menu item Extensions Analysis RMSD Trajectory Tool .
5.2 Ubiquitin in a Water Box

5.2.1 Starting a New Simulation
Dont forget to click FINISH and RESET QwikMD
More realistic MD simulations are performed with explicit representation
of every atom of the solvent, usually a solution of water and salt. The water
box created by QwikMD is somewhat big for most studies and the edge size is
defined by:
p
boxedge = ( x2 + y 2 + z 2 ) + 15
where x, y and z are the dimensions of structure in the three axis. Box dimen-
sions in A. The big water box is adopted as a safety measure. It is common to
see large conformational changes in proteins. These changes can make the water
box too small, which is hard to be observed by someone new in the field. Ideally,
one should work with a box, which is large enough that the protein does not
interact with its image in the next cell if periodic boundary conditions are used.
The use of periodic boundary conditions consists in surrounding the system un-
der study with identical virtual unit cells. The atoms in the surrounding virtual
systems interact with atoms in the real system. These modeling conditions are
effective in eliminating surface interaction of the water molecules and creating
a more faithful representation of the in vivo environment.
Different water molecule models can be used in an MD simulations. NAMD
currently supports the 3-site TIP3P water model, the 4-site TIP4P water model,
and the 5-site SWM4-NDP water model (from the Drude force field). As the
standard water model for CHARMM force field, TIP3P is the model employed
in the simulations prepared with QwikMD.



NMR Structures. Nuclear magnetic resonance spectroscopy of pro-
teins, usually abbreviated protein NMR, is a field of structural bi-
ology in which NMR spectroscopy is used to obtain information of
the structure and dynamics of proteins, nucleic acids, and their com-
plexes. It is not the case of the 1UBQ structure, solved by X-ray
crystallography, but NMR structure in the PDB usually have mul-
tiple steps. In order to start a MD simulation one have to select
one of the steps as the initial coordinates. It is usual, when running
more than one simulation of the same system, to select different
initial steps in each of the simulations to improve sampling of con-
formational structure.
Load. For this tutorial we are going to work with Ubiquitin, PDB code:
1ubq
Structure Manipulation. This button will open a new window titled

Structure Manipulation/Check where the user can perform point
mutations, rename molecules that have wrong names (read more be-
low), change protonation states, delete parts of the molecules and
also inspect the structure with a interactive residue/molecule list.
Structure Manipulation/Check is especially important in cases
where one of the molecules/ions have wrong names, or names that
are different from the name used in the CHARMM force field. For
instance, in the example described further in the Cohesin-Dockerin
complex in the section 7.2, the Ca2+ ions have the residue name
CA, whereas this residues is defined in the CHRAMM force field
with the residue name CAL. Structure Manipulation/Check al-
lows the user to rename CA ions to proper Calcium parameters that
will be compatible with the CHARMM force field.
4 To perform simulations in explicit solvent we need to change part of

QwikMD standard protocol. For that, first select the Explicit solvent
model.
5.2.2 Creating a Salt Solution

Ions should be placed in the water to represent a more typical biological envi-
ronment. They are especially necessary if the protein being studied carries an
excess charge. In that case, the number of ions should be chosen to make the
system neutral. The ions present will shield the regions of the protein, which
carry the charge, and make the entire system more stable. They should be
placed in regions of potential minima, since they will be forced to those regions
during the simulation anyway. The psf file contains the charge of each atom
and may be used to determine the charge of the total system or parts of it.
One must set the desired salt concentration when preparing the simulation
with QwikMD. The default Salt Concentration is 0.15 mol/L. Even if the Salt
Concentration is set to ZERO, QwikMD will add ions to neutralize the total
charge of the system. Remember, in a MD simulation with periodic boundary
condition the total charge of the system should be ZERO.
5 Perform a simulation in explicit solvent is not an easy task for your com-
puter. Change the Simulation Time, under the protocol drop arrow, to
0.1 ns, to be able to perform this tutorial simulation in a more realistic
time. Depending on the processing power of your machine, this simulation
may still take some time. Please notice that Simulation time refers to
the production MD protocol step.
6 To run a simulation in background, make sure the Live View box is de-
select and click Prepare. You will have to select a folder where the files
will be saved. During this step the autopsfgen function of VMD is called
to create the necessary files to run the MD simulation. QwikMD will also
create the necessary NAMD configuration files.
7 To run the simulation click the Start Equilibration Simulation But-

ton. VMD windows will be unavailable until the simulation is finished.
5.3 Running your Simulation outside of QwikMD

To run molecular dynamics simulations with NAMD at least four files are re-
quired: a Protein Data Bank (pdb) file a Protein Structure File (psf), a force
field parameter file and a configuration file. During the preparation steps, where
the system might be solvated, ionized, among other procedures, several files are
created. QwikMD stores the files created during the preparation step in the
setup folder, while files needed to run the MD simulations are stored in the
run folder. These two folders are created inside the folder with the same
name as the QwikMD input file (*.qwikmd) defined by the user when preparing
the system. This input file (with .qwikmd extension) allows the user to load
the simulations performed with QwikMD and also previously structure manip-
ulation operation and simulation options, like amino acid residues mutations or
salt concentration.
To run a simulation prepared with QwikMD in a computer cluster or super-
computer, one needs to copy only the run folder.
1 Running simulation in a command line

6 TACKLING COMMON SCIENTIFIC PROBLEMS 15
In a terminal window, navigate to the run folder of the simulation prepared

by QwikMD and type the following command:
namd2 +p <Number of CPUs to parallelize to>
qwikmd equilibration 0.conf > qwikmd equilibration 0.log &
Analyzing Trajectories
To analyze a simulation executed outside of QwikMD, one can load the file with
.qwikmd extension. If any trajectory was generated, meaning if the files with
the extension .dcd exists in the run folder, QwikMD will load the trajectories
into VMD.
2 Load the file Ubiquitin.qwikmd from the tutorial files using the Load
button in the Simulation Setup.
3 Select Initial Structure, qwikmd equilibration 0 and

qwikmd production 1 trajectories to be loaded.
Once the frames are loaded into VMD, one can performed the analysis
present in both Basic and Advanced tabs.
Image Rendering
QwikMD offers a very simple way to render your VMD scene and save it as a
picture, however the options available are limited. For more advanced rendering
options, please follow the VMD Images and Movies Tutorial in our website:
http : //www.ks.uiuc.edu/Training/Tutorials/.
4 To render the current VMD scene displayed in the VMD display, select the
Basic Analysis tab and beneath the Thermodynamics you can select
the option for the image render.
5 Select the image resolution and the image processing method (render) to
be use to produce the image.
6 Press Render and select the location and filename to save the image.
6 Tackling common scientific problems

After getting familiar with QwikMD by applying it to the toy example of ubiq-
uitin you will now apply QwikMD to a real scientific scenario. One common
problem in structural biology is that experiments reveal mechanistically rele-
vant mutations for cellular processes but the detailed atomic structural changes
evoked by the mutation often remain elusive. Here MD simulations can help.
Instead of time consuming and expensive mutational studies through X-ray
or NMR spectroscopic experiments, MD simulations are often a much faster
and cheaper way to gain atomic insights in structural and dynamical changes
through mutations. Prerequisite to solve this problem by MD simulation is only

one experimentally resolved atomic structural model of the wild type structure
as initial model.
6.1 Cancer mutation in Ras

Here, we will use the example of the small GTPase Ras, a crucial switch regulat-
ing cellular signal transduction for cell growth. By GTP hydrolysis to GDP and
Pi, Ras is switched from the active GTP-bound on-conformation to the inac-
tive GDP bound off-conformation and signal transduction is terminated. If the
downregulation of Ras by GTP hydrolysis is disturbed, the nucleus receives an
enduring signal for proliferation, resulting in uncontrolled cell growth that may
lead to cancer. Indeed, about 20% of all cancer tumors are associate to mal-
functions of the Ras protein through the so called oncogenic mutations, which
most off them effect the hydrolysis rate. One famous oncogenic mutation, which
inactivates the switching mechanism of the Ras protein, is the G12V mutation.
In order to investigate the atomistic structural differences between wild type
and the oncogenic mutation, you need first to prepare and simulate the wild
type structure in a similar manner as done previously for ubiquitin.
Ras wild type simulation

1 Load the active GTP bound structure of Ras using the PDB code 1QRA
and click Load.
A warning will appear that one or more residues could not be identified. To
prepare the MD simulation of the Ras protein we need to introduce the correct
parameters for the GTP molecule into QwikMD. The Structures in the PDB
often times consist of not only standard amino acids but also DNA, ligands or
other small molecules. If the structure loaded contains non-standard molecules
for which the CHARMM force field parameters are not included in QwikMD,
you will not be able to prepare the MD simulation without further modifica-
tion. If no parameters exist in the literature, you will need to parameterize
the molecule. Alternatively you can delete the non-identified part if it is not
relevant for structural properties or function. But usually this is not the case.
The non-identified parts are colored with throbbing colors in your structure and
highlighted in red in the Structure Manipulation/Check window. In case of
Ras the ligand GTP could not been identified.
Force field parameters. To run an MD simulation force field pa-

rameters are needed. The force field contains mathematical expres-
sion of the potentials which atoms in the system experience and
which leads to the molecular motion. Every force field needs pa-
rameters to describe the potentials between the atoms. There are
force field parameters available for all standard amino acids, DNA,
several small molecules and much more. However, for some small
molecules no parameters are available yet. Details about parame-
terization are described in the tutorial about the force field Tool Kit
(ffTK), which is a VMD plugin that assists in the process of param-
eterization. However, force field parameterization is something for
advanced users and usually a very challenging and time consuming
process. As the outcome of the simulation strongly depends on the
force field parameters, parameterization should be carried out very
carefully.
2 Press OK in the warning window and click on Structure Manipulation.

You can see all non-identified parts highlighted in red (Fig. 1). In this
case, to add parameters for non-identified GTP, click on Add Topo+Param
and then hit + to select toppar all36 na nad ppi gdp gtp.str from the
tutorial files. Change the type of GTP from nucleic to hetero and press
Apply (Fig. 2). Hit Reset in the QwikMD main window and the force
field parameter for GTP is now successfully added to the database.
Figure 1: Unrecognized residues are mark in red in the Structure

Manipulation/Check window.
3 Reload the Ras structure and now all molecules of the structure should
be recognized correctly by QwikMD. Beside the hetero atoms of GTP
Figure 2: In the Topology & Parameters Selection window one can add miss-
ing topology and parameters information for non-standard residues.
(residue ID 167) and the magnesium ion (residue ID 168), represented as

a small sphere, the Ras structure has some waters molecules resolved in
the crystal, which are revealed as spheres in a VDW representation. These
waters can play a crucial functional role. However, for simplicity we here
delete them by clicking on Chain/Type Selection and unchecking the
third line A and water.
4 Now the structure is ready to set up the simulation. Following the same
steps as in section 4.1.1 for ubiquitin you can prepare a live simulation of
GTP bound Ras. For time reasons just choose implicit solvent. Note, if
you want to run a scientific relevant simulation you should consider to use
explicit solvent as solvation effect are in general mechanistically crucial,
in particular to explain mutational effects.
5 Once you start the equilibration run, navigate to the analysis tab in the
top. Here, you can choose to interactively calculate and plot some im-
portant properties during the MD simulation, like RMSD, energies, and
temperature. Open an analysis window that will plot the RMSD of the
backbone atoms by checking the Backbone option in the RMSD panel and
clicking on Calculate. Uncheck Total and check Kinetic and Potential
in the Energies panel and then press Calculate to open a window for the
kinetic and potential energy plot. In order to generate a plot to follow the
temperature press the Plot button in the Conditions panel.
6 Press Finish to stop the equilibration. Navigate in your file

browser to the run folder in your work directory. The file
qwikmd equilibration 0.restart.coor is the last state structure shown
on the screen as you pressed finish. Load the psf file mol new <psf file
name>.psf, followed by the qwikmd equilibration 0.restart.coor
mol addfile qwikmd equilibration 0.restart.coor into the VMD
session and save the coordinates of all atoms to a file named
Ras WT equil.pdb File Save Coordinates... Selected Atoms: all
Save.
Ras G12V mutation simulation

7 Redo steps one and two for the Ras wild type simulation and then go
to the Structure Manipulation/Check window. Select Mutate in the
table options. Click on GLY with the residue ID 12 and choose VAL
from the scroll down menu. In order to introduce the G12V mutation you
need to prepare the system by using the Prepare button. All simulation
parameters should be the same as for the wild type simulation.
8 Run a simulation for the Ras G12V mutant in the same manner as for
the wild type but choose a different work directory. After finishing the
simulation press Reset and close QwikMD. Then navigate to the run folder
in the G12V work directory and preform the same steps described before
to generate the pdb file to Ras G12V equil.pdb.
Comparison of the RasWT and Ras G12V mutation simulations

9 Open VMD and load via File New Molecule the files Ras WT equil.pdb
and Ras G12V equil.pdb. Open the Representation window in Graphics
Representation. Select in the top menu Ras WT equil.pdb and click in
the top window on the selection all. Then change the Drawing Method
to New Cartoon and the Coloring Method to Color ID 7. Create a
new representation (Create Rep) and select only the GTP by typing
resname GTP in the Selected Atom line. Change the Drawing Methode
to Licorice. Create another representation for the mutated residue by
typing resid 12 in the Selected Atoms line. Change the Drawing Methode
to Licorice and the Coloring Methode to Color ID 7. Now select
Ras G12V equil.pdb in the top menu and create the same representations
as for the wild type but choose Color ID 1.
You can observe that there are almost no overall structural differences be-
tween the inactive oncogenic G12V Ras mutation (red) and the active Ras wild
type structure (green). In addition the site chain atoms added by introducing
valine instead of glycine do not have an effect on the structure or positioning
of the GTP. This reveals that the G12V mutant has no influence on the GTP
bound state and implies that it has to have an effect on the intermediate state.
So far there is no experimental method available to resolve the structure of the
intermediate state.
6.2 HIV protease

In this section, we are going to use QwikMD to prepare HIV-1 proteases molec-
ular dynamics simulations. HIV-1 protease is a retroviral aspartyl protease and
it is essential for the HIV life-cycle, the well known retrovirus causing the Ac-
quired Immune Deficiency Syndrome (AIDS). This protease has been extensively
studied, either computationally and experimentally, and several structural prop-
erties can be found in the literature. In this section we will use MD simulations
to analyze the structure in the bound and unbound state. The HIV-1 protease
flaps are a two very distinctive loops that enclose the substrate in the binding
site and are very stable in the presence of the substrate (bound state). How-
ever, in the absence of the substrate (unbound state), the flaps become very
flexible, were the backbone can move as larger as 7 A (Miller et al. Science,
246, pp: 1149-1152, 1989.).


3 To perform a MD simulation of the HIV protease structure, go to Advanced

Run tab. Load the PDB by typing in PDB ID: 1KJF and press the Load
button.
4 A warning will appear that one or more residues could not be identified.
Always make a visual check by rotating and looking at details of the
molecular system. In the Display window, molecules with throbbing colors
were not identified by QwikMD. Fig. 3
5 To facilitate visualizing the complex structure, on can turn off the rep-
resentation of water molecules present in the crystallographic structure.
Click in VDW representation of water and select OFF. Fig. 4
6 A visual inspection of the loaded structure shows a few small molecules

and ions with their colors throbbing. QwikMD did not identify these
molecules and ions as having force field parameters available, at least not
with the name they are presented in the PDB file.
7 Go to Structure Manipulation/Check window to check what problems

are present. Here, certain topologies and parameters are missing, and
that some torsion angles are considered outliers and marginals in the Ra-
machandran plot.
8 Check the molecules and ions that have missing topologies and parameters,
which are marked in red. The name ACT for Acetate is not the same as
the name in the CHARMM36 force field (Fig. 5)
Figure 3: HIV-1 protease (PDB code 1kjf) loaded into VMD by QwikMD.
During the loading process, QwikMD checks the structure for possible problems
and assign representation schemes for easy structure identification. The protein
is represented in NewCartoon and colored in red, blue and gray, the water
oxygen atoms as VDW and colored in red and the acetate ion molecules is
represented as Licorice and with throbbing colors to mark the unrecognized
residues.
9 To rename the Acetate, select Rename and click on the name ACT. A
list of possible substitutes will become available, select Acetate. QwikMD
automatically detects that other molecules have the same name (resname)
and asks if they should all be renamed. Click Yes.
10 Now that QwikMD recognizes the residue, another check routine compares
if the atoms name in the structure correspond with the atom names defined
in the force field for the selected residue name. In this case, none of the
atoms have the correct name, so QwikMD ask if the user wants to assign
the correct name. Click Yes and rename the atoms as shown in the Fig.
6.
11 After renaming ACT to Acetate and change the atoms name, check if
the problems were fixed by clicking the Check button. Here, the Ra-
machandran Outliers can be ignored. Click on Ignore. To learn how to
fix a problem found by QwikMD, click on the hyperlink (in blue) when a
problem is found.
12 Close the Structure Manipulation/Check window and select the

Advanced Run tab.Here, select Explicit Solvent as solvent model, check
the box Minimal Box and Buffer of 20 A. Then, select 0.15 mol/L as con-
centration of salt NaCl. Set the Temperature to 27 C. In this section
Figure 4: QwikMD main window after loading the HIV-1 protease (PDB code
1kjf). The main table presents the information regarding the different chains
of the structure and the correspondent representation and coloring mode. One
can use this table to change the chain representation by clicking in the value
intended to change and selecting one of the options from the drop-down menu.
we will execute the default protocol steps in the MD tab (Minimization,

Annealing and Equilibration), and changing the n Stpes of the MD
step from 500,000 to 10,000,000, meaning from 1 ns to 20 ns (the default
integration step is 2 fs).
Note: To reduce the size of the files generated by NAMD, you can select one
of the protocol steps and click on the Edit button in the protocol table to edit
the configuration files manually and change the saving frequency values. Be
careful when changing the saving frequencies of DCD (trajectory file), energies
or SMD forces, since the amount of data generated can be insufficient for a
proper statistical analysis of the property to be computed.
13 Select the MD protocol step and then click on the + button to extend
the simulation for another 20 ns. Depending on where you are running
your simulations, some time is beneficial to split one big simulation in sev-
eral smaller steps to overcome, for instance, overcrowded supercomputer
queues and clusters walltime limitations.
Figure 5: The user must use Structure Manipulation/Check window to

change the name of the unrecognized residues (marked in red) to use the cor-
rect parameters in the CHRAMM force field. To rename the residue, select
Rename, click on the residue in the table and select the correct name from
the drop-down menu.
Figure 6: In the Edit Atom window, one can change the residue identifier num-
ber (Res ID) and the atoms name to match the topology file of the CHARMM
force field. This window also display the topological information found based on
the residue name in the Structure Manipulation/Check window and isolate
the molecule to be edited in VMD OpenGl window with the matching atoms
indexes for reference.
The duration of the MD simulations must be according to the system in

study and its stability must be evaluated before collecting frames for analysis.
Simulations stability can be evaluated by analyzing, for instance, the RMSD,
the different energies components, temperature and pressure. For tutorial pur-
poses we will simulate 40 ns, whereas in a scientific application research project,
one should extend this time for timescales encompassing the structural motions
in study.
Warning!! The Minimal Box option reduces significantly the size

of the water box. This option rotates the system to reduce water box
volume and add water molecules to complete a padding of at least
the value selected in Buffer on the surface of the structure. This
option must be used carefully to ensure the protein does not interact
with its image in the next cell if periodic boundary conditions are
used.
14 To prepare all the files necessary for the simulation, click Prepare. This
step can take a few minutes.
15 A visual inspection is always important. Make sure everything looks as

expected. QwikMD automatically makes a very simple water box repre-
sentation, helping users to better observe the protein and the substrates
inside the water box.
You prepared the bound state of the HIV-1 Protease. To prepare the un-
bound state, Reset QwikMD and follow steps from 2 to 13. In this next step
we want to delete the substrate, which can be simply done by:
16 On the Select Chain/Type menu of QwikMD main window, deselect the

chain P and protein. In this way, the substrate will not be considered
during the preparation phase.
17 Prepare the system with a different name for future comparison between
bound and unbound state.
18 To start the simulation in the local computer, click on the Start Mini-
mization Simulation button.
The final results of the simulations prepared previously are located in
the supporting files of this tutorial under HIVProtease folder with
the names of HIVprotease with substrate.qwikmd and HIVpro-
tease without substrate.qwikmd.
Note: The resulting trajectories were stripped of waters and solvent ions
mole for file size reduction. You can remove the water molecules and sol-
vent ions molecules from your simulations by selecting the options Dont
load water molecules and Dont load solvent ion molecules in the Loading
Trajectories window. This process can take a few minutes, however is exe-
cuted once. Options only available for the Unix (Linux & Mac) VMD versions.
In this window, you can also chose to not load all the frames saved in the tra-
jectory files by increasing the Stride number present in the bottom of the
window (available for all platforms).
19 Load the HIVprotease with substrate.qwikmd file using the Load but-
ton in the Simulation Setup section.
20 Select the Initial Structure, MD and MD.1.
21 Calculate the RSMD, with the Align Structure option on and the selec-
tion (chain A B) and protein for both RMSD calculation and alignment
entries, and export the plot using the option Export to ASCII matrix
under the File menu in the plot window.
22 Perform the same analysis for the unbound state by loading the HIVpro-
tease without substrate.qwikmd. You can perform this analysis in the
same VMD session or opening another VMD so more analysis can be
performed to both systems simultaneously.
23 Plot both calculations for better comparison using your favorite plotting
program.
The RMSD analysis returns larger RMSD values and large amplitude of the
RMSD values for the unbound state when comparing to the bound state. You
can isolate the RMSD calculation only to the flaps region (residues numbers
35 to 57 of the chain A and B) by using the selection (resid 35 to 57) and
protein in the RMSD calculation atom selection. You can also calculate the
RMSF (Root Mean Square Fluctuation) to have an insight of which residues
are more flexible.
6.3 Proton transport through a membrane by bactheri-

orhodopsin
Bacteriorhodopsin is a membrane protein in archaea in charge of pumping pro-
tons across a membrane. Powered by the green light, bacteriorhosopsin produces
an electrochemical proton gradient across the membrane which in turn is used by
the ATP synthase to produce ATP. Deep inside bacteriorhodopsin, a molecule
called retinal is buried. Retinal, which contains a string of carbons, changes
conformation under light absorption from a straight form to a bent form. This
conformational change makes the proton transport across the membrane possi-
ble.
In the next section, QwikMD is used to prepare a membrane embedded
bacteriorhodopsin model for MD simulations.
Bacteriorhodopsin simulation
1 Restart VMD.

3 Open Structure Manipulation window.

4 Press Add Topo+Param. We need to add topology information and force

field parameters for retinal.
5 Press + and choose RetinalTop.str in the browser.
6 In the Topology & Parameters Selection window for type choose

protein from the drop-down menu.
7 Press Apply and continue with Yes.
8 Reset QwikMD.
9 Load the structure of bacteriorhodopsin by typing the PDB ID 1QHJ in

the blank space and then clicking Load.
10 Proceed with OK.
11 Select Advanced Run.
12 Click on the Structure Manipulation window.
13 In bacteriorhodopsin, residue Lysine (LYS) 216 of the Bacteriorhodopsin

is linked to a protonated Schiffs base, which by its turn is bound to a
retinal molecule. In the PDB file, the Schiffs base and the retinal are
considered one residue with the resid 300 and resname RET. However, in
the topology file you loaded earlier, the complex Lysine, Shiffs base and
Retinal is defined as one single residue with the resname LYR. In order to
treat LYS + Shiff base + Retinal as one residue, they have to be defined
as the same residue. In the following few steps we will change the Res
NAME for both LYS 216 to LYR and change the Schiff base + Retinal
residue from 300 to 216.
14 Rename the residue LYS 216 to LYR by checking the Rename option (on
the top right corner of the window) and clicking on LYS next to Res ID
216 and choose LYR from the drop-down menu.
15 Press No when prompt to rename all residues named LYS.
16 Press Yes when prompt to change the atoms names to match CHARMM
topologies.
17 Rename the nitrogen atom with index 9 from NZ to N16 to be consistent

with the topology and parameters provided for LYR.
18 Press Ok.
19 Change the table option to Type to change the residue type of Res ID 300
from hetero to protein.
20 Rename the Res ID 300 from RET to LYR.

21 Check option next to Edit Atoms.
22 Select Res ID 300 and press Edit.
23 Change the Res ID from 300 to 216 for all the atoms by clicking on 300
and typing in 216.
24 Press Ok. You can find the compound residue (resname LYR) in Fig. 7
shown in Licorice; the carbon chain is colored in cyan.
Figure 7: Retinal (in licorice representation) embedded deep inside bacteri-

orhodopsin.
25 To delete the lipids, check the circle next to Delete and select the lipids
(Res NAME PH1) and press Apply.
26 Also delete the water molecules that are not internally bound to the pro-
tein, which are RES ID 405 and RES IDs 410 to 444.
27 In the same window, open the Membrane menu on the right panel and for
the Box dimensions choose 100x100. Click on Box to display the borders
of the membrane and make sure the protein is placed reasonably in the
membrane (See Fig. 9). You can use Translate and/or Rotate and the
, , +, ++ buttons to adjust the box to the protein transmembrane
region. Click on Generate once you are satisfied with the boxs position.
Figure 8: The Stucture Manipulation window once all the topologies and
parameters exist.
28 In order to prepare a 10 ns simulation for bacteriorhodopsin in a membrane

surrounded environment with explicit solvent, in the QwikMD main menu
choose Explicit from the Solvent drop-down and 15 A for buffer.
Note: When the protein is inserted in the membrane, QwikMD selects the
solvent model Explicit Solvent and disables the Minimal Box option, since
the solvent box for membrane complexes is created in a slightly different way.
To create the effect of a infinite membrane, the periodic boundary conditions
are defined to be the limits of the membrane (in the membrane axis) and any
water molecules is added to the membrane axis. The solvent molecules are only
placed above and beneath the membrane, with a buffer of the value selected by
the user.
29 Go to the Protocol tab and click on n Steps of the MD protocol and

change the number of steps from 500,000 (1 ns) to 5,000,000 (10 ns).
30 Press the Prepare button under the Simulation Setup section of

QwikMD main window.
31 You can either run the simulation on your machine by pressing Start
MD Simulation in the Simulations Control. QwikMD automatically
recognizes the maximum number of processors and uses them as default
but you can use fewer if you want. Press OK to start the simulation. A
note will pop up stating that the simulation is run in the background so
Figure 9: Bacteriorhodopsin placed in a box, prior to membrane being gener-

ated.
the VMD winow will freeze. Proceed with Yes. Alternatively you can
start each of the steps, for example Minimization, Annealing, Equilibra-
tion, ... in the terminal by going to the run folder and execute NAMD
with the command: namd2 +p <Number of CPUs to parallelize to>
Filename.conf > Filename.log & . Make sure that you execute the
protocol steps in the same order as they appear in the protocol table,
since every step is dependent on the files of the previous step. Note that
whether you start through VMD or the terminal you will have to restart
each of the steps manually. It does not happen authomatically. Running
10 ns for all the systems could take a couple of days, so if you do not want
to wait until the simulations are done you can continue right away with
the analysis using the supporting files of this tutorial, under the Bac-
theriorhodopsin. Note that the dcd writing frequency was set to 100000
steps (every 200 ps) just for practical reasons of the tutorial to avoid pro-
ducing large dcd files. For a real simulation the dcd file should be written
more frequently. One can always remove steps of the trajectory or parts
of the simulation system like the solvent (or hydrogen atoms) during the
trajectories loading process in QwikMD as explained in the section 6.2.
Analysis of the trajectory

32 Go to the Simulation Setup tab and press load and load the .qwikMD
of the system you want to analyze.
33 You can choose the parts of the simulation you want to load in the Loading
Trajectory window. As the trajectory files could become very big you
can reduce the size of data loaded in by the stride option or by filtering
out the parts of the system like solvent molecules or hydrogen atoms.
34 To visualize the changes in secondary structure you have to download

sscache.tcl from the tutorial files into your working directory and open
the TK console in VMD through Extension TK console and source the
file by typing source sscache.tcl and start sscache in the TK console.
35 You can go through the trajectory by using the slider in the VMD main
window.
36 Go to the Basic Analysis tab in QwikMD window and open the RMSD
menu. Check the Align Structure box and press Calculate. Then open
the Plot menu and you can analyze the RMSD there. You can export the
plot to an postscript file, an xmgrace file or to an ASCII matrix or vector
file for further analysis in any plot program.
Vacuum and explicit solvent simulations

In this section, we will set up two other independent simulations of 10 ns in
different environments: one in vacuum and one in explicit solvent (without the
membrane environment).
37 In order to prepare the systems in both cases follow steps 1 to 25. Fol-
low step 26 for the vacuum simulation but not for the explicit solvent
simulation.
38 To set up the simulations: For the vacuum simulation, choose Vacuum

in the drop-down menu for Solvent (Note that for vacuum simulation
NVT ensemble is considered.). For the explicit solvent simulation, choose
Explicit in the drop-down menu for Solvent, check Minimal Box and
select 20 A for Buffer. And then follow steps 29 to 31 for both the
simulations.
Warning!! The Minimal Box option reduces significantly the size
of the water box. This option rotates the system to reduce water box
volume and add water molecules to complete a padding of at least
the value selected in Buffer on the surface of the structure. This
option must be used carefully to ensure the protein does not interact
with its image in the next cell if periodic boundary conditions are
used.
39 Follow steps 32 to 36 for each of these simulations to load the correspond-

ing trajectory and perform basic analysis, e.g. calculate RMSD.
Comparing Vacuum, explicit solvent and membrane + explicit solvent

simulation results for the Bacteriorhodopsin
40 Plot RMSD for all the three simulation systems. You will observe that for
any simulation RMSD increases along the trajectory before it plateaus.
Also, the smallest RMSD is obtained for the membrane-embedded Bac-
teriorhodopsin simulation. The reason is the following. RMSD measures
7 STEERED MOLECULAR DYNAMICS 31
the deviation from the initial structure which in this case was the crystal
structure 1QHJ. In order to grasp the most natural conformation for Bac-
teriorhodopsin, which is a membrane protein, the protein was expressed
in a lipid environment for crystallography. Therefore, the simulation of
membrane-embedded Bacteriorhodopsin results in a more similar struc-
ture to the crystal structure (i.e. smaller RMSD values) than the the
structures obtained from simulations in vacuum or water environment
(i.e. larger RMSD values).
7 Steered Molecular Dynamics

Among MD methods, steered molecular dynamics (SMD) simulations in which
external forces are used to explore the response and function of proteins have
become a powerful tool especially when combined with single molecule force
spectroscopy SMFS employing atomic force microscopy. SMD has been suc-
cessfully employed in a wide range of biological systems, from the investigation
of protein mechanotransduction, to permeability of membrane channels, and
the characterization of proteinreceptor interactions. SMD simulations have also
been used to study force propagation through proteins.
7.1 Biomolecular interactions during protein unfolding



SMFS. Single-molecule force spectroscopy has emerged as a pow-

erful tool to investigate the forces and motions associated with bi-
ological molecules and enzymatic activity. The most common force
spectroscopy techniques are optical tweezers, magnetic tweezers and
atomic force microscopy (AFM). In AFM, a microscopic tip that is
located on the end of a cantilever picks up molecules adsorbed on
a surface. Usually the cantilever tips are covalently functionalized
with the molecules of interest. A piezoelectric controller then pulls
up the cantilever. If some force is acting on the elastic cantilever
(for example because some molecule is being stretched between the
surface and the tip), this will deflect upward (repulsive force) or
downward (attractive force). According to Hookes law, this de-
flection will be proportional to the force acting on the cantilever.
Deflection is measured by the position of a laser beam reflected by
the cantilever.
7.1.1 Preparing a SMD system

3 Load the Deca-alanine structure located in the supporting files for this
tutorial by clicking on the Browser button, select the file smd ini.pdb
and then click on Load.
4 A warning will appear showing possible problems found in the structure.

Click in Structure Manipulation to open the Residue Selection window
and check the identified problems in the Structure Check section. The
possible problem identified in this structure is related with the Ramachan-
dran plot outliers, which is not problematic for our case study.
5 To perform SMD simulations, click on the Steered Molecular Dynamics

tab within the Easy Run tab.
6 Click in Protocol options to select SMD parameters.
7 First we will run a simulation with Implicit Solvent, at 0.15 mol/L of

Salt and 27C.
8 To unfold Deca-alanine, set Pulling Distance to 20 Aand Pulling Speed

to 1000 A/ns. That will prepare a simulation of 0.02 ns.
9 To perform a SMD simulation we need to select an amino acid residue to

be pulled. Click in Pulling Residues and a window will open. Select
the first residue (Res ID 1) and click Apply.
10 To perform a SMD simulation we also need to select an amino acid residue

to be anchored, otherwise our molecule would just be translate during the
simulation. Click in Anchoring residues and a window will open. Select
the last residue of Deca-alanine (Res ID 10) and click Apply.
SMD atoms: The VMD display now shows the amino acid residues that
are your pulling atoms and your anchoring atoms.
11 Since we are running the simulation in implicit solvent and the system in
study is relative simple, we will skip the Equilibration and MD and
execute the SMD in the first place. Deselect the steps Equilibration
and MD under the protocol section and leave only the SMD selected.
Note: The SMD forces profile is strictly connected with the intra-molecular
interactions of the different residues. Make sure that you start a SMD simulation
from a well equilibrated structure, particularly if explicit solvent model is used.
12 To run a Live Simulation, make sure the Live View box is selected and
click Prepare. You will have to select a folder where the files will be
saved. During this step the autopsf function of VMD is called to create
the necessary files to run the MD simulation. QwikMD will also create
the necessary NAMD configuration files.
13 To run the simulation click in the Start Simulation Button. Interactive
Molecular Dynamics (IMD) used in the Live Simulation may take a few
seconds to start. In the console window is possible that the message mes-
sage ERROR) Error connecting to localhost on port 3000 shows
up until the connection between VMD and NAMD is established.
NOTE:! The connection between VMD and NAMD during the
initiation process of an IMD simulation may take from a few sec-
onds to minutes, particularly in explicit solvent simulations. If the
connection is never established, you may want to make sure the
network port 3000 is not blocked by the firewall definitions or if any
errors were introduced during the preparation phase by checking the
simulation log file in the run folder.
7.1.2 Analyzing during a Live Simulation

VMD is a powerful tool for analysis of structures and trajectories and should
be used as a tool to think. Numerous tools for analysis are available under the
VMD Main menu item Extensions Analysis. In addition to these built-in
tools, VMD users often use custom-written scripts to analyze desired properties
of the simulated systems. VMD Tcl scripting capabilities are very extensive,
and provide boundless opportunities for analysis. QwikMD provides the user
with some of the most employed analysis tools, allowing also the analysis while
performing live NAMD sections.
During the simulation, clicking on the Basic Analysis or Advanced
Analysis tab on the top you can perform a few analysis during the Live Simu-
lation.
14 To check how much force is necessary to unfold this peptide you can, in
the Advanced Analysis tab, click on Calculate the SMD Force.
15 At the same time, check the number of Hydrogen bonds you are breaking
by clicking on, in the Advanced Analysis tab, Calculate the Hydrogen
Bonds.
Do you see any correlation between the number of Hydrogen

Bonds and the force necessary to unfold the complex?
What if you pull a little slower? How different the Force profile
will be?
How is the force behaving at the end of the simulation? Why?
1000 times slower: Prepare the same simulation, but this time select a pulling
speed 1000 times slower, meaning 1 A/ns and execute the simulation in the back-
ground (Live View option off). Analyze the forces profile and compare with
the simulations performed previously. You can find the final results for this
simulation in the supporting files for this tutorial, under the SMD folder,
Deac-alanine.qwikmd. You can read more about Deca-alanine SMD experi-
ments in the paper also located in the supporting files for this tutorial.
7.2 Setting-up steered molecular dynamics to study pro-

tein complex interaction
To elucidate the molecular mechanisms at play that enable extreme mechanos-
tability in some protein complexes, one can carry out all-atom steered molecular
dynamics simulations of such complexes. One of the strongest protein-protein
interaction ever found, namely cellulosomes cohesin-dockerin interaction, was
characterized using this simulation approach in addition to single molecule force
spectroscopy. SMD results showed the force increased with distance until the
complex rupture. Forces observed were in the range of hundreds to thousands
of pico-Newtons in both simulation and experiment.
Ultrastable protein interactions in cellulsomes. Cellulosomes

are protein networks designed by nature to degrade lignocellulosic
biomass. These networks comprise intricate assemblies of conserved
subunits including catalytic domains, scaffold proteins, carbohydrate
binding modules (CBMs), cohesins (Cohs), dockerins (Docs) and X-
modules (XMods) of unknown function. Coh:Doc pairs form com-
plexes with high affinity and specificity, and provide connectivity to
a myriad of cellulosomal networks with varying Coh:Doc network
topology. To know more check: Schoeler, C., et al. Ultrastable
cellulosome-adhesion complex tightens under load. Nature commu-
nications 5 (2014).


Load. The ultrastable cellulosomal Cohesin:Dockerin complex that we are
studying here can be loaded from PDB code: 4iu3
4 A warning will appear that one or more residues could not be identified.
In the Display window, a few ions and small molecules with throbbing
colors were not identified by QwikMD.
5 On QwikMD window click in the Chain/Type Selection and deselect wa-

ter molecules. As you can see in the Display window, sulfate, the molecule
that was not recognized, is present over the surface of our protein com-
plex. The sulfate in these positions is only present due to crystallization
effect. To remove the sulfate, go to Structure Manipulation and click
in Delete in the right-hand side and click in the residue you want to delete
and click Apply.
6 Calcium ions are presented as CA, which is the CHARMM name for alpha-
carbons. In the right-hand side, click in Rename and then click in the CA
and change the Res NAME to Calcium. QwikMD will offer to change the
other molecules with same name that are also marked in RED. If you are
confident about it click Apply.
7 The Structure Check at the right bottom shows possible structural

problems including missing topologies and parameters. To learn how
to fix a problem found by QwikMD, click on the hyperlink (in blue).
To ignore outliers, click in Ignore. Before closing the Structure
Manipulation/Check window, click in Check in the Structure Check to
confirm that no other problems are present. After check the structure,
press Ignore so mark that you aknowlwdge the cispeptide bond present
in the structure
8 To perform SMD simulations, click on the Steered Molecular Dynamics

tab.
9 Click in Protocol options to select SMD parameters.
10 We will run the simulation with similar parameters that were used in
the Nature Communications manuscript described above. For that, select
Explicit Solvent, at 0.075 mol/L of Salt and 27C.
11 To break Coh:Doc apart we set Puling Distance to 250 Aand Pulling

Speed to 0.25 A/ns. That will prepare a simulation of 1000 ns. Note that,
if no unfolding is observed, the protein complex will break apart much
earlier. In this kind of simulation, always check the results periodically to
avoid bad usage of computer time.
12 To perform a SMD simulation we need to select an amino acid residue to

be pulled. Click in Pulling residues and a window will open. Select
the last residue of chain A - Cohesin (Res ID 210) and click Apply.
13 To perform a SMD simulation we also need to select an amino acid residue

to be anchored, otherwise our molecule would just be translate during the
simulation. Click in Anchoring residues and a window will open. Select
the first residue of chain B - Dockerin (Res ID 5) and click Apply.
SMD atoms: The VMD display now shows the amino acid residues that
are your pulling atoms and your anchoring atoms.
14 To run a simulation like this you will need a computer cluster or a su-
percomputer, so make sure the Live View box is deselected and click
Prepare. You will have to select a folder where the files will be saved.
During this step the autopsf function of VMD is called to create the neces-
sary files to run the MD simulation. QwikMD will also create the necessary
NAMD configuration files.
Note: In very long simulations, like the one prepared in this example, one
can split the SMD simulations in several small steps to overcome, for instance,
overcrowded supercomputer queues and clusters walltime limitations. For that,
use the Advanced Run tab and create extensions of the SMD simulations of 10
or 100 ns.
Running your Simulation outside of QwikMD. To run long steered
molecular dynamics simulations with NAMD at least four files are required:
a Protein Data Bank (pdb) file a Protein Structure File (psf), a force field
parameter file and a configuration file. During the preparation steps, where
the system might be solvated, ionized, among other procedures, several files
are created. QwikMD separates the files created in the preparation step in a
setup folder, while files needed to run the MD simulations are in a run
folder. These two folders are created inside the folder defined by the user in
working directory window. With the same name as the folder created by the
user, a file with .qwikmd extension allows the user to load simulations performed
with QwikMD and also previously created preparation steps, like amino acid
residues mutations or salt concentration.
To run a simulation prepared with QwikMD in a computer cluster or super-
computer, one needs to copy only the run folder.
15 Running: Run your simulation by typing in a Terminal window:
namd2 qwikmd equilibration 0.conf > qwikmd equilibration 0.log

&
namd2 qwikmd production 1.conf > qwikmd production 1.log &
namd2 qwikmd production smd 2.conf > qwikmd production smd 2.log
&
8 STRUCTURE REFINEMENT WITH MDFF 37
16 To perform the analysis after running in another computer, you use the
saved QwikMD file to load your simulation in QwikMD and perform your
analysis with the same tools as well as any other VMD tools.
Force Propagation. A combination of network-based correlation

analysis and thermodynamic fluctuation theory allows for the visu-
alization of stiff paths through protein complexes along which force
is transmitted. For the presented here Coh:Doc complex from cellu-
losome, analysis of the force path suggests specific force-propagation
routes nonparallel to the pulling axis that are advantageous for
achieving high dissociation forces. To learn how to perform net-
work analysis check out the Dynamical Network Analysis tutorial.
To know more check: Schoeler, C., et al. Mapping Mechanical
Force Propagation through Biomolecular Complexes Nano Letters
(2015).
8 Structure refinement with MDFF

MDFF for cryo-EM The molecular dynamics flexible fitting (MDFF) method
can be used to flexibly fit atomic structures into density maps. The method
consists of adding external forces proportional to the gradient of the density
map into a molecular dynamics (MD) simulation of the atomic structure. In this
section, we are going to apply MDFF to Adenylate Kinase. The files necessary
for this section are located in the supporting files of this tutorial under the
QwikMDFF folder.
Windows Installation! Currently there is an issue with using MDFF

related tools on Windows computers. In order to use MDFF (and
the volutil plugin) on Windows, you will need to work in the following
directory: C:\Users\Username\AppData\Local where Username is
your Windows username. You can create a new folder to work in,
but it must be under that directory. Also, this may not be under
the C: drive on your installation, but should be under whatever drive
your Users folder is located. To see the AppData folder, you may
need to turn on Hidden items under the View menu in File
Explorer. You should cd to the directory on the TkConsole or set
the directory as your working directory in the MDFF GUI.


3 To prepare a MDFF simulation, remeber to go to Advanced Run tab. Load

the PDB by using Browser to load the file 1ake docked-noh.pdb. Fig. 10
Figure 10: Initial model for the Adenylate Kinase
4 A warning will appear that one or more potential problems were detected.
Always make a visual check by rotating and looking at details of the
molecular system. In the Display window, molecules with throbbing colors
were not identified by QwikMD. Fig. 11
5 Go to Structure Manipulation window, click in Check in the Structure

Check to confirm that no problems are present. Here, the Ramachandran
Outliers can be ingored. Fig. 12
6 To perform MDFF, click on the MDFF tab. We are using the default values
under the Protocol Option, which means that none of the protein atoms
are fixed and restraints are applied to maintain the secondary structure,
cis piptide bonds, and the chirality during MDFF.
7 Check the Live View and click Prepare to set up a working directory
where the input files for MDFF are created.
8 After that, QwikMD closes and directs you to the MDFF GUI.
9 In the MDFF GUI, click on the MDFF File menu. Add the target density
for the MDFF run. Set up a Grid Scaling Factor of 0.3. After the density
is loaded, click on Generate NAMD files. Fig. 13
Figure 11: QwikMD main window after loading 1ake docked-noh.pdb.
10 You may want to change the represenation of the molecule. Open the
Graphical Representaion Window of VMD. Change the representaion fo
the density to an Isosurface drawn as Solid Surface with a Transparent
Material. Fig. 14
11 In the MDFF GUI, go to the IMDFF Connect tab.

12 To check the quality of the fit to the density on the fly, click on the Cross
Correlation Analysis menu and specify the mol ID of the density file
and its resolution (5 A).
13 To start the iMDFF run, press first Submit and then Connect. Depending
on your hardware, it might take between minutes up to half an hour until
the initial structure is successfully fitted to the density.
14 Compare your fitted result with the experimentally observed model for
the target conformation by adding the experimental struture to the VMD
through File New Molecule, Browse and Load 4ake target.pdb, from the
folder you have downloaded in step 1.
9 CLOSING REMARKS 40
Figure 12: The Structure Manipulation window. Check if any problem exists.
Figure 13: In the MDFF GUI, load the density and prepare MDFF simulations.
15 For better comparison, change the Drawing Method to New Cartoon , the
Coloring Method to Color ID and the Color to green (7). Fig. 15
9 Closing Remarks
Computer simulations of biomolecular systems have grown rapidly over the past
few decades, passing from simulating very small proteins in vacuum to simu-
Figure 14: Visualization of the target density (in Isosurface representation) for
the fitting.
lating large protein complexes in a solvated environment. All-atom MD sim-

ulations, employing classical mechanics, allowed the study of a broad range of
biological systems, from small molecules such as anesthetics or small peptides,
to very large protein complexes such as the ribosome or virus capsids. Hybrid
classical/quantum MD simulations allowed the study of enzymatic activity or
polarizable molecules in biological membranes. However, despite its success,
Figure 15: Comparison of the MDFF result (blue) with the experimentally
observed model (green).
MD simulations are still limited in two regards, inaccuracy of force fields and
high computational cost. Such limitations can lead to inadequate sampling of
conformational states, which in turn limits the ability to analyze and reveal func-
tional properties of the systems being examined. All relevant states of a system
must be reached in simulations in order for its dynamics and function to be
meaningfully characterized. Molecular dynamics simulations have always been
viewed as a general sampling method for the study of conformational changes
of biomolecules. However, biological molecules are known to have rough energy
landscapes, with many local minima frequently separated by high-energy barri-
ers, making it easy to fall into a non-functional state that is hard to jump out
of in most conventional simulations.
As discussed, when running MD simulations it is very important to run
more than one replica of the same system. Long trajectories usually also
helps one to sample different conformations. Therefore a long simulation is
important if big conformational changes are expected. If you want to learn
more about sampling and molecular dynamics check our group publications at:
http://www.ks.uiuc.edu/Publications/Papers/

Qwikmd Tutorial

Uploaded by

Copyright:

Available Formats

Qwikmd Tutorial

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Qwikmd Tutorial

Uploaded by

Copyright:

Available Formats

University of Illinois at Urbana-Champaign

Beckman Institute for Advanced Science and Technology

QwikMD - Easy Molecular

Tutorial by Rafael C. Bernardi, Till Rudack, Joao V. Ribeiro,

4 Installing the Required Programs 6

5 Running my First Molecular Dynamics Simulation 7

6 Tackling common scientific problems 15

7 Steered Molecular Dynamics 31

8 Structure refinement with MDFF 37

a Protein Structure File (psf) which stores structural information of the

Warning! The goal of this tutorial is to introduce QwikMD and

NAMD: Available at http://www.ks.uiuc.edu/Research/namd/ (for

VMD: Available at http://www.ks.uiuc.edu/Research/vmd/ (for all

2.1 For Linux/Mac Users:

Plotting Program: We will use the free program xmgrace, available at

2.2 For Windows Users:

4 Installing the Required Programs

1 Open VMD by typing vmd in the Terminal window.

Launch VMD. Open VMD graphically by double-clicking on the VMD

2 Open QwikMD by clicking Extensions Simulation QwikMD menu

For LINUX/MAC users:

5 Running my First Molecular Dynamics Sim-

Histidine Residues. Of the 20 amino acids, histidine is the only

5.1 Ubiquitin in implicit solvent

1 Open VMD by typing vmd in the Terminal window.

Launch VMD. Open VMD graphically by double-clicking on the VMD

2 Open QwikMD by clicking Extensions Simulation QwikMD menu

Representations. To help the user to recognize the patterns of

To perform MD simulations one has to mimic the environment of the protein,

An implicit solvent model is a simulation technique that eliminates the need

Generalized Born Implicit Solvent

X-ray Crystallography. X-ray crystallography methods utilize

5.1.1 Preparing structures and starting simulations

When the Prepare button is pressed, a series of operations is performed

QwikMD helps the user to prepare NAMD configuration files to run in a

6 To run the simulation click in the Start Equilibration Simulation

Warning!! Always click Finish, Reset or close QwikMD win-

5.1.2 Analyzing during a Live Simulation

8 Change the representation of the protein to licorice and the color to

9 Follow the evolution of the Temperature while the system undergoes a

10 Calculate the Root-Mean-Square Deviation (RMSD). In the RMSD sec-

Root-Mean-Square Deviation. The root-mean-square deviation

5.2 Ubiquitin in a Water Box

1 Open VMD by typing vmd in the Terminal window.

Launch VMD. Open VMD graphically by double-clicking on the VMD

2 Open QwikMD by clicking Extensions Simulation QwikMD menu

Structure Manipulation. This button will open a new window titled

4 To perform simulations in explicit solvent we need to change part of

5.2.2 Creating a Salt Solution

7 To run the simulation click the Start Equilibration Simulation But-

5.3 Running your Simulation outside of QwikMD

1 Running simulation in a command line

In a terminal window, navigate to the run folder of the simulation prepared

3 Select Initial Structure, qwikmd equilibration 0 and

6 Tackling common scientific problems

through mutations. Prerequisite to solve this problem by MD simulation is only

6.1 Cancer mutation in Ras

Ras wild type simulation

Force field parameters. To run an MD simulation force field pa-