University of Illinois at Urbana–Champaign

NIH Center for Macromolecular Modeling and Bioinformatics

Beckman Institute
Computational Biophysics Workshop

NAMD TUTORIAL
Unix/MacOSX Version

NAMD Developers: James Phillips, David Hardy

NAMD Tutorial Contributors: Tim Isgro, James Phillips, Marcos

Sotomayor, Elizabeth Villa, Hang Yu, David Tanner, Yanxin Liu,
Zhe Wu, David Hardy

April 2017

A current version of this tutorial is available at

http://www.ks.uiuc.edu/Training/Tutorials/
Join the [email protected] mailing list for additional help.
CONTENTS 2

Contents
1 Basics of NAMD 7
1.1 What is Needed . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2 Generating a Protein Structure File (PSF) . . . . . . . . . . . . . 7
1.3 Solvating the Protein . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.1 Ubiquitin in a Water Sphere . . . . . . . . . . . . . . . . 12
1.3.2 Ubiquitin in a Water Box . . . . . . . . . . . . . . . . . . 14
1.4 Ubiquitin in a Water Sphere: Simulation with Non-Periodic Bound-
ary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4.1 Configuration File . . . . . . . . . . . . . . . . . . . . . . 16
1.4.2 Run your Simulation . . . . . . . . . . . . . . . . . . . . . 24
1.5 Ubiquitin in a Water Box: Simulation with Periodic Boundary
Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.5.1 Configuration File . . . . . . . . . . . . . . . . . . . . . . 25
1.5.2 Run your Simulation . . . . . . . . . . . . . . . . . . . . . 28
1.6 Ubiquitin in Generalized Born Implicit Solvent . . . . . . . . . . 29
1.6.1 Configuration File . . . . . . . . . . . . . . . . . . . . . . 29
1.6.2 Run your Simulation . . . . . . . . . . . . . . . . . . . . . 30
1.7 Output: Water Sphere Log File . . . . . . . . . . . . . . . . . . . 31
1.8 Analysis of Water Sphere Equilibration . . . . . . . . . . . . . . . 33
1.8.1 RMSD for Entire Protein . . . . . . . . . . . . . . . . . . 33
1.8.2 RMSD for Protein without Last 5 Residues . . . . . . . . 37

2 Analysis 39
2.1 Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.1.1 RMSD for individual residues . . . . . . . . . . . . . . . . 39
2.1.2 Maxwell-Boltzmann Energy Distribution . . . . . . . . . . 42
2.1.3 Energies . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.1.4 Temperature distribution . . . . . . . . . . . . . . . . . . 47
2.1.5 Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.2 Non-equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.2.1 Heat Diffusion . . . . . . . . . . . . . . . . . . . . . . . . 55
2.2.2 Temperature echoes . . . . . . . . . . . . . . . . . . . . . 59

3 Steered Molecular Dynamics 72

3.1 Removing Water Molecules . . . . . . . . . . . . . . . . . . . . . 72
3.2 Constant Velocity Pulling . . . . . . . . . . . . . . . . . . . . . . 73
3.2.1 Fixed and SMD Atoms . . . . . . . . . . . . . . . . . . . 74
3.2.2 Configuration File . . . . . . . . . . . . . . . . . . . . . . 77
3.2.3 Running the First SMD Simulation . . . . . . . . . . . . . 79
3.3 Constant Force Pulling . . . . . . . . . . . . . . . . . . . . . . . . 80
3.3.1 The SMD Atom . . . . . . . . . . . . . . . . . . . . . . . 81
3.3.2 Configuration File . . . . . . . . . . . . . . . . . . . . . . 82
3.3.3 Running the Second SMD Simulation . . . . . . . . . . . 83
3.4 Analysis of Results . . . . . . . . . . . . . . . . . . . . . . . . . . 83
CONTENTS 3

3.4.1 Force Analysis for Constant Velocity Pulling . . . . . . . 83

3.4.2 Distance Analysis for Constant Force Pulling . . . . . . . 86

4 Acknowledgments 88

A PDB Files 89

B PSF Files 90

C Topology Files 93

D Parameter Files 100

E NAMD Configuration Files 108

F NAMD Standard Output 114

G Water Sphere tcl Script 117

CONTENTS 4

Introduction
This tutorial provides a first introduction to NAMD and its basic capabilities.
It can also be used as a refresher course for the non-expert NAMD user.
The tutorial assumes that you already have a working knowledge of VMD
and that NAMD 2.8 or later has been installed correctly on your computer.
For installation instructions, please refer to the NAMD User’s Guide. For the
accompanying VMD tutorial go to
http://www.ks.uiuc.edu/Training/Tutorials/
The tutorial is subdivided in three sections. The first one covers the basic
steps of a molecular dynamics simulation, i.e., preparation, minimization, and
equilibration of your system. The second section introduces typical simulation
techniques and the analysis of equilibrium properties. The last section deals
with Steered Molecular Dynamics and the analysis of unfolding pathways of
proteins. Finally, brief descriptions of all files needed for the simulations are
provided in the appendices.
For a detailed description of NAMD the reader is referred to the NAMD
User’s guide located at
http://www.ks.uiuc.edu/Research/namd/current/ug/
The examples in the tutorial will focus on the study of ubiquitin – a small
protein with interesting properties. Throughout the text, some material will
be presented in separate “boxes". Some of these boxes include complementary
information to the tutorial, such as information about the biological role of
ubiquitin, and tips or shortcuts for using NAMD. These boxes are not required
for understanding the tutorial and may be skipped if you are short on time.
Boxes with an exclamation sign are especially important and should not be
skipped.
Warning! The goal of this tutorial is to introduce NAMD by per-
forming some short molecular dynamics simulations. Therefore, the
examples provided are optimized so simulations can be done in a rea-
sonable period of time on a common computing facility. 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 alternatives or more ap-
propriate parameters/conditions will be provided in case you want
to improve the simulations and/or you have more computer power
available.

Computer Related Material. To aid in completing this tutorial,

a web page of basic UNIX commands has been made available at
http://www.ks.uiuc.edu/Training/Tutorials/Reference/unixprimer.html.
CONTENTS 5

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

• 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/ (Purchase re-
quired), and gnuplot, http://www.gnuplot.info/ (Free download).

Getting Started
• If you are performing this tutorial at a Computational Biophysics Work-
shop offered by the Theoretical and Computational Biophysics Group, a
copy of the files needed for this tutorial have been set up for you. The
files exist in the directory called Workshop in your home directory. In a
Terminal window type:
cd ~/Workshop/namd-tutorial/namd-tutorial-files
This will place you in the directory containing all the necessary files.

• If you have downloaded this tutorial at home, you will also need to down-
load the appropriate files, unzip them, and place them in a directory of
your choosing. You should then navigate to that directory in a similar
manner as described directly above. The files for this tutorial are avail-
able at http://www.ks.uiuc.edu/Training/Tutorials/.
CONTENTS 6

Mac Users. If you have downloaded this tutorial and are not working
at a Computational Biophysics Workshop, Mac-specific instructions are
provided for you in these boxes. These instructions relate to starting
VMD. The tutorial instructs users to start VMD via the command line,
but this is only possible on a Mac if the user has set up “aliases" be-
forehand. Thus, to avoid changing system settings, we recommend you
simply follow the instructions in these boxes instead. However, if you
are able to set up command line aliases on your computer to launch
VMD, you may do so and then ignore these boxes.

Figure 1: Directory Structure for tutorial exercises. Output for all simulations
is provided in an “example-output” subdirectory within each folder shown.
1 BASICS OF NAMD 7

1 Basics of NAMD
In this section you will learn how to use NAMD to set up basic molecular dy-
namics (MD) simulations. You will learn about typical NAMD input and output
files, in particular, those for protein energy minimization and equilibration in
water.
NOTE: You will be generating output data in this section by performing
simulations and using other features of NAMD. These files are needed for Units
2 and 3. If you are not able to produce the output, correct versions have been
provided for each section and may be found in the example-output folder inside
the directory corresponding to the given simulation or exercise.

1.1 What is Needed

In order to run any MD simulation, NAMD requires at least four things:
• a Protein Data Bank (pdb) file which stores atomic coordinates and/or
velocities for the system. Pdb files may be generated by hand, but they are
also available via the Internet for many proteins at http://www.pdb.org.
More in Appendix A.
• a Protein Structure File (psf) which stores structural information of the
protein, such as various types of bonding interactions. More in Appendix
B.

• 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 GROMACS are four types of force fields, and NAMD is able
to use all of them. The parameter file defines bond strengths, equilibrium
lengths, etc. More in Appendix D.

• 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. More in Appendix E.

Force Field Topology File. Later, you will make a psf file for your
system. In doing so, a force field topology file is necessary. This file
contains information on atom types, charges, and how the atoms
are connected in a molecule. Note that the pdb file contains only
coordinates, but not connectivity information! More in Appendix C.

1.2 Generating a Protein Structure File (PSF)

Of the four files mentioned above, an initial pdb file will typically be obtained
through the Protein Data Bank, and the parameter and topology files for a
given class of molecule may be obtained via the Internet at
http://mackerell.umaryland.edu/charmm_ff.shtml.
1 BASICS OF NAMD 8

Notice that the latest release (c31b1) also contains the updated CMAP cor-
rection, used in this tutorial. The psf file must be created by the user from
the initial pdb and topology files. The NAMD configuration file is also created
by the user, with commands based on the specific requirements for the MD
simulation.

CMAP. CMAP is an energy correction map based on quantum

mechanical calculations and is one of the latest additions to the
CHARMM force field. It improves protein backbone behavior and
thus yields more accurate dynamic properties for the protein. For ex-
ample, CHARMM22 with CMAP gives the experimentally observed
α-helix while the CHARMM22 force field without CMAP gives a
π-helix for certain model peptides. More in Appendix D.

1 Go to the 1-1-build directory. In a Terminal window, you can change

directories using the cd command. Type cd 1-1-build. You can see the
contents of the directory you are in by typing ls. In this folder, you will
find many files that you will use later.

First, you will remove the water molecules from 1UBQ.pdb, and create a pdb
file of the protein alone.

2 Open VMD by typing vmd in the Terminal window.

namd-tutorial-files/1-1-build> vmd

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

VMD icon under Applications in the Finder. Then, navigate
to the 1-1-build directory by using the cd command in the
VMD command terminal. For instance, if you placed the tutorial
files in your home directory in a folder called Workshop, you would type:

cd ~/Workshop/namd-tutorial/namd-tutorial-files

3 Load 1UBQ.pdb by clicking File → New Molecule... menu item in the VMD
Main window. In the Molecule File Browser use the Browse... button to
find the file 1UBQ.pdb. Load it by pressing the Load button.

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 generate 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.
1 BASICS OF NAMD 9

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 wavelength
of that radiation. X-rays are diffracted by the electron clouds in
molecules, since both the wavelength of the X-rays and the diam-
eter 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
0.9Å.

4 Choose the Extensions → Tk Console menu item and in the VMD TkCon
window. Be sure you are in the 1-1-build directory. If you are not,
navigate there using the ls command to list files and directories and the
cd command to change directories. Then, type the following commands:
set ubq [atomselect top protein]
$ubq writepdb ubqp.pdb
(Hit the Return key after each command.)

In the previous step you have created the file ubqp.pdb, which contains the
coordinates of the ubiquitin alone without hydrogens, in the 1-1-build direc-
tory.

5 Quit VMD by choosing File → Quit.

6 Now, you will create the psf file of ubiquitin. Note that VMD offers an
automatic psf file builder via the VMD Main menu by clicking Extensions
→ Modeling → Automatic PSF Builder. We will create the psf file manually
to teach you exactly how it is done. The psfgen package of VMD is very
useful in this regard. In order to create a psf, you will first make a pgn
file, which will be the target of psfgen.
In a Terminal window type nedit to open the text editor. Type in the
following lines:
package require psfgen
topology top_all27_prot_lipid.inp
pdbalias residue HIS HSE
pdbalias atom ILE CD1 CD
segment U {pdb ubqp.pdb}
coordpdb ubqp.pdb U
guesscoord
writepdb ubq.pdb
writepsf ubq.psf
1 BASICS OF NAMD 10

7 After typing this, save the file by clicking File → Save. Be sure that you
are in the 1-1-build directory and enter the file name as ubq.pgn. Quit
the text editor by clicking File → Exit.

The file you just created contains the necessary commands to create the psf
file of ubiquitin with hydrogen atoms and without water. Each command of the
pgn file is explained:
• Line 1: You will be running psfgen within VMD. This line requires that
the psfgen package is available to be called by VMD.
• Line 2: Load the topology file top_all27_prot_lipid.inp

• Line 3: Change the residue name of histidine to the proper name found
in the topology file. HSE is one of three names for histidine, based on the
protonation state of its side group. See the science box below for more
information.

• Line 4: The atom named “CD1" (δ carbon) in isoleucine residues is re-

named as “CD", its proper name from the topology file. Since isoleucine
contains only one δ carbon atom, the psf file does not use the number
label after “CD".
• Line 5: A segment called U is created, containing all atoms from ubqp.pdb.
The segment command also adds hydrogen atoms.
• Line 6: Coordinates are read from ubqp.pdb and residue and atom names
are matched. Old segment labels will be overridden with the new segment
label “U".
• Line 7: Coordinates of missing atoms (like hydrogens) are guessed based
on residue definitions from the topology file.
• Line 8: A new pdb file with the complete coordinates of all atoms, includ-
ing hydrogens, is written.
• Line 9: A psf file with the complete structural information of the protein
is written.

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 δ 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").
1 BASICS OF NAMD 11

8 In a Terminal window (again, be sure that you are in the 1-1-build

directory), type the following command:
> vmd -dispdev text -e ubq.pgn
This will run the package psfgen on the file ubq.pgn and generate the psf
and the pdb file of ubiquitin with hydrogens.

Launch VMD in text mode. Open VMD in text mode and run the
package psfgen on the file ubq.pgn by typing the following command
in a Terminal window in the 1-1-build directory:

/Applications/VMD\ <version>.app/Contents/vmd/vmd_MACOSX
-dispdev text -e ubq.pgn

Substitute the version number of VMD installed on your machine for

“<version>".

In your screen you will see different messages. Warnings are related to the ends
of your molecule and are normal. Your system should have 1231 atoms and 631
with guessed coordinates.

9 Type exit in the Terminal to exit VMD.

Having run psfgen, two new files will now appear in your 1-1-build directory:
ubq.pdb and ubq.psf. Check this by typing ls in a Terminal. You have created
the psf! You may want to inspect ubq.pdb using nedit .

Multiple Chains. Often, pdb files will contain structural information

for polymers of proteins. Each monomer is typically labeled as a
separate chain in the pdb file. Chain ID is indicated in the column
of the pdb after the residue name. In those cases, you can use
VMD to make separate pdb files for each chain, using the following
commands in the VMD TkCon window:

set chainA [atomselect top "chain A"]

$chainA writepdb chainA.pdb

These commands should be used for each chain in the sys-

tem. The psfgen package requires a pdb file for each chain.
1 BASICS OF NAMD 12

Figure 2: Flowchart indicating the role of files as used by VMD, NAMD, and
psfgen.

1.3 Solvating the Protein

Now, the protein needs to be solvated, i.e., put inside water, to more closely
resemble the cellular environment. You will do so in two ways, placing ubiquitin
in:

• a water sphere in surrounding vacuum, in preparation for minimization

and equilibration without periodic boundary conditions.
• a water box, in preparation for minimization and equilibration with peri-
odic boundary conditions.

1.3.1 Ubiquitin in a Water Sphere

You will utilize a prepared tcl script to create the water sphere. It is called
wat_sphere.tcl and is located in your 1-1-build directory. The syntax of the
script itself is given in Appendix G.

1 In a Terminal window in your 1-1-build directory, type

vmd -dispdev text -e wat_sphere.tcl
This will call the script, which will place ubiquitin in the smallest possible
water sphere which completely immerses the protein.
1 BASICS OF NAMD 13

Launch VMD in text mode. Open VMD in text mode and run the
script wat_sphere.tcl by typing the following command in a Terminal
window in the 1-1-build directory:

> /Applications/VMD\ <version>.app/Contents/vmd/vmd_MACOSX

-dispdev text -e wat_sphere.tcl

Substitute the version number of VMD installed on your machine for

“<version>".

Water Sphere. The creation of a water sphere is not a necessary step

if one wishes to examine a protein in water surrounded by vacuum.
However, it is a sensible one. If a water box is used, minimization
and equilibration will simply lead the box to deform into the most
stable shape possible, minimizing the surface tension. That shape
is a sphere! Creating the sphere before beginning simulations re-
duces the amount of time needed to achieve equilibration and saves
computational effort.

2 The output of the wat_sphere.tcl script will be the center and ra-
dius of the water sphere. Record these numbers. You will need them
later. The script will have created the pdb and psf files ubq_ws.pdb and
ubq_ws.psf, respectively, of ubiquitin in a water sphere. They will be in
your 1-1-build directory. Check this by typing ls .

3 Type exit in the Terminal to exit the text version of VMD.

4 Launch the graphical version of VMD by typing vmd in the Terminal win-
dow.

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

icon under Applications in the Finder. Then, navigate to the
1-1-build directory by using the cd command in the VMD command
terminal.

We will check the appearance of the immersed molecule by loading the molecule
with both structural and coordinate information.

5 Load the structural information first. Click File → New Molecule... in

the VMD Main window. In the Molecule File Browser use the Browse...
button to find the file ubq_ws.psf. Load it by pressing the Load button.

6 Load the coordinates into the PSF file. Make sure the Load files for: field
in the Molecule File Browser says ubq_ws.psf. Then, click Browse. . ., find
the file ubq_ws.pdb, and load it by pressing the Load button.
1 BASICS OF NAMD 14

You should be able to see your ubiquitin molecule in the center of a beautiful
sphere of water. Changing the representations in VMD will enable you to see it
better.

7 Delete your ubiquitin molecule in a water sphere using the VMD Main
window by clicking on ubq_ws.psf in the window, and selecting Molecule
→ Delete Molecule.

1.3.2 Ubiquitin in a Water Box

1 In the VMD Main window, click Extensions → Tk Console. In the VMD
TkCon window of your VMD session type:
package require solvate
solvate ubq.psf ubq.pdb -t 5 -o ubq_wb

The package require solvate command loads the solvate package, so that
VMD will be able to call it.
The solvate package will put your protein (described in ubq.psf and ubq.pdb)
in a box of water. The -t option creates the water box dimensions such that
there is a layer of water 5Å in each direction from the atom with the largest
coordinate in that direction. The -o option creates the output files ubq_wb.pdb
and ubq_wb.psf for ubiquitin with the water box. You will also obtain the files
combine.pdb and combine.psf which may be eliminated. They are intermedi-
ate files used by the solvate program to transport information into VMD.

Size of Water Box. The water box created is somewhat small for
this study, especially considering that you will pull and unravel the
protein in Unit 3 of this tutorial. 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.
Periodic boundary conditions will be explained in section 1.5. Fur-
thermore, if the protein is being pulled, the box should be large
enough that the water will still significantly immerse the protein
when it is fully extended.

Ions. Ions should be placed in the water to represent a more typical

biological environment. 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.

2 In the VMD Main window, click File → New Molecule. In the Molecule
File Browser use the Browse... button to find the file ubq_wb.psf. Load
1 BASICS OF NAMD 15

it by pressing the Load button. This will load the structural information
into VMD.

3 In the same Molecule File Browser, browse for ubq_wb.pdb and load it. In
the display, check if the water box surrounds ubiquitin, and demonstrate
to yourself that the hydrogen atoms now exist.

4 In the VMD TkCon window type:

set everyone [atomselect top all]
measure minmax $everyone
This analyzes all atoms in the system and gives you the minimum and
maximum values of x, y and z coordinates of the entire protein-water
system.

5 These values are defined relative to the origin of the coordinate system,
set by the initial pdb file, 1UBQ.pdb. The center of the water box may be
determined by calculating the midpoint of each of the three sides in the
coordinate system. For example, if minimum and maximum values of x
are returned as 10.44 and 51.12, respectively, x-coordinate of the center
of the box would be 51.12+10.44
2 = 30.78. Alternatively, one can use tcl
scripting in VMD to find the center of the box by typing
measure center $everyone
in the VMD TkCon window. Determine the coordinates of the center of
your water box and record these numbers.

6 Close VMD by choosing File → Quit in the VMD Main window.

Figure 3: Ubiquitin in a water box and in a water sphere. Hydrogen atoms are
colored black for contrast.

Now that you have the pdb and psf files for solvated ubiquitin, along with the
parameter file for proteins in general, you only need a NAMD configuration file
1 BASICS OF NAMD 16

to be able to minimize and equilibrate your ubiquitin-water system. In the next

two sections, you will edit a NAMD configuration file provided to you in order
to gain a beginner’s understanding of its function.

Minimization and Equilibration. Minimization and equilibration dif-

fer from each other by the nature in which they implement the
molecular dynamics force field. Energy minimization involves search-
ing the energy landscape of the molecule for a local minimum, i.e.,
place in which the molecule is relaxed, by systematically varying the
positions of atoms and calculating the energy. Equilibration involves
molecular dynamics, whereby Newton Second Law is solved for each
atom in the system to dictate its trajectory. Achievement of equilib-
rium is judged by how well velocities, pressure, etc. are distributed
in the system over a given amount of time.

7 Before beginning the next section, you must copy the necessary pdb and
psf files into your common folder. We will store the files there so we have
a single directory from which to access them, and so that we do not need
to keep multiple copies of them. In a Terminal window, type cp ubq.pdb
ubq.psf ubq_w* ../common/ and hit Enter. Then, type ls ../common/
to see which files you have copied.

1.4 Ubiquitin in a Water Sphere: Simulation with Non-

Periodic Boundary Conditions
In this section, you will examine the minimization and equilibration of ubiquitin
in a water sphere placed in vacuum.

1.4.1 Configuration File

Configuration Warning. Be very careful when typing commands

into your configuration files. Even the slightest typographical error
will cause the simulation not to run. The point is not simply to
run a simulation, but to run it correctly. Be sure your file is correct
before you submit a simulation to be run!

Configuration File Explanation. This section will introduce a con-

figuration file to you. The file is complete and ready for submission
as is. Since you will be examining the file line by line in the section,
you may skip to section 1.4.2 if you are very short on time, and
examine the contents of this section later.

1 Go to your 1-2-sphere directory by typing cd ../1-2-sphere . Here,

you will find a configuration file for the minimization and equilibration of
ubiquitin in a water sphere.
1 BASICS OF NAMD 17

All output files for the minimization and equilibration of your ubiquitin in a
water sphere system will be placed in this directory. The configuration file is the
only input file placed here, since it is particular to this simulation. The pdb, psf,
and parameter files, which may be used by many other simulations, are placed
in your common directory and are called by each respective configuration file.

2 Open the configuration file, ubq_ws_eq.conf by typing nedit ubq_ws_eq.conf.

The configuration file may seem complex at first, but it will be examined line
by line to determine its function in your simulation.
Note that when “#" appears at the beginning of a line, the entire line is treated
as a comment and ignored by NAMD. In the middle of a line, “;#" is used to
comment out the remainder of the line.

3 The “Job Description" section contains only comments, and its purpose
is to inform those who view the configuration file about what it is meant
for. Your comment should read
# Minimization and Equilibration of
# Ubiquitin in a Water Sphere

4 The “Adjustable Parameters" section contains five commands:

• structure: calls the psf file describing the system (../common/ubq_ws.psf

).
• coordinates: calls the initial coordinate data from the file listed
next to it (in this case ../common/ubq_ws.pdb ).
• set temperature: creates a variable called “temperature" in which
to store a value for the initial temperature of the system. If you place
the text “$temperature" in the configuration file, NAMD simply
reads it as a label for the number “310". (Creating variables is useful
since you can alter the value of that variable in a single place if it is
listed many times in the configuration file.)
• set outputname: creates a variable called “outputname" in which to
store a generic name for output files. If you place the text “$outputname"
in the configuration file, NAMD simply reads it as a label for “ubq_ws_eq".
• firsttimestep: simply sets a number value for the first time step
of the simulation. It is typically useful when restarting a simulation.
For instance, if a previous simulation ended at time step 552, the
command firsttimestep 553 would be used.

5 The “Simulation Parameters" section contains many commands, com-

mented into different categories:

• Input
1 BASICS OF NAMD 18

– paraTypeCharmm: indicates whether or not the parameter file is

in the format used by the CHARMM force field. on indicates
that it is; off indicates that it is not. (If this command is not
specified, NAMD assumes the file is in X-PLOR format by de-
fault.)
– parameters: calls the force field parameters from the file listed
next to it (in this case ../common/par_all27_prot_lipid.inp
).
– temperature: sets the initial temperature of the system in Kelvin
(K), with the value listed next to it (in this case $temperature,
or 310). This is done by assigning random velocities to atoms
picked from a Maxwell distribution such that their average ki-
netic energy accurately represents the given temperature.
• Force-Field Parameters
– exclude: specifies which atomic interactions are to be excluded
from consideration. See Figure 4 for general atom labels. The
scaled1-4 value indicates that interactions between any such
atoms 1 and 2 and 1 and 3 are neglected and interactions between
atoms 1 and 4 are weakened. The van der Waals interaction for
“1-4" atoms are modified using special 1-4 parameters defined in
the parameter files, and electrostatic interaction is modified as
shown in the next command.

Figure 4: Number labels for atoms which are a given amount of bonds away
from one another

– 1-4scaling: specifies the degree to which the electrostatic in-

teraction between 1-4 atoms is to be taken into account. It may
be a decimal between 0 and 1 (here, it is 1) and indicates how
much the interaction is “turned off" or “on", respectively.
– cutoff: indicates the distance in Å beyond which electrostatic
and van der Waals interactions are cut-off. Otherwise, those in-
teractions are considered over the entire volume of your system.
This can be computationally too costly. When also employing a
1 BASICS OF NAMD 19

fast solver for electrostatics such as Particle Mesh Ewald (Sec-

tion 1.5.1) or Multilevel Summation Method (Section 1.4.1), the
cutoff parameter instead defines the splitting distance for the
1/r interaction potential, in which a short-ranged part is evalu-
ated exactly between atoms within the cutoff distance, and the
remaining long-ranged part is approximated by the solver.
– switching: indicates whether or not switching functions are
used to smoothly take electrostatic and van der Waals interac-
tions to zero at the cutoff distance. You may list “on" or “off"
next to it as a yes/no answer.
– switchdist: indicates the distance in Å at which the functional
form of electrostatic and van der Waals interactions is modified
to allow their values to approach zero at the cutoff distance. A
visual explanation of this is useful and may be found in Figure
5.
– pairlistdist: is designed to make computation faster. It spec-
ifies a distance in Å. NAMD will only search within this distance
for atoms which may interact by electrostatic or van der Waals
interactions. This way, NAMD does not have to search the entire
system. The distance must be greater than the cutoff distance,
and the list must be updated during the simulation. See Figure
5.

Figure 5: Cutoff and switching distances indicated on the left. Pair list distance
indicated on the right.

• Integrator Parameters
– timestep: indicates the value of the time step size used in the
simulation. MD simulations solve Newton’s laws in a discrete
approximation to determine the trajectories of atoms. The time
step tells NAMD how to discretize the particle dynamics. It is
specified in femtoseconds (here, 2 fs).
– rigidBonds: specifies which bonds involving hydrogen are con-
sidered to be rigid (non-vibrating). The value all specifies all
1 BASICS OF NAMD 20

linear bonds involving hydrogen and any other atoms.

– nonbondedFreq: specifies in number of time steps how often
nonbonded interactions should be calculated. It is useful for
saving computational time.
– fullElectFrequency: specifies in number of time steps how of-
ten full electrostatic interactions should be calculated.
– stepspercycle: Atoms are reassigned pair list identities (as ex-
plained above) once every cycle. This command specifies how
long one cycle lasts, i.e. the number of time steps in one cycle.
• Constant Temperature Control
– langevin: indicates whether or not the simulation uses Langevin
dynamics; uses values on and off. See the science box below for
more on Langevin dynamics.
– langevinDamping: sets the value of the Langevin coupling co-
efficient, which quantifies the friction applied to the system, re-
moving energy from the system, slowing atoms down, etc. It is
specified in ps−1 .
– langevinTemp: Langevin dynamics may be applied to all atoms
or only non-hydrogen atoms in the system. This command spec-
ifies the temperature at which to keep those atoms, even though
friction and random forces will be acting on them. (Remember
$temperature is a variable for the value 310.)
– langevinHydrogen: indicates whether or not Langevin dynamics
will be applied to hydrogen atoms in the simulation; uses values
on and off.
• Electrostatics with MSM
Multilevel Summation Method (MSM) is a useful method for dealing
with electrostatic interactions in a system that does not have fully pe-
riodic boundary conditions along all three dimensions, in other words,
when employing non-periodic or semi-periodic boundaries. Although
MSM also supports fully periodic boundary conditions, the Particle
Mesh Ewald (PME) method discussed in Section 1.5 is more efficient
for smaller periodic systems, such as ubiquitin simulated in a water
box. In the present simulation involving a water sphere, MSM is the
best alternative for electrostatics.
MSM uses particle meshes, similar to PME, arranged as 3-D grids for
evaluating long-range electrostatic interactions. Charges are assigned
to nearby grid points, and the interaction of these gridded charges
result in gridded potentials that are interpolated back to the atoms
to determine the long-range atomic forces. Unlike PME which uses
a single grid resolution, MSM uses a nesting of grids having coarser
resolution at each successive “level” in real space for resolving the
long-ranged tail of the 1/r Coulomb potential. Rather than evalu-
ating the Ewald sum as PME does, MSM uses a continued splitting
1 BASICS OF NAMD 21

of the interaction potential and performs interpolation from these

nested grid levels.
– MSM: indicates whether or not the simulation uses the Multilevel
Summation Method; uses values yes and no.
– MSMGridSpacing: The grid spacing between the finest level grid
points determines, in part, the accuracy and efficiency of MSM.
According to an error versus cost analysis, the optimal value is for
the grid spacing to be a bit larger than the average interatomic
spacing; since the performance is extremely sensitive to the grid
spacing, it is best to leave this parameter set to its default value
of 2.5Å.
– MSMxmin, MSMxmax, MSMymin, MSMymax, MSMzmin, MSMzmax: For
non-periodic simulation, MSM needs to know the expected max-
imum and minimum coordinates so as to establish grids that will
contain all atoms throughout the entire simulation. To keep the
atoms from leaving the simulated space, it might be necessary
to use a restraining potential, such as the spherical boundary
restraints used for the water sphere simulation.
Additional MSM parameters that further control accuracy and per-
formance are documented in the NAMD User’s Guide.
• Output
– outputName: Several types of output data may be written by
NAMD for any simulation. This command specifies the prefix
for output filenames. NAMD always returns the following two
output files for every simulation: a pdb file containing the final
coordinates of all atoms in the system and a pdb file contain-
ing the final velocities of all atoms in the system; the extensions
for these files are .coor and .vel, respectively. Thus, this con-
figuration file will write two files named ubq_ws_eq.coor and
ubq_ws_eq.vel
– restartfreq: During the simulation, NAMD can also create
restart files, one of which is a pdb file which stores atomic co-
ordinates, and the other of which stores atomic velocities. This
command specifies the amount of time steps between writing to
the restart file (here, every 500 steps, or 1000fs, or 1 ps). If this
command is not set, NAMD will not create a restart file. Fur-
thermore, NAMD will store the file from the previous cycle each
time it writes a new file. The filename is appended with a .old
extension; it is created in case NAMD fails in writing the new
restart file.
– dcdfreq: The dcd file contains only atomic coordinates, and
they are written to the file several times over the course of a
simulation. Thus, it provides a trajectory of the system over
the runtime. This command specifies the number of time steps
1 BASICS OF NAMD 22

between writing new coordinates to the dcd output file. If this

command is not set, NAMD will not create a dcd file.

In addition to the output files described, NAMD also prints a

log of the simulation (which we will redirect into a .log file).
This output is explained further below.
– outputEnergies: specifies the number of time steps between
each output of system energies (for various force field interac-
tions) into the .log file (here, every 100 steps, or 200 fs).

Langevin Dynamics. Langevin dynamics is a means of controlling

the kinetic energy of the system, and thus, controlling the system
temperature and/or pressure. The method uses the Langevin equa-
tion for a single particle:
2
mi d xi (t)
dt2
= Fi {xi (t)} − γi dxdt
i (t)
mi + Ri (t)

Here, two additional terms on the right hand side accompany the
ordinary force the particle experiences. The second term represents
a frictional damping that is applied to the particle with frictional
coefficient γi mi . The third term represents random forces which
act on the particle (as a result of solvent interaction). These two
terms are used to maintain particle kinetic energy to keep system
temperature, for instance, at a constant value. Because an unnec-
essarily high damping constant can significantly slow the system’s
dynamics, one should always find the minimum langevinDamping
coefficient sufficient to maintain the temperature. A value of 1.0 is
often a good starting point.

Rigid Bonds. The time step used in any MD simulation should

be dictated by the fastest process (i.e. movement of atoms) taking
place in the system. Among the various interactions, bond stretch-
ing and angle bending are the fastest, with typical bond stretching
vibrations occurring on the order of once every 10-100 femtosec-
onds. Using a time step of 2 fs, which is close to the vibrational
period (10 fs) of linear bonds involving hydrogen (fastest vibrations,
since hydrogen has small mass), requires that these bonds be fixed,
and only slower vibrations may be free to move, such as dihedral
angle bending. For large molecules, these slower vibrations typically
govern the behavior of the molecule more than the quicker ones,
so bond fixing is somewhat acceptable, but should be avoided for
accurate simulations. One prefers to use an MD timestep which is
∼ 1/10 of the fastest interactions in the simulation. For simulations
with time step of 1 fs, one should use rigidBonds for water because
water molecules have been parametrized as rigid molecules.
1 BASICS OF NAMD 23

Electrostatics. Long-range electrostatic interactions play a crucial

role in many biomolecular processes, making it important to accu-
rately represent them in simulations. However, direct calculation of
the full electrostatic interaction potential between all pairs of atoms
is not feasible when simulating protein or larger systems. In order to
efficiently calculate long-range electrostatics, NAMD provides the
choice between two different fast electrostatic solvers: the Parti-
cle Mesh Ewald (PME) method requires using periodic boundary
conditions and is preferable for constant pressure simulation; the
Multilevel Summation Method (MSM) supports either periodic or
non-periodic boundaries along any dimension and is preferable for
simulations using non-periodic or semi-periodic boundaries.

6 The “Extra Parameters" section contains commands which are applicable

to more specific simulations. Included here are commands which char-
acterize the spherical boundary conditions on the water sphere. These
conditions prevent the sphere from undergoing evaporation or diffusion.
• Spherical Boundary Conditions
– sphericalBC: indicates whether or not the simulation uses spher-
ical boundary conditions; uses values on and off.
– sphericalBCcenter: sets the x, y, z coordinates of the center of
the sphere to which spherical boundary conditions are applied.
This is why you need to record the center of your water sphere
after you solvate!
The next three commands are parameters which must be specified
for spherical boundary conditions to work properly.
– sphericalBCr1: sets the distance in Å at which the first bound-
ary potential begins to act.
– sphericalBCk1: The first potential applied by spherical bound-
ary conditions to keep the sphere together (or pull it apart) is a
harmonic one. This command sets the value of the force constant
for that potential. Its value is specified in kcal/mol·Å2 .
– sphericalBCexp1: sets the value of the exponent used to formu-
late the potential; must be a positive, even integer.

7 The “Execution Script" section contains three commands, the first two of
which apply to minimization and the last one of which applies to equili-
bration.
• Minimization
– minimize: sets the number of iterations over which to vary atom
positions to search for a local minimum in the potential (in this
case 100).
– reinitvels: Minimization is performed on the system after all
atomic velocities have been set to zero. This command resets
1 BASICS OF NAMD 24

atomic velocities such that the system starts at the temperature

specified (in this case, $temperature, or 310K).
• run: sets the number of time steps over which to run the MD equi-
libration (in this case 2500, which corresponds to 5,000 fs or 5 ps,
since a 2 fs time step has been used).

8 Now, close the configuration file by clicking File → Exit.

NAMD User’s Guide. The explanations of configuration file

commands in the previous sections is far from complete. For
an explanation of a more general configuration file, see Ap-
pendix E. For a more in-depth explanation of configuration
file commands and for acceptable values which may be sup-
plied with those commands, see the NAMD User’s Guide at
http://www.ks.uiuc.edu/Research/namd/current/ug/.

Required Configuration File Parameters. The following pa-

rameters are required by NAMD to appear in every configu-
ration file: coordinates, structure, parameters, exclude,
outputname, numsteps and exactly one of the following:
temperature, velocities, or binvelocities.

1.4.2 Run your Simulation

Running. Run your simulation by typing in a Terminal Window:

namd2 ubq_ws_eq.conf > ubq_ws_eq.log &

Setting the Path. In order to run namd2, you can provide the full
path to the binary executable (e.g., /home/me/installed/namd2)
or you can add the directory to your PATH environment variable (e.g.,
for bash you would export PATH=/home/me/installed:$PATH).
An even better approach might be to create a namd2 alias that
specifies the number of cores you plan to use for your simula-
tions. For example, setting up a namd2 alias (again using bash)
to run NAMD on four cores of your workstation, you would spec-
ify alias namd2="/home/me/installed/namd2 +p4". The most
common Linux distributions as well as Mac OS X default to using
bash (Bourne again shell) for command line access.

Your simulation should take about 20 minutes to complete, and NAMD will
produce the output files described in Section 1.7.
1 BASICS OF NAMD 25

1.5 Ubiquitin in a Water Box: Simulation with Periodic

Boundary Conditions
In this section, you will examine the minimization and equilibration of ubiquitin
in a water box with periodic boundary conditions.

Periodic Boundary Conditions. The use of periodic boundary con-

ditions involves surrounding the system under 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 environ-
ment than a water sphere surrounded by vacuum provides.

1.5.1 Configuration File

1 Go to your 1-3-box directory by typing cd ../1-3-box . Here, you will
find a configuration file for the minimization and equilibration of ubiquitin
in a water box. All output files for the minimization and equilibration of
your ubiquitin in a water box system will be placed in this directory.

2 Open the configuration file, ubq_wb_eq.conf by typing nedit ubq_wb_eq.conf.

The configuration file contains some commands which are different than the
water sphere configuration file. Here, these differences are pointed out and
explained.

3 The only differences in this file lie in the “Simulation Parameters" section,
where three new categories of parameters have been added. The “Output"
section has also been modified. The new commands are listed:
• Periodic Boundary Conditions
– Three periodic cell basis vectors are to be specified to give the
periodic cell its shape and size. They are cellBasisVector1,
cellBasisVector2, and cellBasisVector3. In this file, each
vector is perpendicular to the other two, as indicated by a single
x, y, or z value being specified by each. For instance, cellBasisVector1
is x = 42Å, y = 0Å, z = 0Å. With each vector perpendicular, a
rectangular 3-D box is formed.
– cellOrigin: specifies the coordinates of the center of the peri-
odic cell in Å. This is why you need to calculate the center of
your water box after you solvate!
– wrapWater: This command may be used with periodic boundary
conditions. If a water molecule crosses the periodic boundary,
leaving the cell, setting this command to on will translate its
coordinates to the mirror point on the opposite side of the cell.
Nothing can escape. The command may be set to on or off.
1 BASICS OF NAMD 26

– wrapAll: same as wrapWater, except this applies to all molecules.

• Electrostatics with PME
Particle Mesh Ewald (PME) is a useful method for dealing with elec-
trostatic interactions in a system when periodic boundary conditions
are present. The Ewald sum is an efficient way of calculating long
range forces in a periodic system. The particle mesh is a 3-D grid
created in the system over which the system charge is distributed.
From this charge, potentials and forces on atoms in the system are
determined. As a result, your grid size should be chosen such that
it is fine enough to accurately represent the configuration of your
system.
– PME: indicates whether or not the simulation uses the Particle
Mesh Ewald Sum method; uses values yes and no.
– PMEGridSpacing: sets the minimum ratio between the number
of PME grid points along each cellBasisVector and the phys-
ical dimensions. Since the grid will replicate the charge distri-
bution in your system, PMEGridSpacing should be chosen to be
large enough so that the grid spacing accurately reproduces your
charge distribution. However, it should not be so large that it
will slow down your simulation to provide unnecessary precision.
Typically, a grid density of slightly more than 1/Å is a good
choice to reproduce charge distribution in biological systems,
where the closest atoms have a bond separation on the order
of 1 Å. This corresponds to a PMEGridSpacing of 1.0. NAMD
will then automatically set the PME grid sizes (see below) such
that there is always less than 1.0 Å between grid points and the
sizes contain only small prime factors (e.g. 2, 3, and 5).
Alternatively, one can define the PME grid sizes manually, using
PMEGridSizeX, PMEGridSizeY, and PMEGridSizeZ. These set the size
of the PME grid along cellBasisVector1, 2 and 3, respectively
(not the x, y, and z directions as implied). For speed in computing
Fast Fourier Transforms, PMEGridSizeX should be chosen so that it
can be factorized by 2, 3, or 5. If your cellBasisVector1 = (60,
0, 0), a good choice for PMEGridSizeX might be 64, since 60 Å /
64 = 0.9375 Å and 64 = 26 . Note that since cellBasisVector is
defined with slightly different values in each direction, the size of the
mesh spacing (in length) will be different in each direction.
Note also that when using the PME method, the command cutoff
dictates the separation between long and short range forces for the
method; it does not simply turn off interactions.
• Constant Pressure Control (variable volume)
– useGroupPressure: NAMD calculates system pressure based on
the forces between atoms and their kinetic energies. This com-
mand specifies whether interactions involving hydrogen should
1 BASICS OF NAMD 27

be counted for all hydrogen atoms or simply between groups of

hydrogen atoms; uses values yes and no and must be set to yes
if rigidBonds are set.
– useFlexibleCell: specifies whether or not you want to allow
the three dimensions of the periodic cell to vary independently;
uses values yes or no.
– useConstantArea: NAMD allows you to keep the x − y cross
sectional area constant while varying the z dimension; uses values
yes and no.
– langevinPiston: indicates whether or not the simulation uses
a Langevin piston to control the system pressure; uses values on
and off.
– langevinPistonTarget: specifies, in units of bar, the pressure
which the Langevin piston tries to maintain. (1 atm = 1.013
bar)
The following are specifications for the Langevin piston which
NAMD allows you to specify.
– langevinPistonPeriod: sets the oscillation time constant in fs
for the Langevin piston.
– langevinPistonDecay: sets the damping time constant in fs for
the Langevin piston.
– langevinPistonTemp: sets the “noise" temperature in K for the
Langevin piston; should be set equal to the target temperature
for the temperature control method (here, set by langevinTemp).
• Output
– xstFreq: The extended system trajectory file contains a record of
the periodic cell parameters, essentially recording the trajectory
of the cell boundaries over the run time. This command specifies
how often, in time steps, the configuration will be recorded. If
this command is set, three xst files will be output: 1 final and 2
restarts.
– outputPressure: specifies the number of time steps between
each output of system pressure into the .log file.

Note that the commands which specify spherical boundary conditions have
been completely removed, since this simulation is using periodic boundary con-
ditions.
1 BASICS OF NAMD 28

Minimization and Equilibration. Typically, MD minimization

and equilibration simulations involve more than one minimization-
equilibration cycle, often fixing and releasing molecules in the sys-
tem. For instance, one typically minimizes the system and then
equilibrates with the atoms in the protein fixed in space, and then
minimizes the system again and equilibrates again, this time with
the protein free to move. Fixing the protein allows the water, which
typically responds much faster to forces than the protein, to do the
relaxing in the first step. This saves computational effort and pre-
vents the introduction of artifacts from an unstable starting struc-
ture.

1.5.2 Run your Simulation

Running. Run your simulation by typing in a Terminal window:

namd2 ubq_wb_eq.conf > ubq_wb_eq.log &

A Note on Restarting Simulations. Often, it is necessary or use-

ful to run your simulation in steps. For instance, you may fully
equilibrate a protein system and then desire to perform several
different types of simulations using this equilibrated structure as
a starting point. In that case, you wish to restart simulating
from the equilibrated structure. You can use the configuration
file namd-tutorial-files/common/sample.conf as a template for
restarting simulations if you note the following points:
• Change the first instance of “if {0} {" to “if {1} {" so
that NAMD recognizes that it should take coordinates, veloc-
ities, and system dimensions from restart files. By bracketing
a set of commands inside an if {n} statement, NAMD will
ignore the commands if n=0 and read them if n6=0.
• Make sure you do not specify a temperature if you are us-
ing a velocity restart file. You should comment out the
temperature command by placing a # in front of it.
• For periodic boundary conditions, make sure you do not re-
specify the periodic cell if you are using restart files, since
NAMD will read the dimensions from the restart file.
• Be certain all other simulation conditions accurately match
what you would like to accomplish. For example, you may
no longer need minimization if your system is already equili-
brated.
1 BASICS OF NAMD 29

1.6 Ubiquitin in Generalized Born Implicit Solvent

In this section, you will examine the minimization and equilibration of ubiquitin
in generalized Born implicit solvent (GBIS).

Implicit Solvent. An implicit solvent model is a simulation tech-

nique which eliminates the need for explicit water atoms by including
many of the efects of solvent in the inter-atomic force calculation.
For example, polar solvent acts as a dielectric and screens (lessens)
electrostatic interactions. The elimenation of explicit water accel-
erates conformational explorations and sometimes increases simula-
tion speed at the cost of not modeling the solvent as accurately as
explicit models.

Accuracy of Implicit Solvent. Because implicit solvent models elim-

inate 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 sol-

vent 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 atom’s 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 won’t be screened much and their Born radii will
be large. Second, inter-atomic electrostatic forces are calculated
based on atom sepparation as well as the geometric mean of the
two atoms’ Born radii.

1.6.1 Configuration File

1 Go to your 1-4-gbis directory by typing cd ../1-4-gbis . Here, you will
find a configuration file for the minimization and equilibration of ubiquitin
in implicit solvent. All output files for simulation of your ubiquitin in
implicit solvent will be placed in this directory.
2 Open the configuration file, ubq_gbis_eq.conf by typing nedit ubq_gbis_eq.conf.
The configuration file contains some commands which are different than the
water sphere and periodic boundary conditions configuration files. Here, the
differences are pointed out and explained.
3 Because there is currently no way of calculating long-range interactions in
implicit solvent, such as through PME, features based on periodic bound-
ary conditions are not used with GBIS as listed below.
1 BASICS OF NAMD 30

• Periodic Boundary Conditions are not used.

• PME is not used.
• Constant Pressure Control (variable volume) is not used.

4 A few additional commands are required for implicit solvent use. The new
commands are listed:

• structure & coordinates: implicit solvent simulations use struc-

tures which do not include any explicit water or ions as these are
represented implicitly.
• gbis: indicates whether or not the simulation uses the generalized
Born implicit solvent model. Default value is no; set to yes to utilize.
• cutoff: because there is no long-range electrostatic calculation, cutoff
should be set higher for GBIS than for PME simulations; a value of
14 Å is sufficient here.
• alphaCutoff: sets the cutoff used to determine the Born radius of
each atom. It is reasonable to set it a few Ås less than cutoff;
alphaCutoff = 12 Å should be sufficient here.
• ionConcentration: sets the concentration of implicit ions. Increas-
ing the ion concentration increases the electrostatic screening; the
default value is 0.2 M.

Explicit Water or Ions. Because solvent and ions are implicitly

modeled with generalized Born, neither water molecules nor ions
should be included in the simulated structure.

1.6.2 Run your Simulation

Running. Run your simulation by typing in a Terminal window:

namd2 ubq_gbis_eq.conf > ubq_gbis_eq.log &

Your implicit solvent simulations should take roughly the same time as the
water sphere simulation. The simulation will produce output files similar to the
ubiquitin in water sphere simulation, which are discussed in Section 1.7.
1 BASICS OF NAMD 31

1.7 Output: Water Sphere Log File

1 Output of your water sphere minimization-equilibration simulation will be
returned to the 1-2-sphere directory. Go there by typing cd ../1-2-sphere
. List the files in the directory by typing ls .

Based on the descriptions in the Output part of the “Simulation Parameters"

section, your present simulation should yield these eleven output files:

• ubq_ws_eq.log
• ubq_ws_eq.coor
• ubq_ws_eq.vel
• ubq_ws_eq.xsc

• ubq_ws_eq.dcd
• ubq_ws_eq.restart.coor
• ubq_ws_eq.restart.vel

• ubq_ws_eq.restart.xsc
• ubq_ws_eq.restart.coor.old
• ubq_ws_eq.restart.vel.old
• ubq_ws_eq.restart.xsc.old

(Your original configuration file should also still be in the directory.)

Seven of the output files contain binary data so that coordinate and velocity
information may be kept to high precision. The files which are not in binary
code are:

* ubq_ws_eq.xsc, ubq_ws_eq.restart.xsc, ubq_ws_eq.restart.xsc.old,

which are extended system configuration files. They store the periodic cell
dimensions of the system and the time step at which this information was
taken. They are not useful when periodic boundary conditions are not
used. (The restart files function in the same way as described before.)

* ubq_ws_eq.log, the log file. It is here that the energy of the various force
field interactions is stored. The log file stores a lot of information and it
will be examined.

2 The first section of the log file, indicated by the Info tag, contains the
parameters used to run the simulation, and other useful information about
the system, such as the number of atoms, the type and number of bonds,
the total mass, and the total charge.
1 BASICS OF NAMD 32

3 After the first section, information about the minimization of the system
is displayed beginning with the line labeled TCL. This line states how long
the minimization lasts.

4 After that, the pressure of the system is given. For the present simulation,
no pressure output is given because NAMD does not define the pressure
for a non-periodic system. For your simulation using periodic boundary
conditions, you will see values for the pressure.

5 The next part gives a listing of energies of interaction, first defining which
are listed with ETITLE, then listing the actual values in kcal/mol every
time step next to ENERGY. (The time step (TS) value is a count of the
number of timesteps NAMD has performed and is not in units of time,
and the temperature (TEMP) value is in K.)

6 The values listed for INITIAL STEP and GRADIENT TOLERANCE are mathe-
matical parameters corresponding to the numerical search for a minimum
in the potential. GRADIENT TOLERANCE will decrease over the course of the
minimization and may be used to roughly judge the extent of minimiza-
tion. See Figure 6.

Figure 6: Typical minimization output in the log file

7 After minimization, equilibration data are presented in the log file. See
Figure 7. The first lines shown in the figure confirm that NAMD is writing
the proper information at the proper time step, which you have specified
in the configuration file. The following space is where pressures would
normally be written. Note that since they are not pressures here, this
1 BASICS OF NAMD 33

space is used to store the total time step of the simulation. Energies at
that time step are then fully reported. In this case, they are output every
100 time steps, as specified in the configuration file.

Figure 7: Typical equilibration output of the log file

8 The end of the file confirms that the final frame of the simulation is being
recorded, and provides computational data which may be used to bench-
mark your run should you wish to change parameters, run on another
computer, etc.

The log file for your simulation of ubiquitin in a water box is quite similar to
the one described above with the exception of pressure output, as described.
See Appendix F for more information on log files.

1.8 Analysis of Water Sphere Equilibration

In this section, you will analyze the extent to which your system has equili-
brated using what is known as the Root Mean Square Deviation, or RMSD. The
RMSD characterizes the amount by which a given selection of your molecule de-
viates from a defined position in space. You will use the output files from your
minimization and equilibration of ubiquitin in a water sphere to calculate RMSD
values and analyze the extent of equilibration of the simulation. RMSD values
will be calculated for all atoms of the protein backbone (without hydrogens) for
the entire protein and for the protein excluding the last five residues.

1.8.1 RMSD for Entire Protein

Your minimization-equilibration simulation generated a trajectory for the sys-
tem called ubq_ws_eq.dcd. That trajectory is key to calculating the RMSD of
1 BASICS OF NAMD 34

the protein over the course of the equilibration. The simulation was performed
in a so-called NVT ensemble in which the number of particles, N, the volume,
V, and the temperature, T, are kept constant.
VMD includes many analysis tools. One of these, the NAMD Plot plugin, may
be used to plot output from NAMD log files. This VMD plugin uses the program
namdplot, which gets the relevant data from the log files for values of energies,
temperature, etc. over time. VMD then uses an internal plotting program to
plot the data for you.

namdplot. The program namdplot is also available separately at

www.ks.uiuc.edu/Research/namd/utilities/. This stand-alone ver-
sion will retrieve the relevant data and then use the package xmgrace
for plotting.

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. Then, navigate to the
1-2-sphere directory by using the cd command in the VMD command
terminal.

2 Launch the NAMD Plot plugin by clicking Extensions → Analysis → NAMD

Plot in the VMD Main window.

3 In the NAMD Plot window, select File → Select NAMD Log File, select the
file ubq_ws_eq.log and click Open.

4 All the thermodynamic variables from the NAMD log file are available
for plotting as shown in the VMD NAMD Plot window. TS stands for
time step, BOND for bond energy, and so forth. Click the box marked
TEMP and then select File → Plot Selected Data. This will get all the
temperature data from the log file and plot it over time.

5 Try using NAMD Plot for TOTAL (total energy). You should see the vari-
able converging to an average value for an equilibrated system.

The thermodynamic variables that you checked above tell you about the ther-
modynamic state of the whole system. But you would also like to know if your
protein is conformationally stable. For this, you will calculate the RMSD of the
protein backbone. This will give you an idea of the stability of the protein. If
the RMSD is still increasing at the end of the run, it means your protein is still
searching for a lower energy state, and thus is not yet equilibrated!

6 When you are finished examining the plots, close the Multiplot window.
1 BASICS OF NAMD 35

Root Mean Square Deviation. The Root Mean Squared Devia-

tion (RMSD) is a numerical measure of the difference between two
structures. It is defined as:
rP
Nα
(~ rα i)2
rα (tj )−h~
α=1
RM SDα (tj ) = Nα

where Nα is the number of atoms whose positions are being com-

pared, Nt is the number of time steps over which atomic positions
are being compared, ~rα (tj ) is the position of atom α at time tj ,
and h~rα i is the average value of the position of atom α to which
the positions ~rα (tj ) are being compared. It is defined as:
1
PNt
h~rα i = Nt j=1
~rα (tj )

Note that the RMSD for your protein will change over time. Also,
note the script you are using takes the first time step of the simu-
lation as a reference, rather than using an average value.

You will use VMD to calculate RMSD values:

7 Load your psf file for the water sphere, ubq_ws.psf, which is in your
common directory, by clicking File → New Molecule... in the VMD Main
window. Click the Browse button in the Molecule File Browser window.
Find ubq_ws.psf in the common directory, and double click on it. Then
click Load.
8 Click the Browse button in the Molecule File Browser window. Click on the
directory 1-2-sphere, and choose ubq_ws_eq.dcd by double clicking on
it. Then click Load. You have loaded the trajectory of your equilibration.
9 The script we are going to use is called rmsd.tcl. This is the script con-
tent:

set outfile [open rmsd.dat w]

set nf [molinfo top get numframes]
set frame0 [atomselect top "protein and backbone and noh" frame 0]
set sel [atomselect top "protein and backbone and noh"]
# rmsd calculation loop
for { set i 1 } { $i <= $nf } { incr i } {
$sel frame $i
$sel move [measure fit $sel $frame0]
puts $outfile "[measure rmsd $sel $frame0]"
}
close $outfile

10 The script does the following:

• Open file rmsd.dat for writing.
set outfile [open rmsd.dat w]
1 BASICS OF NAMD 36

• Get the number of frames in the trajectory and assign this value to
the variable nf
set nf [molinfo top get numframes]
• Select the first frame of the molecule to be the one other frames will
compare to. The selection contains the atoms in the backbone of the
protein, excluding hydrogens:
set frame0 [atomselect top "protein and backbone and noh" frame 0]
• Make the same selection as before for an undetermined frame.
set sel [atomselect top "protein and backbone and noh"]
• Text after # denotes comment.
# rmsd calculation loop
• Loop over all frames in the trajectory:
for { set i 1 } { $i <= $nf } { incr i } {
• Change the selection $sel to the frame to be compared.
$sel frame $i
• Calculate the matrix that will fit both selections. Apply this matrix
to the second selection to align the molecules:
$sel move [measure fit $sel $frame0]
• Calculate the RMSD value between these two selections, and write
it to file:
puts $outfile "[measure rmsd $sel $frame0]"

You can use the script for the system to test for equilibration.

11 Open the TkCon window of VMD by clicking Extensions → Tk Console in

the VMD Main window. In the TkCon window, type source rmsd.tcl.
This will perform all the commands in the script. The script will write
a file rmsd.dat that will contain the value of the RMSD of the protein
backbone against time.

Outside of VMD, you can use a plotting program to see this data. Examples
of these are gnuplot, xmgrace, Microsoft Excel, Mathematica, and scilab.

12 Use xmgrace to plot the file rmsd.dat. In the Terminal window, type
xmgrace rmsd.dat.

The curve does not reveal very much information because of short runtime and
infrequent writing of the dcd file. Figure 8 shows a more typical RMSD plot
for an equilibrated system along with the plot you have generated. Can you see
the first RMSD curve flattening? This means the system is equilibrated.
1 BASICS OF NAMD 37

Figure 8: RMSD versus simulation time for a system equilibrating with a longer
run time (left) and your system (right)

1.8.2 RMSD for Protein without Last 5 Residues

The last five residues are relatively unstable compared to the rest of the protein.
How do you think the RMSD should differ in this case than with the entire
protein?

1 You can easily modify the rmsd.tcl to plot the RMSD for the protein
without the last five residues. Exit xmgrace by clicking File → Exit in the
program window. Create an identical file to rmsd.tcl with a new name
by typing cp rmsd.tcl rmsd_5.tcl in the Terminal window.

2 Open the new file rmsd_5.tcl using nedit by typing nedit rmsd_5.tcl
in the Terminal window.

3 Edit the file in the following way:

Line 1: open rmsd.dat w → open rmsd_5.dat w
Line 3: ...and noh" → ...and noh and not (resid 72 to 76)"
Line 4: ...and noh" → ...and noh and not (resid 72 to 76)"

4 Save the file and exit nedit by clicking File → Save and then clicking File
→ Exit.

5 Type source rmsd_5.tcl in the TkCon window of VMD. This will per-
form all the commands in the script. The script will write a file rmsd_5.dat
that will contain the value of the RMSD of the protein backbone against
time without the last five residues.

6 Use xmgrace to plot the file rmsd_5.dat in the same manner as before
with file rmsd.dat.
xmgrace rmsd_5.dat

7 When you are done analyzing the plot, close xmgrace.

1 BASICS OF NAMD 38

8 Delete the current displayed molecule in VMD using the VMD Main win-
dow by clicking on ubq_ws.psf in the window, and selecting Molecule →
Delete Molecule.
2 ANALYSIS 39

2 Analysis
This unit presents a series of examples of how to analyze data produced from
NAMD simulations. For these examples you will not run the simulations them-
selves; instead you will analyze already produced data. Nevertheless, the in-
structions to set up and run the simulations are included, so you can actually
produce the data yourself if time permits.
The steps that must be followed for the analysis are marked with an aster-
isk (*). Make sure to go through these steps. We encourage you to at least
read the other steps, so you can refer to them later when setting up future
simulations. The exercises will have detailed explanations of methods used. We
encourage, however, that you be bold and try your own way to do the analysis.

2.1 Equilibrium
This section contains exercises that will let you analyze properties of a sys-
tem in equilibrium. The exercises mostly use pre-equilibrated trajectories from
different ensembles to calculate properties of the ubiquitin system.

2.1.1 RMSD for individual residues

Objective: Find the average RMSD over time of each residue in the
protein using VMD. Display the protein with each residue colored ac-
cording to its average RMSD value.

* 1 Go to the directory 2-1-rmsd/

You will use a script within VMD that will allow you to compute the average
RMSD of each residue in your protein, and assign this value to the User field
for each residue.

The VMD User Field. VMD stores a set of data under the field
“User" for each atom in a molecule for each timestep. One can use
this field to store any numerical values one wishes. VMD’s “User"
coloring mode, which you will soon use, will color atoms according
to their User values, making the desired property stand out from
atom to atom and over the course of a simulation. This is very
useful when you want to show a property of the system that you
have computed yourself.
Properties may be stored in any field that VMD associates with
atoms, such as the Beta and Occupancy fields, which correspond to
the B and O fields of a PDB file, but these two fields do not have
the flexibility to change their values over time (and may actually
contain pertinent PDB information!) like the User field can.

In the previous unit, you had VMD open with the equilibration trajectory of
the ubiquitin periodic system loaded. You will repeat this process:
2 ANALYSIS 40

* 2 Load the structure file. First, open the Molecule File Browser from
the File → New Molecule. . . menu item. Browse for and Load the file
ubq_wb.psf in the directory common. (Click ../ to go up one directory.)

* 3 The menu at the top of the Molecule File Browser window should now show
2: ubq_wb.psf. This ensures that the next file that you load will be added
to that molecule (molecule ID 2). Now, browse for the file ubq_wb_eq.dcd
located in the directory 1-3-box/ and click on Load again. You can close
the Molecule File Browser window.

* 4 Go to the TkCon window in VMD, and navigate to the 2-1-rmsd/ direc-

tory by typing cd ../2-1-rmsd/

* 5 The script you will use is called residue_rmsd.tcl. In the TkCon win-
dow, type:
source residue_rmsd.tcl

This command itself does not actually perform any calculations. It reads the
script residue_rmsd.tcl which contains a procedure called rmsd_residue_over_time.
Calling this procedure will calculate the average RMSD for each residue you se-
lect over all the frames in a trajectory. The procedure is called as:

rmsd_residue_over_time <mol> <sel_resid>

where <mol> is the molecule in VMD that you select (if not specified, defaults
to the top molecule), and <sel_resid> is a list of the residue numbers in that
selection. The procedure employs the equations for RMSD shown in Unit 1.

* 6 For this example, you will select all the residues in the protein. The list
of residue numbers can be obtained by typing in the TkCon window:
set sel_resid [[atomselect top "protein and alpha"] get resid]
The above command will get the residue numbers of all the α-carbons in
the protein. Since there is just one and only one α-carbon per residue, it
is a good option. The command will create a list of residue numbers in
the variable $sel_resid.

* 7 Call the procedure to calculate the RMSD values of all atoms in the newly
created selection:
rmsd_residue_over_time top $sel_resid
You should see the molecule wiggle as each frame gets aligned to the initial
structure (This may happen very fast in some machines). At the end of
the calculation, you will get a list of the average RMSD per residue. This
data is also printed to the file residue_rmsd.dat.

The procedure also sets the value of the User field of all the atoms of the
residues in the selection to the computed RMSD value. You will now color the
2 ANALYSIS 41

protein according to this value. You will be able to recognize which residues are
free to move more and which ones move less during the equilibration.
In order to use these new values to learn about the mobility of the residues, you
will create a representation of the protein colored by User value.
* 8 Open the Representations window using the Graphics → Representations. . .
menu item.
* 9 In the Atom Selection window, type protein. Choose Tube as Drawing
Method, and in the Coloring Method drop-down menu, choose submenu
Trajectory, then User, and again User. Now, click on the Trajectory tab,
and in the Color Scale Data Range, type 0.40 and 1.00. Click the Set
button.
You should now see the protein colored according to average RMSD values.
The residues displayed in blue are more mobile while the ones in red move less.
This is somewhat counterintuitive, so you will change the color scale.
* 10 Choose the Graphics → Colors. . . menu item. Choose the Color Scale tab.
In the Method pull-down menu, choose BGR. Your figure should now look
similar to Fig. 9.

Figure 9: Ubiquitin colored by the average RMSD per residue. Red denotes
more mobile residues, and blue denotes residues which moved less during equi-
libration.

* 11 Plot the RMSD value per residue by typing in a Terminal window (make
sure you are in the directory 2-1-rmsd/):
xmgrace residue_rmsd.dat
2 ANALYSIS 42

Take a look at the RMSD distribution. You can see regions where a set of
residues shows less mobility. Compare with the structural feature of your pro-
tein. You will find a correlation between the location of these regions and
secondary structure features like α helices or β sheets.
* 12 When you are finished investigating the plot, close xmgrace, saving your
plot if you wish.
* 13 Delete the current displayed molecule in VMD using the VMD Main win-
dow by clicking on ubq_wb.psf in the window, and selecting Molecule →
Delete Molecule.

2.1.2 Maxwell-Boltzmann Energy Distribution

Objective: Confirm that the kinetic energy distribution of the atoms
in a system corresponds to the Maxwell distribution for a given temper-
ature.
* 1 In a Terminal window, type cd ../2-2-maxwell to change to the direc-
tory 2-2-maxwell/
For this exercise, you need a file that contains the velocities of every atom for
one instant of the simulation. You will use a file corresponding to the last frame
of an equilibration of ubiquitin. The simulation that generated this file was run
exactly as the one you performed in directory 1-3-box/, except that the time
step used here was 1 fs. You can compare the configuration file used for that
simulation to the one in the example-output/ of this directory. As discussed in
the first unit, one should use rigidBonds for water in any production simulation
since water molecules have been parametrized as rigid molecules. But for the
illustration of Maxwell-Boltzmann energy distribution, the effect of force field
can be negligible. To be simple, we are not using the option rigidBonds. Here,
we provide you with the file ubq_wb_eq_1fs.restart.vel. Alternatively, you
can use any velocity file, as long as constraints were not used in the simulation.
2 You will now load the structure file. First, open the Molecule File Browser
from the File → New Molecule. . . menu item. Browse for and Load the
file ubq_wb.psf in the directory common.
3 The menu at the top of the window should now show 3: ubq_wb.psf. This
ensures that the next file that you load will be added to that molecule
(molecule ID 3). Now, browse for the file ubq_wb_eq_1fs.restart.vel
in the directory 2-2-maxwell/. This time you need to select the type
of file. In the Determine file type: pull-down menu, choose NAMD Binary
Coordinates, and click on Load again. Close the Molecule File Browser.
The molecule you loaded looks terrible! That is because VMD is reading the
velocities as if they were coordinates. It does not matter how it looks, because
what we want is to make VMD write a useful file: one that contains the masses
and velocities for every atom.
2 ANALYSIS 43

4 In the VMD TkCon window, navigate to the 2-2-maxwell/ directory by

typing cd ../2-2-maxwell/

5 First, you will create a selection with all the atoms in the system. Do this
by typing in the TkCon window:
set all [atomselect top all]

6 Now, you will open a file energy.dat for writing:

set fil [open energy.dat w]

7 For each atom in the system, calculate the kinetic energy k = 12 mv 2 , and
write it to a file
foreach m [$all get mass] v [$all get {x y z}] {
puts $fil [expr 0.5 * $m * [vecdot $v $v] ]
}

8 Close the file:

close $fil

9 Outside VMD, take a look at your new energy.dat file. In a Terminal

window, make sure you are in the 2-2-maxwell/ directory and type:
nedit energy.dat
It should look something like the list below (although the values will not
necessarily be the same):
1.26764386127
0.189726622306
0.268275447408
...

10 Quit nedit and VMD.

* 11 If you chose not to go through the steps above, there is a VMD script
that reproduces them and will create the file energy.dat. In a Terminal
Window, type:
vmd -dispdev text -e get_energy.tcl
This will run VMD in text mode, source the script get_energy.tcl, and
produce the energy.dat file. When it is finished, type exit to exit VMD.

You now have a file of raw data that you can use to fit the obtained energy
distribution to the Maxwell-Boltzmann distribution. You can obtain the tem-
perature of the distribution from that fit. In the rest of this section, we will
show how to do this with the 2-D graphics package xmgrace. If you are more
familiar with another tool, you may use it instead.

* 12 Quit VMD.
2 ANALYSIS 44

* 13 In a Terminal window, type xmgrace.

* 14 Once you see the xmgrace window, choose the Data → Import → ASCII
menu item. In the Files box, select the file energy.dat. Click on the OK
button. You should see a black trace. This is the raw data. You can now
close the Read sets window.

* 15 To look at the distribution of points, you will make a histogram of this

data. Choose the Data → Transformations → Histograms. . . menu item.
In the Source → Set window, click on the first line, in order to make a
histogram of the data you just loaded.

* 16 Click on the Normalize option.

* 17 Fill in Start: 0, Stop at: 10, and # of bins: 100. Click Apply.

* 18 You have created a plot, but you cannot see it yet. Use the right button of
your mouse to click on the first set (in the Source → Set window). Click
on Hide. Now, go to the main Grace window and click on the button
labeled AS, that will resize the plot to fit the existing values. This is your
distribution of energies.

Now, you will fit a Maxwell-Boltzmann distribution to this plot and find out the
temperature that the distribution corresponds to. (Hopefully, the temperature
you wanted to have in your simulation!)

Maxwell-Boltzmann distribution for kinetic energy.

2 1 √ 
k

f (k ) = √  k exp −
π (kB T ) 32 kB T

* 19 Choose the Data → Transformations → Non-linear curve fitting. . . menu

item. In the Source → Set window, click on the last line, which corre-
sponds to the histogram you created.

* 20 In the Main tab, type in the Formula window:

y = (2/ sqrt(Pi * a0∧3 )) * sqrt(x) * exp (-x / a0)
This will fit the curve and get a fit for the parameter a0, which corresponds
to kB T in units of kcal/mol. (These are the energy units that NAMD uses)

* 21 In the Parameters drop-down menu, choose 1. Fill-in forms will appear

below. You can give an initial value to A0, so that the iterations will look
for a value in the vicinity of your initial guess. In the A0 window, type
0.6, which in kcal/mol corresponds to a temperature of ∼ 300K.
2 ANALYSIS 45

Figure 10: Maxwell-Boltzmann distribution for kinetic energies.

* 22 Click on the Apply button. This will open a window with the new value
for a0, as well as some statistical measures, including the Correlation
Coefficient, which is a measure of the fit (better fit as it approaches
1). You can click on the Apply button of the Non-linear fit window several
times to obtain a better fit.

* 23 The value of a0 you obtained corresponds to kB T . Obtain the temperature

kcal
T for this distribution with kB = 0.00198657 molK . Your result is hopefully
close to 310K!

The deviation from your value and the expected one is due to the poor sampling
as well as to the finite size of the protein; the latter implies that temperature
can be “measured” only with an accuracy of about ±10K even in case of perfect
sampling.

* 24 When you are finished investigating the histogram and fit, close xmgrace,
saving your work if you wish.
2 ANALYSIS 46

2.1.3 Energies

Objective: Calculate the averages of the various energies (kinetic and

the different internal energies) over the course of a simulation.

You will look at the kinetic energy and all internal energies: bonded (bonds,
angles and dihedrals) and non-bonded (electrostatic, van der Waals) over the
course of the water box simulation which you ran in the previous unit. This
section will introduce you to a tcl script called namdstats.tcl which is very
useful for calculating average energy data for a simulation or extracting energy
data over time.

* 1 In a Terminal, go to the directory 2-3-energies/

We will investigate the simulation you ran in directory ../1-3-box/. Namdstats.tcl

will search all the information in the NAMD output file and extract the energy
data for you. You should investigate it with your text editor to familiarize
yourself with it.

Tcl and namdstats. namdstats.tcl was written in tcl so as to be

callable directly from VMD. As such, it is platform independent.
However, there are more efficient ways to perform some of the tasks
in the script for Unix users. A version of namdstats for the c shell
is available at
http://www.ks.uiuc.edu/Research/namd/utilities/

* 2 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.

* 3 In the VMD TkCon window, navigate to the 2-3-energies/ using the cd

command if you are not already there.

* 4 To use the script, type

source namdstats.tcl
data_avg ../1-3-box/ubq_wb_eq.log 101 last
The first line will source the script itself, and the second will call a pro-
cedure which will calculate the average for all output variables in the log
file from the first logged timestep after 101 to the end of the simulation.
2 ANALYSIS 47

Namdstats.tcl. The tcl script namdstats.tcl contains two

procedures that you may wish to call depending on what you would
like to calculate. The procedures are called after sourcing the script.

The first procedure is data_avg. It will output the average

of all energies that NAMD calculates over any timestep range you
specify. It is called in the following way:

data_avg <logfile> [<first timestep> <last timestep>]

where <logfile> is the NAMD logfile of your run and spec-

ifying a first and last timestep together is optional. Omitting them
will run the analysis from the start of the simulation to its end, and
specifying them as “first" or “last" will have the analysis performed
from the first timestep or to the last timestep of the simulation.

The second procedure is data_time. It will output a stream

of timestep and energy values for whichever energy type you choose
and over any timestep range you specify. The output will be placed
in a file <data stream>.dat. It is called in the following way:

data_time <data stream> <logfile> [<first timestep>

<last timestep>]

where <data stream> is one of the energy values that NAMD

outputs.

* 5 The output of namdstats.tcl should appear similar to that in Table 1.

Note that the standard units used by NAMD are Å for length, kcal/mol
for energy, Kelvin for temperature, and bar for pressure.
The temperature of your system should be close to 310 K, the temperature we
specified for the water box simulation. It will not be exact because temperature
fluctuations occur due to the finite size of the system. We will investigate these
fluctuations in the next section.

2.1.4 Temperature distribution

Objective: Simulate ubiquitin in an NVE ensemble, and analyze the
temperature distribution.
* 1 In a Terminal window, go to the directory 2-4-temp/.
For this exercise, you need a simulation that is long enough to sample the quan-
tity of interest very well. Here, we need to sample the fluctuations in kinetic en-
ergy, and hence in the temperature. Instead of doing such a simulation yourself,
we recommend that you use the provided file example-output/ubq-nve.log.
However, if you choose to run your own simulation, you can find all the nec-
essary files and instructions below. We recommend that you read the whole
section even if you don’t perform the simulations.
2 ANALYSIS 48

BOND: 232.561128
ANGLE: 690.805988
DIHED: 397.649676
IMPRP: 41.250136
ELECT: -23674.863868
VDW: 1541.108344
BOUNDARY: 0.0
MISC: 0.0
KINETIC: 4579.147764
TOTAL: -16192.340832
TEMP: 313.429264
TOTAL2: -16173.821268
TOTAL3: -16174.663916
TEMPAVG: 313.140436
PRESSURE: -47.513352
GPRESSURE: -19.105588
VOLUME: 71087.006196
PRESSAVG: 3.340104
GPRESSAVG: 3.4207

Table 1: Typical output from the script namdstats.tcl.

Units are Å for length, kcal/mol for energy, Kelvin for tem-
perature, and bar for pressure.

Temperature. The temperature of an ensemble is defined as:

N
2 X1 −
T = mj →
vj 2
3N kB 2
j=1

This expression follows from the equipartition theorem of statistical

mechanics applied to the kinetic energy
N
X 1 3
h mj −
→
vj 2 i = N kB T
2 2
j=1

where N is the number of atoms in the system.

The simulation is performed in the microcanonical ensemble (NVE, i.e., con-

stant number, volume, and energy). The configuration file provided to start the
NAMD run is called ubq-nve.conf. The main features of this configuration file
are:

• The simulation takes the restart files from the sphere equilibration simula-
tion performed in Unit 1. The files are located in directory 1-2-sphere/
and are named ubq_ws, i.e., ubq_ws.restart.vel, etc.
2 ANALYSIS 49

• The initial temperature is determined by the velocity restart file as shown

in equation above. This initial temperature corresponds to T ∼ 313K, the
temperature of the water sphere simulation.
• The time step is 2 fs; with rigidBonds turned on, this time step is ac-
ceptable.
• The simulation will run for 1 ns.

2 Run the Molecular Dynamics simulation

Running. Run your simulation by typing in a Terminal window:

namd2 ubq-nve.conf > ubq-nve.log &

* 3 If you do not run the simulation yourself, you will need to obtain the data
for the temperature from the provided log file. Copy this log file from the
directory example-output/.
cp example-output/ubq-nve.log .
If you run the simulation yourself, your log file will be generated by
NAMD, and you do not need to copy it.

* 4 In the VMD TkCon window, navigate to the 2-4-temp/ directory by

typing cd ../2-4-temp/

* 5 In order to obtain the data for the temperature from the log file we will
again use the script namdstats.tcl, which we have already sourced. Do
this by typing:
data_time TEMP ubq-nve.log
Each timestep and the corresponding temperature of the system is now in
the file TEMP.dat. The script will include all timesteps and it may take
some time to finish. Note there is no minimization of the system since
this was already performed in the initial simulation.

In order to plot the data, you can use xmgrace.

* 6 Open xmgrace.

* 7 Choose the Data → Import → ASCII menu item. Select the file TEMP.dat
You should see a black trace. This is a plot of the temperature over time.

* 8 To look at the distribution of points, you will make a histogram with this
data. Choose the Data → Transformations → Histograms. . . menu item.
In the Source → Set window, click on the first line, in order to make a
histogram of the data you just loaded.
2 ANALYSIS 50

* 9 Click on the Normalize option, and fill in Start: 300, Stop at: 330, and #
of bins 100. Click Apply.

* 10 You created a plot, but you cannot see it yet. Use the right button of
your mouse to click on the first set (in the Source → Set window). Click
on Hide. Now, go to the Main window and click on the button labeled AS,
that will resize the plot to fit the existing values. This is your distribution
of temperatures.

Does this distribution look any familiar? Your distribution should look like a
Gaussian distribution.

Temperature fluctuations. The Maxwell distribution for individual

kinetic energies n in its differential form is:
1 −n
dP (n ) = 1 exp dn
(πT0 n ) 2 kB T0

One can derive:

1
hn i = T0 kB
2
3
h2n i = (T0 kB )2
4
1
h2n i − hn i2 = (T0 kB )2
2
1
mj vj2 , ac-
P
The distribution of the total kinetic energy Ek = j 2
cording to the central limit theorem, is approximately Gaussian
−(Ek −hEk i)2
3N k2 T 2
P (Ek ) = c e B 0

The distribution function for the temperature T = 2Ek /3kB fluc-

tuations ∆T = T − T0 is then
−(∆T )2
P (∆T ) = c e 2σ 2 , σ 2 = 2T 2 /3N

Note that for N → ∞ the distribution is very sharp, but for a finite
system it is always broad. For T = 300K and N = 1000 one has
σ∼8K.

* 11 You will now fit your distribution to a normal distribution. For this, you
will use the Non-linear curve fitting feature of xmgrace.

* 12 Choose the Data → Transformations → Non-linear curve fitting. . . menu

item. In the Source → Set window, click on the last line, which corre-
sponds to the histogram you created.

* 13 In the Main tab, type in the Formula window:

y= a0 * exp(-(x-a1)∧2/a2)
2 ANALYSIS 51

* 14 In the Parameters drop-down menu, choose 3. Fill-in frames will appear

below. Give A0 an initial value of 1, A1 an initial value of 315 (you can
see that the center of the Gaussian is around that value), and A2 a value
of 2. Click several times on the Apply, to get a better fit.

This will fit the curve and get values for the parameters a0, a1 and a2, where
a0 is a normalization constant, a1 the average temperature, and a2 is σ 2 .

Figure 11: Fluctuation of Temperature in a Finite Sample

The window that appeared contains the new value of the parameters, as well
as some statistical measures, including the Correlation Coefficient, which
is a measure of the fit.

* 15 Compare the average value of the temperature obtained by this method

with the one you would obtain using namdstats by typing in the VMD
TkCon window:
data_avg ubq-nve.log

* 16 Now, look at the deviation σ 2 = 2T 2 /3N which can be calculated from

your data. Note how it is dependent on the size of the system!
2 ANALYSIS 52

Take home message: You may note from this example that single proteins,
on one hand, behave like infinite thermodynamic ensembles, reproducing the
respective mean temperature. On the other hand, they show signs of their
finiteness in the form of temperature fluctuations. In general, it is interesting
to note that the velocities of the atoms and, therefore, the kinetic energy of the
protein serves as its thermometer!
* 17 When you are finished investigating the histogram and fit, close xmgrace,
saving your work if you wish.

2.1.5 Specific Heat

Objective: Find the specific heat of ubiquitin.

Specific heat is an important property of thermodynamic systems. It is the

amount of heat needed to raise the temperature of an object per degree of
temperature increase per unit mass, and a sensitive measure of the degree of
order in a system. (Near a bistable point, the specific heat becomes large.)

Energy fluctuations and specific heat. In the canonical ensemble,

the specific heat is defined as:

hE 2 i − hEi2
cV =
kB T 2

For this exercise, you need to carry out a long simulation in the NVT (canonical)
ensemble that samples sufficiently the averages hE 2 i and hEi that arise in the
definition of cV . To produce the data for the total energy of the molecule over
time, you need to work through the steps presented below. This exercise is long
and computationally intensive. You may wish to work only through the steps
marked with *, that correspond to the analysis of the data. Nevertheless, we
encourage you to read through the whole section since it will provide you with
tools for future analysis.
* 1 In a Terminal windows, change to the directory 2-5-spec-heat/.
In order to produce an NVT run, you will restart the equilibration you per-
formed in Unit 1 for the water sphere. All necessary files are present in your
directory.
2 Run the Molecular Dynamics simulation

Running. Run your simulation by typing in a Terminal window:

namd2 ubq-nvt.conf > ubq-nvt.log &

2 ANALYSIS 53

Trajectory writing frequency. This configuration file will save frames

to the dcd very often. This will result in large output files. Make
sure you revise this configuration file for later use, as normally you
won’t need this massive kind of output.

* 3 If you do not run the simulation yourself, you will need to obtain the
trajectory that has been provided. Copy the simulation output from the
directory example-output/ by typing in a Terminal window:
cp example-output/ubq-nvt* .
If you run the simulation yourself, your output will be generated by
NAMD, and you do not need to copy it.

You now have an equilibration run of ubiquitin in the NVT ensemble, and you
want to look at the fluctuation in the energy of ubiquitin. In order to calculate
the specific heat for the protein only (as opposed to the whole system), you
need the total energies for the atoms in the protein only, however the NAMD
output for the run you just performed contains the energies for all atoms in the
system.
You would like NAMD to write the energies for only a part of the system, rather
than for the entire system, as is reported in the log file. One way to do this is
to run a separate simulation with only the part of the system you are interested
in, but this may be unrealistic from a biological standpoint. Fortunately, VMD
has a plugin which will call NAMD and allow you to calculate energies for any
part of your system that you specify.

* 4 In the VMD TkCon window, navigate to the 2-5-spec-heat/ directory:

cd ../2-5-spec-heat/

* 5 If you did not run the simulation, and copied output from the
example-output directory, please skip this step. If you ran the
simulation yourself, perform this step and skip the next one.
Load the psf file and trajectory for your system.
mol new ../common/ubq_ws.psf
mol addfile ubq-nvt.dcd

* 6 Load the psf file and trajectory for the example-output system.
mol new ../common/example-output/ubq_ws.psf
mol addfile ubq-nvt.dcd

* 7 In the VMD Main window, open the NAMD Energy plugin by clicking
Extensions → Analysis → NAMD Energy.

* 8 In the NAMDEnergy window, select the following inputs:

2 ANALYSIS 54

Figure 12: NAMD Energy GUI (Graphical User Interface)

• Molecule: ubq_ws.psf
• Selection 1: protein
• Energy Type: ALL
• Output File: myenergy.dat
• Parameter Files: ../common/par_all27_prot_lipid.inp
All other parameters are chosen by default and are consistent with the
specifications of your simulation. For more information on NAMD En-
ergy’s options, click Help → Help...

* 9 Click Run NAMDEnergy to perform the analysis. Output will be written in

the vmd console window. When you are finished, close the NAMDEnergy
window. (Note that NAMD Energy may give you an error message if it
cannot find NAMD installed. You may need to point it to the location of
NAMD on your computer.) This analysis may take several minutes to be
completed.

Now you have a file called myenergy.dat which contains the energy of only
the ubiquitin protein atoms in your system. In order to calculate the specific
heat of the molecule, we must derive both the average and squared-average total
energy of the molecule from this data. We will use a tcl script to do so.

* 10 In the VMD TkCon window, source the script to calculate the average
total energy of ubiquitin for your simulation:
source average.tcl

The script will search your output file for the total energy data for the system,
calculate both the average and squared-average total energy over all timesteps,
and output the values for you. You should become familiar with the script to
fully utilize the power of Tcl scripting in analyzing NAMD output. This script
will teach you how to read input files, extract data from them, and perform
calculations on that data.
2 ANALYSIS 55

* 11 Calculate the specific heat for ubiquitin using the equation provided in the
Energy fluctuations and specific heat box at the beginning of this
kcal
section. Consider kB = 0.00198657 molK and T = 310K. What is your
result? Try to convert your results to units of specific heat used everyday,
i.e., J/(kg˚C) using the conversion factor 1 J = 1.43846 × 1020 kcal/mol.
Note that this units correspond to the specific heat per unit mass. You
calculated the fluctuation in the energy of the whole protein, so you must
take into account the mass of your protein by dividing your obtained value
by m = 1.4219×10−23 kg. Compare with some specific heats that are given
to you in Table 2.

Substance Specific Heat

J/(kg˚C)
Human Body(average) 3470
Wood 1700
Water 4180
Alcohol 2400
Ice 2100
Gold 130
Strawberries 3890

Table 2: Specific heat of sev-

eral substances. Data obtained from
http://www.yesican.yorku.ca/home/sh_table.html and
http://www.eng.auburn.edu/users/wfgale/usda_course

* 12 Delete the current displayed molecule in VMD using the VMD Main win-
dow by clicking on ubq_ws.psf in the window, and selecting Molecule →
Delete Molecule.

2.2 Non-equilibrium
In this section, you will analyze two non-equilibrium properties of proteins,
the diffusion of heat and a phenomenon known as temperature echoes. This
section requires some knowledge of the principles of statistical physics, but the
results of the simulations can be understood intuitively even without a deep
understanding of their principles.

2.2.1 Heat Diffusion

Objective: Simulate cooling of the ubiquitin molecule and determine,
from the results of your simulation, the thermal diffusivity of your sys-
tem.
2 ANALYSIS 56

For this exercise, you will use the last step of the equilibration of ubiquitin
in a water sphere, file ubq_ws_eq.restart.coor. You will use the temperature
coupling feature of NAMD to set the temperature of the molecules in the outer
layer of the sphere to 200 K, while the rest of the bubble is set to a temperature
of 300 K. In this way, you will determine the thermal diffusivity by monitoring
change of the system’s temperature and comparing it to the theoretical expres-
sion.
Heat diffusion. Time dependence of the local temperature T (~r, t)
is governed by the heat diffusion equation
∂T (~r, t)
= D∇2 T (~r, t),
∂t
subject to the initial and boundary conditions, which for a sphere of
radius R are

T (~r, 0) = hTsim i(0), for |~r| < R,

T (~r, t) = Tbath , for |~r| ≥ R,

where D is the thermal diffusivity, Tsim is the initial temperature of

the sphere, and Tbath is the temperature of the sphere’s boundary.
The average temperature of such a system depends on time as
∞  2  
Tsim − Tbath X 1 nπ
hT i(t) = Tbath + 6 exp − Dt .
π2 n2 R
n=1

* 1 Change to this problem’s directory in a Terminal window:

cd ../2-6-heat-diff/

In order to use the temperature coupling feature of NAMD, you need to create
a file which marks the atoms subject to the temperature coupling. The following
lines
tCouple on
tCoupleTemp 200
tCoupleFile ubq_shell.pdb
tCoupleCol B
in the provided configuration file ubq_cooling.conf will enable this feature,
and set the atoms coupled to the thermal bath as the ones that have a value of
1.00 in the B column of the ubq_shell.pdb, which you will create below.
* 2 In the VMD TkCon window, navigate to the 2-6-heat-diff/ directory.
cd ../2-6-heat-diff/

3 Load the system into VMD by typing the following in the VMD TkCon
window:
mol new ../common/ubq_ws.psf
mol addfile ../1-2-sphere/ubq_ws_eq.restart.coor
2 ANALYSIS 57

4 Select all atoms in the system:

set selALL [atomselect top all]

5 Find the center of the system:

set center [measure center $selALL weight mass]

6 Find x, y and z coordinates of the system’s center and place their values
in the variables xmass, ymass, and zmass:
foreach {xmass ymass zmass} $center { break }

7 Select atoms in the outer layer:

set shellSel [atomselect top "not ( sqr(x-$xmass)
+ sqr(y-$ymass) + sqr(z-$zmass) <= sqr(22) ) "]

8 Set beta parameters of the atoms in this selection to 1.00:

$shellSel set beta 1.00

9 Create the pdb file that marks the atoms in the outer layer by “1.00” in
the beta column:
$selALL writepdb ubq_shell.pdb

10 Run the Molecular Dynamics simulation

Running. Run your simulation by typing in a Terminal window:

namd2 ubq_cooling.conf > ubq_cooling.log &

* 11 If you do not run the simulation yourself, you will need to obtain the data
for the temperature from the provided log file. Copy this log file from the
directory example-output/.
cp example-output/ubq_cooling.log .
If you run the simulation yourself, your log file will be generated by
NAMD, and you do not need to copy it.

* 12 Use namdstats.tcl to find out how the system’s temperature changes

with time. In the VMD TkCon window, type:
source ../2-3-energies/namdstats.tcl
data_time TEMP ubq_cooling.log
The file TEMP.dat will be written with the temperature of the system over
time.

* 13 You can now quit VMD.

2 ANALYSIS 58

* 14 In a Terminal window, type xmgrace TEMP.dat

* 15 In the xmgrace program, choose the Data → Transformations → Geomet-

rical transform . . . menu item. In the Scale X: window, type 2.0, which
corresponds to the time step used in your simulation. Press Accept, which
will rescale the plot such that the x-axis is now in units of fs. Now, press
the AS button in the upper left hand corner of the xmgrace window to
auto-scale your plot.

* 16 Double-click on the y-axis of the plot and enter a value of 320 in the Stop
field. Click Accept.

* 17 You will now fit the simulated temperature dependence to the theoretical
expression. For this, you will use the Non-linear curve fitting feature
of xmgrace.

* 18 Choose the Data → Transformations → Non-linear curve fitting. . . menu

item. In the Source → Set window, click on the last line, which corre-
sponds to the simulated time dependence.

* 19 In the Main tab, type or copy and paste over the Formula window (make
sure it is on one line):
y = 200 + 66.87*(exp(-0.0146*a0*x) +0.25*exp(-0.25*0.0146*a0*x)
+1/9*exp(-1/9*0.0146*a0*x) +1/16*exp(-1/16*0.0146*a0*x)
+1/25*exp(-1/25*0.0146*a0*x) +1/36*exp(-1/36*0.0146*a0*x)
+1/49*exp(-1/49*0.0146*a0*x) +1/64*exp(-1/64*0.0146*a0*x)
+1/81*exp(-1/81*0.0146*a0*x) +1/100*exp(-1/100*0.0146*a0*x))

* 20 In the Parameters drop-down menu, choose 1. Windows will appear below.

Click several times on Apply, to get a better fit. This will fit the curve
and get a value for the parameter a0, which is the thermal diffusivity.

* 21 Multiply the value of the a0 parameter by 0.1 to get the thermal diffusivity
in cm2 s−1 . You should get a value of around 0.45×10−2 cm2 s−1 . How
does your result compare to the thermal diffusivity of water D = 1.4 ×
10−3 cm2 s−1 ?

* 22 Compute the thermal conductivity of the system, K, by the following

formula: K = ρ cV D, where cV is the specific heat (that was calculated in
Section 2.1.5) and D is the thermal diffusivity (assume that the density
of the system, ρ, is 1 g/mL).

* 23 When you are finished investigating the plot, close xmgrace, saving your
plot if you wish.
2 ANALYSIS 59

Figure 13: Cooling of ubiquitin in a water sphere. The red line shows a non-
linear fit of the data to the theoretical expression.

2.2.2 Temperature echoes

Objective: Demonstrate the coherent dynamics of proteins by gener-
ating temperature echoes using MD simulations.

The motions of atoms in globular proteins (e.g., ubiquitin), referred to as

internal dynamics, comprise a wide range of time scales, from high frequency
vibrations about their equilibrium positions with periods of several femtoseconds
to slow collective motions which require seconds or more, leading to deformations
of the entire protein.
The internal dynamics of these proteins on a picosecond time scale (high fre-
quency) can be described as a collection of weakly interacting harmonic oscil-
lators referred to as normal modes. Since normal modes are formed by linear
superposition of a large number of individual atomic oscillations, it is not sur-
prising that the internal dynamics of proteins on this time scale has a delocal-
ized character throughout the protein. The situation is similar to the lattice
vibrations (phonons) in a crystalline solid. Experimentally, there exist ways
to synchronize, through a suitable signal or perturbation, these normal modes,
forcing the system in a so-called (phase) coherent state, in which normal modes
oscillate in phase. The degree of coherence of the system can be probed with a
second signal which through interference with the coherent normal modes may
lead to resonances, referred to as echoes, which can be detected experimentally.
However, the coherence of atomic motions in proteins decays through non-linear
2 ANALYSIS 60

contributions to forces between atoms. This decay develops on a time scale τd

which can be probed, e.g., by means of temperature echoes, and can be described
by employing MD simulations.
In a temperature echo the coherence of the system is probed by reassigning
the same atomic velocities the system had at an earlier time and then looking
to an echo in the temperature at time τe , as a result of such reassignment. An
example is shown in Fig. 14. At time t1 = 0 the velocities of all atoms in the
system are quenched; then, at t2 = τ the atomic velocities are reassigned again
(quenched) to their values at time t1 = 0. As a result, a temperature echo, i.e.,
a sharp dip in T (t), is detected at a subsequent time τe = τ after t2 (i.e., at
time t = 2τ ).
In this section, you will employ MD simulations to generate temperature echoes
in ubiquitin by applying the velocity reassignment just described. By modeling
ubiquitin as a large collection of weakly interacting harmonic oscillators (normal
modes), you will find that:

• the dephasing time τd of the oscillators is about one picosecond;

• the temperature echo can be expressed in terms of the temperature auto-

correlation function; and
• the depth ∆T (τe ) of the echoes decay exponentially with the delay time
τe , i.e., ∆T (τe ) ∝ exp (−τe /τd ), the exponential decay being determined
by the dephasing time τd .

T / T0
1

1/2
1/4
1/8
t
0 τ 2τ

Figure 14: Temperature quench echo.

First you will generate temperature echoes in MD simulations, and then you
will analyze the echoes in the framework of the normal mode approximation.
2 ANALYSIS 61

Temperature echoes. Temperature echoes are generated by reas-

signing twice, with a given time delay τ , the same (or scaled with
the same constant factor) set of velocities to all the atoms of the
protein. As a result, in general, the time evolution of the kinetic
temperature
3N −6
2 X mi v 2
i
T (t) =
(3N − 6)kB T 2
i=1

will acquire sharp features at t = 0 and t = τ , due to the correspond-

ing energy release or absorption. The spontaneous reappearance of
a similar sharp, but less intense, feature in T (t) at time(s) t > τ is
referred to as temperature echo(es) (Fig. 14).

The phenomenon of temperature echoes has a simple explanation within the

normal modes description of a protein. The first velocity reassignment enforces
phase coherence for the oscillators. Because of the frequency dispersion of the
normal modes (as well as the deviation from the harmonic approximation) the
phase coherence of the protein will decay in time with a characteristic dephasing
time τd ∼ 1ps. The second velocity reassignment after a delay time τ is a
“probing signal” which will test the degree of coherence of the system at the
instant of time it was applied. The depth of the echo and the instant of time at
which it occurs are quantitative characteristics of the coherence of the internal
dynamics of proteins.
In order to generate temperature echoes, one needs to equilibrate ubiquitin at
the desired initial temperature T0 = 300K, e.g., by using the velocity rescaling
method. Once the system is equilibrated, the thermostat is removed and all the
following simulations are carried out in the microcanonical (NVE) ensemble.
In this section, you will consider two types of temperature echoes. In the
first, you will reassign all atomic velocities to zero, and in the second, you will
reassign the atomic velocities to a 300 K distribution (the same as the initial
temperature of the system) and the second reassignment will be exactly the
same as the first.

Temperature quench echo

The temperature quench echo is obtained by resetting to zero the velocity of
all atoms of the protein at both times t1 = 0 and t2 = τ . You should start
from the pre-equilibrated protein in vacuum at T0 = 300K; the required files
are located in the common directory.
You need to run the simulations in the microcanonical ensemble (NVE) by using
the configuration file equil.conf in the 2-7-echoes/01_equil_NVE directory.
The simulation will run for 500 time steps (fs).

* 1 Change to the 2-7-echoes/01_equil_NVE directory in a Terminal win-

dow.
2 ANALYSIS 62

Let’s take a look at some features in the equil.conf configuration file. You
can read the file using nedit.
• The simulation takes restart files, i.e., the coordinate file and the velocity
file, from an equilibration simulation at T = 300K.
# Continuing a job from the restart files
set inputname ../common/ubi_equil
binCoordinates $inputname.restart.coor
extendedSystem $inputname.xsc

• Notice that the temperature line is commented out. The initial temper-
ature is pre-defined by the initial velocities.
#temperature $temperature
binvelocities $inputname.restart.vel

• Since you are running an NVE simulation, there is no temperature cou-

pling or pressure control.
• As discussed previously, one should use rigidBonds for water in any pro-
duction simulation since water molecules have been parametrized as rigid
molecules. But for the illustration of temperature quench echo, the effect
of force field can be negligible. To be simple, the rigidBonds option is
turned off.
* 2 Close the configuration file and nedit.
3 Run the Molecular Dynamics simulation

Running. Run your simulation by typing in a Terminal window:

namd2 equil.conf > equil.log &

* 4 If you do not run the simulation yourself, you will need to obtain the data
for the temperature from the provided log file. Copy this log file from the
directory example-output/.
cp example-output/equil.log .
If you run the simulation yourself, your log file will be generated by
NAMD, and you do not need to copy it.
* 5 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 ANALYSIS 63

* 6 Once your job is done, you need to get the temperature data from the out-
put file. In the VMD TkCon window (Extensions → Tk Console), navigate
to the 2-7-echoes/01_equil_NVE directory if you are not already there.

* 7 Get the temperature data at every timestep using our namdstats.tcl

script:
source ../../2-3-energies/namdstats.tcl
data_time TEMP equil.log
This will write the data into the file TEMP.dat.

* 8 You will now calculate the temperature autocorrelation function, which

will be used later to analyze the temperature echoes.

Temperature Autocorrelation Function.

< T (t)T (0) > − < T (0) >2

CT,T (t) =
< [T (t)]2 > − < [T (0)] >2
≈ exp(−t/τ0 )

* 9 A tcl script has been provided for you to calculate the temperature au-
tocorrelation function. Source the script by typing in the VMD TkCon
window:
source auto-corr.tcl
This will put the temperature autocorrelation function in the file auto-corr.dat.

* 10 In a Terminal window, type

xmgrace auto-corr.dat
A plot of the temperature autocorrelation function will appear.

You will now fit the simulated temperature autocorrelation function to a de-
caying exponential approximation (see the Science Box). For this, you will use
the Non-linear curve fitting feature of xmgrace.

* 11 Choose the Data → Transformations → Non-linear curve fitting. . . menu

item. In the Source → Set window, click on the only data set shown,
which corresponds to the autocorrelation function.

* 12 In the Main tab, type the following into the Formula window:
y = exp(-x/a0)

* 13 In the Parameters drop-down menu, choose 1. Windows will appear below.

Click several times on Apply, to get a better fit. This will fit the curve and
get a value for the parameter a0. This parameter is the decay (temperature
autocorrelation) time, τ0 , and it will be important for a later analysis of
your simulations.
2 ANALYSIS 64

14 If you prefer, double-click on the title area (above the graph), the x-axis,
and the y-axis to enter suitable labels for your graph.

* 15 When you are finished investigating the plot, close xmgrace, saving your
plot if you wish.

Figure 15: Temperature autocorrelation function from simulation (black) and a

decaying exponential approximation (red).

* 16 Go to the next directory, 02_quencha, by typing in a Terminal window:

cd ../02_quencha

* 17 You will perform the temperature quenching experiment for the value
τ = 200f s. You are going to quench the temperature (set it to zero),
and monitor the recovery of the system as it runs for τ time steps. If you
would like to probe other values of τ , you should edit your configuration
file echo.conf using nedit, changing the value of tau:
set tau 200

18 Run the Molecular Dynamics simulation

2 ANALYSIS 65

Running. Run your simulation by typing in a Terminal window:

namd2 echo.conf > echo.log &

* 19 If you do not run the simulation yourself, you will need to obtain the data
for the temperature from the provided log file. Copy this log file from the
directory example-output/.
cp example-output/echo.log .
If you run the simulation yourself, your log file will be generated by
NAMD, and you do not need to copy it.

* 20 Once your job is done, you need to again get the temperature data from
the output file. In the VMD TkCon window, navigate to the 2-7-echoes/02_quencha
directory.

* 21 Get the temperature data at every timestep using our namdstats.tcl

script, which we have already sourced:
data_time TEMP echo.log
This will write the data into the file TEMP.dat.

Now, you will quench the system for the second time.

* 22 Change to the directory 03_quenchb in a Terminal window.

Note that the configuration file echo.conf indicates that the simulation will
run for 3τ steps: run [expr 3*$tau]

23 Run the Molecular Dynamics simulation

Running. Run your simulation by typing in a Terminal window:

namd2 echo.conf > echo.log &

* 24 If you do not run the simulation yourself, you will need to obtain the data
for the temperature from the provided log file. Copy this log file from the
directory example-output/.
cp example-output/echo.log .
If you run the simulation yourself, your log file will be generated by
NAMD, and you do not need to copy it.
2 ANALYSIS 66

* 25 Again, get the temperature data from the output file. In the VMD TkCon
window, navigate to the 2-7-echoes/03_quenchb directory and get the
temperature data at every timestep using our namdstats.tcl script:
data_time TEMP echo.log
This will write the data into the file TEMP.dat.

* 26 In a Terminal window, change the name of this output file so as not to

confuse it with the others:
mv TEMP.dat temp3.dat

* 27 Copy the previous temperature output to this directory, changing their

names as you do so:
cp ../01_equil_NVE/TEMP.dat ./temp1.dat
cp ../02_quencha/TEMP.dat ./temp2.dat

* 28 Merge the three temperature data files into temp.dat:

cat temp[1-3].dat > temp.dat

You now have the temperature data for the whole temperature quench experi-
ment. You will use xmgrace to analyze the data.

* 29 Examine the temperature profile across all three simulations by typing at

a Terminal window:
xmgrace temp.dat

Note anything unusual? You quenched the temperature twice, and yet you see
three drops in temperature! The third one is a temperature echo. See Fig. 16.
2 ANALYSIS 67

Harmonic Approximation. The temperature echo phenomenon can

be understood by the harmonic approximation of protein motions.
In this approximation, the protein dynamics are described by a col-
lection of harmonic oscillators, with the overall motion is governed
by the statistical average of them. Here we will only give the final
expression of the temperature echoes in the harmonic approxima-
tion. If we set t1 as the temporal origin, for time after t2 , i.e., t > τ ,
it can be shown that the expression of the temperature echoes from
harmonic approximation is

1 + λ21 + 2λ22 1 + λ21 − 2λ22
T (t) ≈ T0 − CT,T (t − τ )
4 4

λ1 λ2 3τ 1 − λ21
− CT,T (|t − |) − CT,T (|t − 2τ |) , (1)
2 2 8
p p
where λ1 = T1 /T0 and λ2 = T2 /T0 . CT,T is the tempera-
ture auto-correlation function obtained from simulations. For the
temperature quench echo T1 = T2 = 0, and the above equation
simplifies to
T0 1
h i
T (t) ≈ 1 − CT,T (t − τ ) − CT,T (|t − 2τ |) . (2)
4 2

Next, you will compare T (t) in the vicinity of the echo (t = 2τ ) obtained from
the MD simulation with the theoretical prediction (2) involving the temperature
autocorrelation function obtained also from the MD simulations. The degree
of agreement between these two results is a measure of the accuracy of the
harmonic approximation.

* 30 You will use a tcl script to calculate the harmonic approximation of the
temperature echo, given by Eq. 1. Source the script by typing in the VMD
TkCon window:
source harmonic.tcl
This will write the data for the harmonic approximation to the file harmonic.dat.

* 31 In xmgrace, view the data by clicking Data → Import → ASCII. . . and

choosing the file harmonic.dat. Click OK. Then, double click on the
x-axis and change the axis Stop value to 1400 for better visualization.

You need to find the depth of the echo. From the graph, you can see that the
echo occurs around 900 fs.

* 32 Open the file temp.dat by typing less temp.dat.

* 33 Scroll down to the temperature data near 900 fs, and find the minimum lo-
cal value of temperature around that time. Record the value and calculate
the depth of the echo by subtracting the value from 75:
2 ANALYSIS 68

Echo depth = 75 - Minimum Temperature

75 is the average temperature after the second quench, and it corresponds
to 1/4 of the initial temperature T0 =300K.

Figure 16: A comparison of the simulation and the expression from harmonic
approximation at the vicinity of the echo.

You have calculated the echo depth for a particular value of the delay time
τ = 200f s. You may wish to repeat the experiment using other values of τ
(100-800 fs). (You must alter the harmonic.tcl script.) If this is done and the
data is plotted and fitted to a single exponential exp(−τ /τd ), one obtains the
so-called dephasing time τd , which is an inherent property of the system.

* 34 When you are finished investigating the plot, close xmgrace, saving your
plot if you wish.

Velocity replacement echo

Here you will generate a temperature echo by replacing the velocities of the
atoms at time t2 with the ones you assign to them at time t1 . Moreover,
by choosing these velocities according to the Maxwell-Boltzmann distribution
corresponding to T0 = 300K, i.e., the temperature of the equilibrated system
2 ANALYSIS 69

(t < 0), there will be no discontinuities (jumps) in T (t) at t1 and t2 , yet you will
observe a temperature echo at a later time, i.e., a sharp dip in the temperature
at time t = τ + τe . Your goal is to determine τe and the depth ∆T of the
echo. To this end, you will need to follow a similar procedure to the one in the
previous exercise.

* 35 In a Terminal window, navigate to the directory 04_consta.

* 36 Reassign the velocities at t1 according to a Maxwell-Boltzmann distribu-

tion for a temperature T1 = 300K, saving the reassigned velocities to a
file 300.vel. In order to do this, the following is added to the end of your
configuration file echo.conf:
run 0
output 300
These settings will not really run a simulation, but will assign initial ve-
locities with the desired distribution and save them to 300.vel. Following
these, the normal run command is used to perform a simulation for tau
time steps:
run $tau

37 Run the Molecular Dynamics simulation

Running. Run your simulation by typing in a Terminal window:

namd2 echo.conf > echo.log &

* 38 If you do not run the simulation yourself, you will need to obtain the data
for the temperature from the provided log file. Copy this log file from the
directory example-output/.
cp example-output/echo.log .
If you run the simulation yourself, your log file will be generated by
NAMD, and you do not need to copy it.

* 39 In the VMD TkCon window, navigate to the 04_consta directory. Use

the script namdstats.tcl to get the temperature data as before and put
it in the file TEMP.dat. Note that at the beginning there is a duplicate
entry for the first time step. You should delete one of these using your
text editor:
500 300.5656
500 300.5656
501 301.0253
502 302.5395
...
2 ANALYSIS 70

* 40 In a Terminal window, go to the directory 05_constb.

41 For this simulation, you should use the file 300.vel you generated be-
fore as the velocity restart file. This is included in the configuration file
echo.conf as:
velocities ../04_consta/300.vel
This will reassign the velocities to the exact same distribution they had
at the beginning of the previous simulation. This simulation will run for
3τ time steps.

42 Run the Molecular Dynamics simulation

Running. Run your simulation by typing in a Terminal window:

namd2 echo.conf > echo.log &

* 43 If you do not run the simulation yourself, you will need to obtain the data
for the temperature from the provided log file. Copy this log file from the
directory example-output/.
cp example-output/echo.log .
If you run the simulation yourself, your log file will be generated by
NAMD, and you do not need to copy it.

* 44 Again, in the VMD TkCon window, navigate to the 05_constb directory

and use the script namdstats.tcl to output the temperature data to the
file TEMP.dat.

* 45 In a Terminal window, change the name of this output file and copy the
other temperature file to this directory:
mv TEMP.dat temp2.dat
cp ../04_consta/TEMP.dat ./temp1.dat

* 46 Merge the two temperature data files into temp.dat:

cat temp[1-2].dat > temp.dat

* 47 Note that there are now two temperature values at 700 fs. You should
open the file temp.dat with your text editor and delete one.

* 48 View the temperature data and observe the temperature echo using
xmgrace:
xmgrace temp.dat
2 ANALYSIS 71

* 49 Try to locate the time of the echo and find its depth by looking at the
temperature data as before:
less temp.dat
This time, the depth of the echo is measured by:
Echo depth = 300 - Minimum Temperature

* 50 Compare your result with the pre-made results stored in the directory
example_output for τ = 200 fs. How would you explain your findings?

* 51 You can now close VMD and xmgrace.

3 STEERED MOLECULAR DYNAMICS 72

3 Steered Molecular Dynamics

In this section you will learn how to use Steered Molecular Dynamics (SMD).
In particular, you will explore elastic properties and unfolding pathways while
becoming accustomed to constant velocity and constant force SMD simulations
performed by NAMD.

• The basic idea behind any SMD simulation is to apply an external force
to one or more atoms, which we refer to as SMD atoms. In addition, you
can keep another group of atoms fixed and study the behaviour of your
protein under various conditions.
• In the following simulations, you will fix one atom of ubiquitin and pull
on another one and, thereby, stretch and unfold ubiquitin. Before running
the actual simulations, you will prepare your system and create the files
needed.

SMD. Steered Molecular Dynamics allows you to explore biological

processes on time scales accessible to molecular dynamics simu-
lations. For instance, unbinding of ligands and conformational
changes in biomolecules (like those explored using Atomic Force
Microscopy, AFM) can be studied using this technique.

3.1 Removing Water Molecules

Water Molecules. In actual SMD simulations water plays a funda-
mental role in unfolding pathways for proteins. Since the scope of
this tutorial is to introduce NAMD by performing short simulations,
we will remove water molecules in the next simulations. However,
the reader should be aware that this step is done in order to save
time when doing the tutorial. Removing water should be in principle
avoided when doing simulations.

1 In a Terminal window, set your current directory as namd-tutorial-files

using the cd command. Open a new session of VMD by typing vmd.

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

icon under Applications in the Finder. Then, navigate to the
namd-tutorial-files directory by using the cd command in the VMD
command terminal.

2 Choose the File → New Molecule... menu item of VMD. In the Molecule
File Browser use the Browse... button to find the file ubq_ws.psf in the
common directory. Load it by pressing the Load button. If you do not find
the file, look for it in the common/example-output/ directory.
3 STEERED MOLECULAR DYNAMICS 73

3 In the VMD Main window your molecule should appear (ID 0).

4 In the Molecule File Browser window, use the Browse... and the Load
buttons load the file 1-2-sphere/ubq_ws_eq.restart.coor into your psf
file. Be certain that the Load files for: field says 0:ubq_ws.psf. If you
have not done that exercise, a copy of the corresponding output files is
provided in the subdirectory 1-2-sphere/example-output/. When you
are done, close the Molecule File Browser window.

Now you have the equilibrated ubiquitin in a water sphere loaded in VMD. The
next step is to eliminate the water.

5 In the VMD TkCon window, type the following commands:

set selprotein [atomselect top protein]
$selprotein writepdb common/ubq_ww_eq.pdb

You have created a pdb file that contains the equilibrated protein without
water.

6 Delete the current molecule by selecting it in the VMD Main window

and choosing the Molecule → Delete Molecule menu item and keep VMD
opened.

3.2 Constant Velocity Pulling

In this first simulation you will stretch ubiquitin through the constant velocity
pulling method.
In this type of simulation the SMD atom is attached to a dummy atom via a
virtual spring. This dummy atom is moved at constant velocity and then the
force between both is measured using:

F~ = −∇U (3)
1
U = k[vt − (~r − ~r0 ) · ~n]2 (4)
2
where:
• U→ Potential energy.

• k→ Spring constant.

• v→ Pulling velocity.

• t→ Time.

• ~r → Actual position of the SMD atom.

• ~r0 → Initial position of the SMD atom.

• ~n → Direction of pulling.
3 STEERED MOLECULAR DYNAMICS 74

t0

t1

vt1
t2

vt2

t3
r3

vt3
t4
r4
vt4

Figure 17: Constant velocity pulling in a one-dimensional case. The dummy

atom is colored red, and the SMD atom blue. As the dummy atom moves at
constant velocity the SMD atom experiences a force that depends linearly on
the distance between both atoms.

Pulling more than one atom. The constant velocity method of

pulling can also be used to pull multiple atoms. In that case, the
position of the center of mass of the SMD atoms is attached to the
dummy atom through the virtual spring.

3.2.1 Fixed and SMD Atoms

NAMD uses a column of a pdb file to determine which atoms are fixed: all
atoms with a value of 1 (or a number different of 0) in a predetermined column
will be fixed; atoms with a value of 0 in the same column will not be affected.
Here you will use the B column of the pdb file to designate fixed atoms. Now
you need to build the respective pdb file using VMD.

1 Choose the File → New Molecule... menu item and using the Browse. . .
and Load buttons load the file common/ubq.psf that you created in the
first unit for ubiquitin in vacuum. (If you did not succeed in generating
this file in unit 1, look for common/example-output/ubq.psf)

2 In the Molecule File Browser window, use the Browse... and the Load
buttons load the file common/ubq_ww_eq.pdb into your psf file. Be certain
that the Load files for: field says 1:ubq.psf. When you are done, close the
Molecule File Browser window. The equilibrated ubiquitin without water
is now available in VMD.

3 In the VMD TkCon window, write the following command:

set allatoms [atomselect top all]
This creates a selection called allatoms which contains all atoms!
3 STEERED MOLECULAR DYNAMICS 75

4 Type $allatoms set beta 0. In this way you set the B column of all
atoms equal to 0.

5 Type set fixedatom [atomselect top "resid 1 and name CA"]. This
creates a selection that contains the fixed atom, namely the Cα atom of
the first residue.

6 Type $fixedatom set beta 1. This sets the value of the B column for
the fixed atom to 1 and, therefore, NAMD will keep this atom now fixed.

Likewise, NAMD uses another column of a pdb file to set which atom is to be
pulled (SMD atom). For this purpose it uses the occupancy column of the pdb
file.

7 Type $allatoms set occupancy 0, so the occupancy column of every

atom becomes 0.

8 Type set smdatom [atomselect top "resid 76 and name CA"]. The
smdatom selection you created contains the Cα atom of the last residue.

9 Type $smdatom set occupancy 1. Now, the occupancy column of the

smdatom selection is 1, and NAMD will pull this atom.

10 Finally, type $allatoms writepdb common/ubq_ww_eq.ref. This will

write a file in pdb format that contains all the atoms with the updated B
and occupancy columns. Keep VMD opened.

11 Take a look at the pdb file you have just created. Open a new Terminal
window, and change the directory to namd-tutorial-files, by typing cd
Workshop/namd-tutorial-files.

12 Open the file by typing nedit common/ubq_ww_eq.ref. Find the line or

row that corresponds to the Cα atom as shown in Fig. 18 (a) for the
methionine (b) residue number 1 (c) or N terminus. In that line you
should be able to see how the B column was switched to 1 (e), while the
B column for all the other atoms is 0 (d).

Figure 18: Snapshot of the file ubq_ww_eq.ref showing the fixed atom row.
3 STEERED MOLECULAR DYNAMICS 76

13 The file you just created (common/ubq_ww_eq.ref) also contains the ap-
propriate information required by NAMD to identify the SMD atom.
Again, check this by finding the row or line that corresponds to the Cα
atom as shown in Fig. 19 (a) for the glycine (b) residue number 76 (c) or
C terminus. This atom, that should have occupancy 1 (e), will be pulled
in the simulation.

Figure 19: Snapshot of the file ubq_ww_eq.ref showing the SMD atom row.

14 When you are done with the comparison of your file with Figs. 18 and
19, close nedit. Note that coordinates (columns 7, 8, and 9) in your file
will not necessarily match the ones shown in the figures because initial
velocities in the equilibration were chosen randomly.

SMD atoms. Be sure to use the coordinates from the equilibrated

protein for the file common/ubq_ww_eq.ref, so that the initial posi-
tion of the dummy atom matches the initial coordinates of the SMD
atom.

Fixed atoms. It is possible to use a column different from the B

column to indicate to NAMD the appropriate fixed atoms. In order
to do this you have to specify in the configuration file which column
you want to use (X, Y, Z, O or B) with the parameter fixedAtomsCol.

Now that you have defined the fixed and SMD atom, you need to specify the
direction in which the pulling will be performed. This is determined by the
direction of the vector that links the fixed and the SMD atoms.

15 Type the following commands in the VMD TkCon window:

set smdpos [lindex [$smdatom get {x y z}] 0]
set fixedpos [lindex [$fixedatom get {x y z}] 0]
vecnorm [vecsub $smdpos $fixedpos]
This gives you three numbers that are the x, y, and z- components of the
normalized direction between the fixed and the SMD atom! Keep these
nx , ny and nz saved for the next section.
3 STEERED MOLECULAR DYNAMICS 77

nx = ny = nz =

In the provided example (look at Figs. 18 and 19), the result is:

(38.843, 39.309, 33.494) − (26.241, 24.928, 3.279) = (12.602, 14.381, 30.215)

√
12.6022 + 14.3812 + 30.2152 = 35.757

12.602 14.381 30.215

nx = 35.757 = 0.352 ny = 35.757 = 0.402 nz = 35.757 = 0.845

16 Delete the current molecule by choosing the Molecule → Delete Molecule

menu item and keep VMD opened.

3.2.2 Configuration File

Now you have one of the files required for the intended SMD simulation, namely
common/ubq_ww_eq.ref. The next step is to create the NAMD configuration
file or modify the provided template. Exercise care at this stage and check for
misspelling of names or commands.

1 Be sure that the current directory in the Terminal window is

namd-tutorial-files.

2 Copy the file common/sample.conf to the directory 3-1-pullcv and re-

name it by typing cp common/sample.conf 3-1-pullcv/ubq_ww_pcv.conf.

3 Open the configuration file using nedit by typing

nedit 3-1-pullcv/ubq_ww_pcv.conf.

4 As a job description you may write

# N- C- Termini Constant Velocity Pulling
Note that this is just a comment line.

5 In the Adjustable Parameters section you need to change:

structure mypsf.psf → structure ../common/ubq.psf
coordinates mypdb.pdb → coordinates ../common/ubq_ww_eq.pdb
outputName myoutput → outputName ubq_ww_pcv
In this way you are using the equilibrated protein without water in your
upcoming simulation. The output files of your simulation will have the
prefix ubq_ww_pcv in their names.

6 In the Input section change:

parameters par_all27_prot_lipid.inp
→ parameters ../common/par_all27_prot_lipid.inp
3 STEERED MOLECULAR DYNAMICS 78

There is no need to modify the Periodic Boundary Conditions, Force Field

Parameters, Integrator Parameters or PME sections (The latter is disabled since
you are not using periodic boundary conditions).

7 The temperature control should be disabled in order to disturb the move-

ment of the atoms as little as possible. Switch off the Constant Tempera-
ture Control by changing:
langevin on → langevin off

8 Again, there is no need to modify Constant Pressure Control and IMD

Settings since they will not be used. However, you have to enable the
Fixed Atoms Constraint by changing the following lines:
if {0} { → if {1} {
fixedAtomsFile myfixedatoms.pdb → fixedAtomsFile
../common/ubq_ww_eq.ref
NAMD will keep fixed the atoms which have a B value of 1 in the file
ubq_ww_eq.ref.

9 In the Extra Parameters section add the following lines:

SMD on
SMDFile ../common/ubq_ww_eq.ref
SMDk 7
SMDVel 0.005
With these configuration file modifications, NAMD will pull atoms which
have occupancy 1 in the file ubq_ww_eq.ref. The virtual spring between
the dummy atom and the SMD atom will have a spring constant of 7
kcal/mol/Å2 , where 1 kcal/mol = 69.479 pN Å. The pulling will be per-
formed at a constant velocity of 0.005 Å/timestep, equivalent to 2.5 Å/ps
in the present case where the time step is 2 femtoseconds.

How to choose the spring constant. The setup of the simulation

should be as close as possible to AFM experiments. The spring
constant must be high enough that the local unbinding potential is
sampled, but must not be too high or the measured force will be
dominated by noise.

Pulling Velocity. The velocity of pulling is determined basically by

the following issues: how close you want to be to experiment, what
the total pulling distance is, and how much computer power you have
available. Therefore, you should choose the smallest velocity that
allows you to perform long enough simulations to see the unfolding
pathway.

10 In the Extra Parameters section you need to write also the direction of
pulling by adding the following line:
SMDDir nx ny nz
3 STEERED MOLECULAR DYNAMICS 79

Where nx , ny and nz should be replaced by the coordinates you calculated

at the end of section 3.2.1. (In our example, the direction is nx = 0.352,
ny = 0.402, nz = 0.845.)

11 In the same section, set the frequency in time steps with which the current
SMD data values will be printed out by typing:
SMDOutputFreq 10

12 Finally, in the Execution Script section of your configuration file be sure

that the minimization is disabled and change the number of time steps of
your simulation by replacing:
run 50000 → run 20000
This is equivalent to 40 ps (and the simulation should not take more than
20 minutes).

13 Now your configuration file is ready. SAVE IT and close your text editor.

Actual Simulations. As pointed out before, parameters of these

simulations are selected so you can run a short SMD simulation.
Actual scientific simulations are run for several nanoseconds using
pulling velocities of the order of 0.1 Å/ps or smaller in systems that
include water molecules (more than 10000 atoms).

3.2.3 Running the First SMD Simulation

All the files you need to launch your simulation are now ready. You should
have a file called ubq_ww_pcv.conf in the 3-1-pullcv directory and files:
• ubq.psf
• ubq_ww_eq.pdb
• ubq_ww_eq.ref
• par_all27_prot_lipid.inp
in the common directory. In case that you have not generated these files you can
use prepared files available in the directories 3-1-pullcf/example-output and
common/example-output.

1 Check that you actually have these files in the respective directories and
change to directory 3-1-pullcv to run the simulation.

Running. Run your simulation by typing in a Teminal Window:

namd2 ubq_ww_pcv.conf > ubq_ww_pcv.log &

Be sure you are in the 3-1-pullcv directory.

3 STEERED MOLECULAR DYNAMICS 80

While your simulation is running, some files will be created (as you already
learned in Unit 1). The only difference in this case is that in the output file
(ubq_ww_pcv.log) you will find specific information about the SMD atom and
the applied force.
• The additional output line starts with the word SMD in the first column
as shown in Fig. 20 (a).
• The second column (b) is the number of timesteps.
• The third, fourth, and fifth columns (c) are the coordinates of the SMD
atom.
• The last three columns (d) are the components of the current force applied
to the SMD atom in pN.

Figure 20: Typical output of a SMD simulation

Since the simulation will take some time, the analysis of the complete output
will be explained in section 3.4. Thus, while you wait for the result you can set
up your Constant Force simulation.

Restarting an SMD simulation. The position of the dummy atom

depends on the velocity of the pulling and the current time. There-
fore, if you want to restart an SMD simulation, make sure that
you set the firsttimestep parameter as the time step of the saved
previous configuration.

3.3 Constant Force Pulling

Now, you will carry out an SMD simulation that applies a constant force. In
this case, the Cα atom of the first residue is again kept fixed, but the SMD atom
Cα of the last residue now experiences a constant force in the direction defined
by the vector that links both atoms, the fixed and the pulled one. Note that in
this case there is no dummy atom or virtual spring.
3 STEERED MOLECULAR DYNAMICS 81

3.3.1 The SMD Atom

Again, NAMD uses a column of a pdb file to determine which atoms will be
fixed and which atoms will be pulled. In addition, another three columns are
used to specify the direction of the constant force that will be applied to the
SMD atom. Your first task is to build this file.

1 Navigate back to the namd-tutorial-files/ directory by typing in a

Terminal window cd ..

2 In your opened session of VMD, choose the File → New Molecule... menu
item and using the Browse... and the Load buttons load the file common/ubq.psf
located in the common directory, and close the Molecule File Browser win-
dow.

3 Using the mouse select the molecule in the VMD Main window and then
choose the Load Data Into Molecule... menu item. Again, using the
Browse... and the Load buttons load the file ubq_ww_eq.pdb located in
the common directory. Close the Molecule File Browser.

4 As you did in the previous section, define your fixed atom by typing the
following commands in the VMD TkCon window:
set allatoms [atomselect top all]
$allatoms set beta 0
set fixedatom [atomselect top "resid 1 and name CA"]
$fixedatom set beta 1
$allatoms set occupancy 0

5 The occupancy of the SMD atom will contain the force applied to it. Type
set smdatom [atomselect top "resid 76 and name CA"]
$smdatom set occupancy 11.54

You have now set the force to 11.54 kcal/mol/Å by entering the value into
the occupancy field. It is equivalent to 800 pN.

6 The direction of the force will be specified in the coordinates of the SMD
atom. Therefore, you have to write the normal vector in the following
way:
$smdatom set x nx
$smdatom set y ny
$smdatom set z nz
where nx , ny , and nz have to be replaced by the appropriate values you
already calculated above (in our example 0.352, 0.402, and 0.845). Since
the VMD OpenGL Display interprets the numbers just entered as coordi-
nates of the SMD atom, it displays now an inaccurate protein structure.
This is OK and we apologize for the poor appearance.
3 STEERED MOLECULAR DYNAMICS 82

7 Now, save the coordinates into a file by typing $allatoms writepdb

common/ubq_ww_eq2.ref
8 Delete the current molecule by choosing the Molecule → Delete Molecule
menu item and keep VMD opened.

3.3.2 Configuration File

As you did before, in the next step you will modify the configuration file in
order to set up your constant force simulation.
1 Open a new Terminal window, and change the directory to namd-tutorial-files/.
2 Copy the file common/sample.conf to the directory 3-2-pullcf and re-
name it by typing cp common/sample.conf 3-2-pullcf/ubq_ww_pcf.conf.
3 Now, open the configuration file by typing
nedit 3-2-pullcf/ubq_ww_pcf.conf.
4 As a job description you can write
# N-C-Termini Constant Force Pulling
5 In the Adjustable Parameters section you need to change:
structure mypsf.psf → structure ../common/ubq.psf
coordinates mypdb.pdb → coordinates ../common/ubq_ww_eq.pdb
outputName myoutput → outputName ubq_ww_pcf
In this way you are using the equilibrated protein without water in your
simulation. The output files of your simulation will have the prefix ubq_ww_pcf
in their names.
6 In the Input section change:
parameters par_all27_prot_lipid.inp
→ parameters ../common/par_all27_prot_lipid.inp
7 As before, switch off the Constant Temperature Control by changing:
langevin on → langevin off
8 Enable the Fixed Atoms Constraint by changing the following lines
if {0} { → if {1} {
fixedAtomsFile myfixedatoms.pdb → fixedAtomsFile
../common/ubq_ww_eq2.ref
NAMD will keep fixed the atoms which have a B value of 1 in the file
../common/ubq_ww_eq2.ref.
9 In the Extra Parameters section add the following lines:
constantforce yes
consforcefile ../common/ubq_ww_eq2.ref
NAMD will apply a constant force to the atoms that have occupancy
different from 0. The force is calculated from the file as (x, y, z) × O,
where O is the value of the occupancy column.
3 STEERED MOLECULAR DYNAMICS 83

10 Finally, in the Execution Script section of your configuration file be sure

that the minimization is disabled and change the number of time steps
your simulation will run by replacing:
run 50000 → run 20000
This is equivalent to 40 ps.

11 Your second configuration file is done. SAVE IT and close the text editor.

3.3.3 Running the Second SMD Simulation

Now, all the files you need to launch your second simulation are ready. You
should have a file called ubq_ww_pcf.conf in the 3-2-pullcf directory and
files:

• ubq.psf
• ubq_ww_eq.pdb
• ubq_ww_eq2.ref

• par_all27_prot_lipid.inp
in the common directory. In case that you have not generated these files you
can use prepared files available at directories 3-2-pullcf/example-output and
common/example-output.

1 In a Terminal window, change to directory 3-2-pullcf/ to run the sim-

ulation.

Running. Run your simulation by typing in a Teminal Window:

namd2 ubq_ww_pcf.conf > ubq_ww_pcf.log &

Be sure you are in the 3-2-pullcf directory.

Note that the output file has no extra information in this case, since the applied
force is now constant.

3.4 Analysis of Results

3.4.1 Force Analysis for Constant Velocity Pulling
Hopefully your constant velocity SMD simulation has finished successfully
and you can proceed with analyzing the generated data, specifically the tra-
jectory and the force applied to the SMD atom. In case you have not suc-
ceeded with your simulations we provide the respective files in the subdirectories
example-output. For instance, if you do not find a file in the directory common,
you should find it in the directory common/example-output.
3 STEERED MOLECULAR DYNAMICS 84

1 In VMD choose the New Molecule... menu item. Using the Browse... and
the Load buttons, load the file ubq.psf in the common directory.

2 In the Molecule File Browser window, use the Browse... and the Load
buttons load the trajectory file 3-1-pullcv/ubq_ww_pcv.dcd. Close the
Molecule File Browser.

In the screen you should be able to see the trajectory of the molecular dynamics
simulation. By choosing an appropriate representation (e.g. New Cartoon) look
how the β strands behave along the trajectory and how one end of the protein
remains fixed while the other one is pulled as you had set it up (Fig. 21). Note
that secondary structure will be calculated for the New Cartoon representation
based on the frame of the trajectory you are viewing. If you move to another
trajectory frame and would like to view the protein’s secondary structure in that
frame, type mol ssrecalc top in the TkCon window and VMD will update it.

Figure 21: Snapshots of ubiquitin pulling with constant velocity at three di-
fferent time steps.

After inspecting your trajectory you should extract the force applied to the
SMD atom from the NAMD output file:

3 In the VMD TkCon window, set your current directory to 3-1-pullcv by

typing: cd 3-1-pullcv

4 In directory 3-1-pullcv, create the subdirectory called analysis by typ-

ing: mkdir analysis

You will now use a script designed to extract the necessary information from the
NAMD log file and calculate the force of pulling. The NAMD log file provides
SMD information on lines beginning with the tag “SMD": the current timestep,
current position of the center of mass of the restrained atoms, and the current
force applied to the center of mass (in piconewtons).
3 STEERED MOLECULAR DYNAMICS 85

5 Source the provided tcl script for getting force information from the NAMD
log file. Type the following command in the VMD TkCon window:
source ft.tcl

In order to obtain the force in the direction of pulling you need to calculate
f~ · ~n, where ~n is the normalized direction of pulling (in our example it was
0.443, 0.398, 0.803). The script will ask you for these values. Simply enter them
as n_x, n_y, and n_z. The script will create a file ft.dat (in the analysis
directory) which contains just the simulation timestep and the magnitude of the
force on the dummy atom at that time.
6 The contents of the script are shown below. You should understand what
each line does in order to make tcl scripting a part of your repertoire.

### Open the log file for reading and the output .dat file for writing
set file [open ubq_ww_pcv.log r]
set output [open analysis/ft.dat w]

### Gather input from user.

puts "Enter a value for n_x:"
set nx [gets stdin]
puts "Enter a value for n_y:"
set ny [gets stdin]
puts "Enter a value for n_z:"
set nz [gets stdin]

### Loop over all lines of the log file

set file [open ubq_ww_pcv.log r]
while { [gets $file line] != -1 } {

### Determine if a line contains SMD output. If so, write the

### timestep followed by f(dot)n to the output file
if {[lindex $line 0] == "SMD"} {
puts $output "[lindex $line 1] [expr $nx*[lindex $line 5]
+ $ny*[lindex $line 6] + $nz*[lindex $line 7]]"
}
}

### Close the log file and the output .dat file
close $file
close $output

7 Type xmgrace analysis/ft.dat. A plot of force versus time should ap-

pear on your screen (Fig. 22).
8 In VMD, look again at the trajectory at different times and try to identify
the events that are related to peaks in the force. When you are done close
xmgrace.
3 STEERED MOLECULAR DYNAMICS 86

Figure 22: Force vs Time. Your plot should resemble the graph on the left. The
graph on the right was obtained in a simulation where ubiquitin was placed in
a water sphere (∼ 26000 atoms) and pulled at 0.5 Å/ps using a time step of 1
femtosecond. The results of this last simulation are in better agreement with
actual AFM experiments. The force at the end (t > 400ps) increases when the
protein becomes completely unfolded. The first peak in the force is associated
with the breaking of hydrogen bonds between two β strands.

What can we learn from this graph. Intermediates states in the

unfolding pathway of the protein can be identified using this kind of
simulation. A particular useful way to see this is by plotting force vs
extension. In addition, microscopic details missed by experiments,
like the behaviour of hydrogen bonds and the role of water in the
unfolding pathway, can be studied and explored along with the mi-
cromechanical and elastic properties of the protein.

3.4.2 Distance Analysis for Constant Force Pulling

1 In VMD, delete the current molecule by choosing the Molecule → Delete
Molecule menu item.

2 Choose the New Molecule... menu item. Using the Browse... and the Load
buttons load the file ubq.psf located in the common directory.

3 In the Molecule File Browser window, use the Browse... and the Load but-
tons load the trajectory file 3-2-pullcf/ubq_ww_pcf.dcd. As mentioned
before, in case that you have not accomplished the simulation to generate
the file, we provide such file at 3-2-pullcf/example-output/ubq_ww_pcf.dcd.

In the OpenGL Display window you will see the trajectory of the constant force
pulling simulation.
3 STEERED MOLECULAR DYNAMICS 87

Figure 23: Distance vs Time Steps. Your plot should resemble the graph on
the left. The graph on the right was obtained in a simulation where ubiquitin
was placed in a water box and pulled from the C-Terminus while keeping the
residue K48 fixed. In this case we can observe an intermediate state.

4 Once the complete trajectory is loaded, use the slider in the VMD Main
window to go back to the first frame.
5 Choose the Graphics → Representations... menu item. In the Graphical
Representations window choose the New Cartoon drawing method.
6 Press the Create Rep button in order to create a new representation. In
the Selected Atoms text entry delete the word all and type protein and
resid 1 76 and name CA.
7 For this representation choose the VDW drawing method. You should be
able to see the two Cα atoms at the end of your protein as spheres.
8 Hide the first representation by double clicking on it in the Graphical
Representations window.
9 Click in the OpenGL Display window and press the key 2 (Shortcut to
bond labels). Your cursor should look like a cross. Pick the two spheres
displayed. A line linking both atoms should appear with the distance
between them displayed.
10 Choose the Graphics → Labels... menu item. In the Labels window, choose
the label type Bonds. Select the bond displayed, click on the Graph tab
and then click on the Graph button. This will create a plot of the distance
between these two atoms over time (Fig. 23). You also have the option
to write the data and to a file, which can then be read by an external
graphing program.
4 ACKNOWLEDGMENTS 88

Intermediate States. SMD allows you to explore properties of dif-

ferent mechanical proteins, where you can observe multiple plateaus
in plots of separation versus time for constant force simulations. In
some biological systems these intermediate states are functionally
relevant.

This is the end of the NAMD tutorial. We hope you can now make use of
VMD and NAMD in your work.

4 Acknowledgments
The developement of NAMD and the tutorials are funded by the National Insti-
tutes of Health (NIH 9P41GM104601 — Center for Macromolecular Modeling
and Bioinformatics, NIH 5 R01 GM908243-02 — “Commodity Hardware Accel-
eration of Popular Modeling Software for Structural Biology”). Proper citation
is a primary way in which we demonstrate the value of our software to the sci-
entific community and is essential to continued NIH funding for NAMD. The
authors request that all published work which utilizes NAMD include the pri-
mary NAMD citation:

Scalable molecular dynamics with NAMD. James C. Phillips, Rosemary

Braun, Wei Wang, James Gumbart, Emad Tajkhorshid, Elizabeth Villa, Christophe
Chipot, Robert D. Skeel, Laxmikant Kale, and Klaus Schulten. Journal of Com-
putational Chemistry, 26:1781-1802, 2005.
A PDB FILES 89

A PDB Files
The term PDB can refer to the Protein Data Bank (http://www.rcsb.org/pdb/),
to a data file provided there, or to any file following the PDB format. Files in
the PDB include information such as the name of the compound, the species
and tissue from which is was obtained, authorship, revision history, journal
citation, references, amino acid sequence, stoichiometry, secondary structure
locations, crystal lattice and symmetry group, and finally the ATOM and HET-
ATM records containing the coordinates of the protein and any waters, ions,
or other heterogeneous atoms in the crystal. Some PDB files include multiple
sets of coordinates for some or all atoms. Due to the limits of x-ray crystallog-
raphy and NMR structure analysis, the coordinates of hydrogen atoms are not
included in the PDB.
NAMD and VMD ignore everything in a PDB file except for the ATOM and
HETATM records, and when writing PDB files the ATOM record type is used
for all atoms in the system, including solvent and ions. Here are the ATOM
records for the first two residues of ubiquitin from the 1UBQ entry in the PDB:
ATOM 1 N MET 1 27.340 24.430 2.614 1.00 9.67 1UBQ 71
ATOM 2 CA MET 1 26.266 25.413 2.842 1.00 10.38 1UBQ 72
ATOM 3 C MET 1 26.913 26.639 3.531 1.00 9.62 1UBQ 73
ATOM 4 O MET 1 27.886 26.463 4.263 1.00 9.62 1UBQ 74
ATOM 5 CB MET 1 25.112 24.880 3.649 1.00 13.77 1UBQ 75
ATOM 6 CG MET 1 25.353 24.860 5.134 1.00 16.29 1UBQ 76
ATOM 7 SD MET 1 23.930 23.959 5.904 1.00 17.17 1UBQ 77
ATOM 8 CE MET 1 24.447 23.984 7.620 1.00 16.11 1UBQ 78
ATOM 9 N GLN 2 26.335 27.770 3.258 1.00 9.27 1UBQ 79
ATOM 10 CA GLN 2 26.850 29.021 3.898 1.00 9.07 1UBQ 80
ATOM 11 C GLN 2 26.100 29.253 5.202 1.00 8.72 1UBQ 81
ATOM 12 O GLN 2 24.865 29.024 5.330 1.00 8.22 1UBQ 82
ATOM 13 CB GLN 2 26.733 30.148 2.905 1.00 14.46 1UBQ 83
ATOM 14 CG GLN 2 26.882 31.546 3.409 1.00 17.01 1UBQ 84
ATOM 15 CD GLN 2 26.786 32.562 2.270 1.00 20.10 1UBQ 85
ATOM 16 OE1 GLN 2 27.783 33.160 1.870 1.00 21.89 1UBQ 86
ATOM 17 NE2 GLN 2 25.562 32.733 1.806 1.00 19.49 1UBQ 87

The fields seen here in order from left to right are the record type, atom
ID, atom name, residue name, residue ID, x, y, and z coordinates, occupancy,
temperature factor (called beta), segment name, and line number.
If this file is loaded into VMD and then written out as a new file, most of the
extra information will be removed, the HETATM records will become ATOM
records, and the previously empty chain ID field (between residue name and
residue ID) will be set to X (unless present in the original file), and the line
number will be omitted, as seen here:
ATOM 1 N MET X 1 27.340 24.430 2.614 1.00 9.67 1UBQ
ATOM 2 CA MET X 1 26.266 25.413 2.842 1.00 10.38 1UBQ
ATOM 3 C MET X 1 26.913 26.639 3.531 1.00 9.62 1UBQ
B PSF FILES 90

B PSF Files
A PSF file, also called a protein structure file, contains all of the molecule-
specific information needed to apply a particular force field to a molecular sys-
tem. The CHARMM force field is divided into a topology file, which is needed
to generate the PSF file, and a parameter file, which supplies specific numerical
values for the generic CHARMM potential function. The topology file defines
the atom types used in the force field; the atom names, types, bonds, and par-
tial charges of each residue type; and any patches necessary to link or otherwise
mutate these basic residues. The parameter file provides a mapping between
bonded and nonbonded interactions involving the various combinations of atom
types found in the topology file and specific spring constants and similar param-
eters for all of the bond, angle, dihedral, improper, and van der Waals terms in
the CHARMM potential function.
The PSF file contains six main sections of interest: atoms, bonds, angles,
dihedrals, impropers (dihedral force terms used to maintain planarity), and
cross-terms. The following is taken from a PSF file for ubiquitin. First is the
title and atom records:

PSF CMAP

6 !NTITLE
REMARKS original generated structure x-plor psf file
REMARKS 2 patches were applied to the molecule.
REMARKS topology top_all27_prot_lipid.inp
REMARKS segment U { first NTER; last CTER; auto angles dihedrals }
REMARKS defaultpatch NTER U:1
REMARKS defaultpatch CTER U:76

1231 !NATOM
1 U 1 MET N NH3 -0.300000 14.0070 0
2 U 1 MET HT1 HC 0.330000 1.0080 0
3 U 1 MET HT2 HC 0.330000 1.0080 0
4 U 1 MET HT3 HC 0.330000 1.0080 0
5 U 1 MET CA CT1 0.210000 12.0110 0
6 U 1 MET HA HB 0.100000 1.0080 0
7 U 1 MET CB CT2 -0.180000 12.0110 0

The fields in the atom section are atom ID, segment name, residue ID, residue
name, atom name, atom type, charge, mass, and an unused 0. PSF files may
be in either CHARMM or X-PLOR format, with the CHARMM format using
an integer rather than a name for the atom type. NAMD requires the X-PLOR
format, which is also more flexible since it is not tied to the specific order of
atom types in a single parameter file. NAMD and VMD require that the order
of atoms in any PDB, DCD, or other atomic data file exactly matches the order
found in the PSF file.
Notice that hydrogen atoms are included, that multiple atoms in a residue
may share the same type (e.g., HT1, HT2, and HT3 are of type HC), and that
B PSF FILES 91

atoms of the same element may have different types (e.g., CA and CB, HA, and
HT1). Much of the information in the PDB is also included in the PSF file.
Often, the only information from the PDB file used by NAMD is the atomic
coordinates. In some cases the contents of the occupancy and beta fields of the
PDB file are used as parameters to optional simulation methods.
The covalent bond section lists four pairs of atoms per line:
1237 !NBOND: bonds
1 5 2 1 3 1 4 1
5 6 7 5 7 8 7 9
10 7 10 11 10 12 13 10

The angle section lists three triples of atoms per line:

2257 !NTHETA: angles
1 5 6 1 5 18 2 1 5
2 1 4 2 1 3 3 1 5
3 1 4 4 1 5 5 18 19

The dihedral and improper sections list two quadruples of atoms per line:
3293 !NPHI: dihedrals
1 5 7 10 1 5 7 8
1 5 7 9 1 5 18 20
1 5 18 19 2 1 5 7

204 !NIMPHI: impropers

18 5 20 19 20 18 22 21
30 32 27 31 30 27 32 31
32 30 33 34 32 30 34 33

The cross-term sections list two quadruples of atoms per line:

74 !NCRTERM: cross-terms
18 20 22 35 20 22 35 37
35 37 39 54 37 39 54 56
54 56 58 74 56 58 74 76

The order of the atoms within each bond is not significant. The order of
atoms within angles, dihedrals, and impropers is significant, but the order can
be reversed without affecting the resulting potential. In no case is the relative
order of different bonds etc. significant. Since every bond and improper is
explicitly listed in the topology file, the atom order within these terms tend to
match the original order given. Angles and dihedrals are typically generated
automatically and therefore appear in a sorted order.
There is a difference between X-PLOR formatted CHARMM PSF files,
which VMD generates with the psfgen module and NAMD uses in conjunction
with CHARMM parameter files, and PSF files generated by X-PLOR, which
NAMD uses in conjunction with X-PLOR parameter files. Dihedral terms some-
times require multiple sinusoids to represent a torsional potential and therefore
B PSF FILES 92

multiple parameter sets appear in the parameter file. In PSF files generated
by X-PLOR multiple dihedrals would be indicated by duplicate dihedrals in
the PSF file. When using CHARMM PSF and parameter files NAMD extracts
any multiple dihedral terms directly from the parameter file and each dihedral
appears only once in the PSF file.
C TOPOLOGY FILES 93

C Topology Files
A CHARMM forcefield topology file contains all of the information needed to
convert a list of residue names into a complete PSF structure file. It also con-
tains internal coordinates that allow the automatic assignment of coordinates
to hydrogens and other atoms missing from a crystal PDB file.
The current versions of the CHARMM forcefield are CHARMM22 for pro-
teins and CHARMM27 for lipids and nucleic acids including CMAP correction to
proteins. The individual topology files are named, respectively, top_all22_prot_cmap.inp,
top_all27_lipid.rtf, and top_all27_na.rtf. To enable computation on hy-
brid systems, combinations are also provided, named top_all27_na_lipid.rtf,
top_all27_prot_lipid.rtf, and top_all27_prot_na.rtf which can all be
found in the CHARMM31 release. While the tools used with NAMD allow mul-
tiple topology and parameter files to be used simultaneously, it is preferable to
use these pre-combined files. The CHARMM31 release is available for download
from the MacKerell web site:
http://mackerell.umaryland.edu/charmm_ff.shtml
We will examine top_all27_prot_lipid.rtf; the other files are similar.
At the beginning of the file is the header, indicated by lines beginning with
*s, followed in this case by “31 1" to indicate the version of CHARMM that
generated the file:

*> CHARMM22 All-Hydrogen Topology File for Proteins and Lipids <<
*>>>>>> Includes phi, psi cross term map (CMAP) correction <<<<<<
*>>>>>>>>>>>>>>>>>>>>>> July 2004 <<<<<<<<<<<<<<<<<<<<<<<<<<
* All comments to ADM jr. via the CHARMM web site: www.charmm.org
* parameter set discussion forum
*
31 1

Comments in the topology file are indicated by ! anywhere on a line. Im-

portant usage information is often contained in these comments, so it is always
a good idea to inspect the topology file in a text editor when using it to build
a structure. The next part of the file is a long set of comments containing
references to the papers and other sources of the parameters in the file:

! references
!
!PROTEINS
!
!MacKerell, A.D., Jr,. Feig, M., Brooks, C.L., III, Extending the
!treatment of backbone energetics in protein force fields: limitations
!of gas-phase quantum mechanics in reproducing protein conformational
!distributions in molecular dynamics simulations, Journal of
!Computational Chemistry, 25: 1400-1415, 2004.
!
!MacKerell, Jr., A. D.; Bashford, D.; Bellott, M.; Dunbrack Jr., R.L.;
C TOPOLOGY FILES 94

!Evanseck, J.D.; Field, M.J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.;
!Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F.T.K.; Mattos,
!C.; Michnick, S.; Ngo, T.; Nguyen, D.T.; Prodhom, B.; Reiher, III,
!W.E.; Roux, B.; Schlenkrich, M.; Smith, J.C.; Stote, R.; Straub, J.;
!Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. All-atom
!empirical potential for molecular modeling and dynamics Studies of
!proteins. Journal of Physical Chemistry B, 1998, 102, 3586-3616.
!

The topology file must define the type, mass, and charge of every atom in
every residue, so that a PSF file can be constructed. While the partial charges
assigned to atoms of the same type vary between residues, their masses do not.
Therefore, the mass of every atom type is declared once at the beginning of
the file in a MASS statement. This statement also pairs an integer with each
type name, which is used in CHARMM formatted PSF files, but not in the
X-PLOR formatted PSF files used by NAMD. The type indices are unique but
not necessarily consecutive. Notice in the following except that there are many
types of hydrogen and carbon atoms defined, but the atomic masses are the
same:

MASS 1 H 1.00800 H ! polar H

MASS 2 HC 1.00800 H ! N-ter H
MASS 3 HA 1.00800 H ! nonpolar H
MASS 4 HT 1.00800 H ! TIPS3P WATER HYDROGEN
MASS 5 HP 1.00800 H ! aromatic H
MASS 6 HB 1.00800 H ! backbone H
MASS 7 HR1 1.00800 H ! his he1, (+) his HG,HD2
MASS 8 HR2 1.00800 H ! (+) his HE1
MASS 9 HR3 1.00800 H ! neutral his HG, HD2
MASS 10 HS 1.00800 H ! thiol hydrogen
MASS 11 HE1 1.00800 H ! for alkene; RHC=CR
MASS 12 HE2 1.00800 H ! for alkene; H2C=CR
MASS 13 HA1 1.00800 H ! alkane, CH, new LJ params (see toppar_all22_prot_aliphatic_c27.str)
MASS 14 HA2 1.00800 H ! alkane, CH2, new LJ params (see toppar_all22_prot_aliphatic_c27.str)
MASS 15 HA3 1.00800 H ! alkane, CH3, new LJ params (see toppar_all22_prot_aliphatic_c27.str)
MASS 16 HF1 1.00800 H ! Aliphatic H on fluorinated C (see toppar_all22_prot_fluoro_alkanes.str)
MASS 17 HF2 1.00800 H ! Aliphatic H on fluorinated C (see toppar_all22_prot_fluoro_alkanes.str)
MASS 20 C 12.01100 C ! carbonyl C, peptide backbone
MASS 21 CA 12.01100 C ! aromatic C
MASS 22 CT1 12.01100 C ! aliphatic sp3 C for CH
MASS 23 CT2 12.01100 C ! aliphatic sp3 C for CH2
MASS 24 CT3 12.01100 C ! aliphatic sp3 C for CH3

When specifying the connectivity of a chain of residues in a protein, it is nec-

essary to refer to atoms in the previous or succeeding residue. The CHARMM
topology file declares those atom types that will be referenced in adjoining
residues as such:

DECL -CA
C TOPOLOGY FILES 95

DECL -C
DECL -O
DECL +N
DECL +HN
DECL +CA

The first and last residues of a chain obviously have different connectivity
from those in the center, since they have one fewer neighbor. This is handled
by applying patch residues, normally referred to as patches, to the terminal
residues. As will be seen, any residue definition can specify the patch to be
applied when it is the first or last residue in a segment. However, a default
set is declared for the entire file, as in the following where the default patch is
NTER for the first residue of a segment and CTER for the last:

DEFA FIRS NTER LAST CTER

While the covalent bond connectivity between atoms must necessarily be

provided by the topology file, enumerating all of the required angles and dihe-
drals would be tedious and error-prone, as well as enormously complicated since
every combination of residues joined by a peptide bond would require a different
set. Therefore, angles and dihedrals are automatically generated for every pair
or triple of connected bonds when a segment is built. This autogeneration may
be enabled or disabled on a per-segment basis as it should never be used on
segments of water, but the default is defined in the topology file:

AUTO ANGLES DIHE

We are now ready for the actual residue definitions, beginning with alanine,
as shown below. A residue is indicated by the RESI statement with the residue
name (ALA) and total charge (0.00). Next are listed all of the atoms in the
residue in ATOM statements with the atom name (N, HN, CA), type (NH1, H,
CT1), and partial charge (-0.47, 0.31, 0.07). The GROUP statements, dividing
the atoms into integer-charge groups, are not used by NAMD and should not be
confused with the hydrogen groups, each a non-hydrogen atom and all hydrogens
bonded to it, that NAMD uses to accelerate distance-testing for nonbonded
calculations.
RESI ALA 0.00
GROUP
ATOM N NH1 -0.47 ! |
ATOM HN H 0.31 ! HN-N
ATOM CA CT1 0.07 ! | HB1
ATOM HA HB 0.09 ! | /
GROUP ! HA-CA--CB-HB2
ATOM CB CT3 -0.27 ! | \
ATOM HB1 HA 0.09 ! | HB3
ATOM HB2 HA 0.09 ! O=C
ATOM HB3 HA 0.09 ! |
GROUP !
C TOPOLOGY FILES 96

ATOM C C 0.51
ATOM O O -0.51

The ALA residue continues by defining connectivity, with each BOND state-
ment followed by a list of pairs of atoms to be connected with bonds. The
DOUBLE statement is a synonym for BOND and does not affect the resulting
PSF file. Observe that the atom C is bonded to +N, the N of the following
residue. A bond between N and -C will be provided by the preceding residue.
The order of bonds, or of the atoms within a bond, is not significant.
BOND CB CA N HN N CA
BOND C CA C +N CA HA CB HB1 CB HB2 CB HB3
DOUBLE O C

As noted above, the angle and dihedral terms will be autogenerated and
are therefore not listed for this residue. The less common improper dihedrals
(normally just called impropers), however, must be listed explicitly. In this case
there are two impropers, which maintain the planarity of the peptide bonds.
As with dihedrals, the order of atoms within an improper may be reversed. As
shown below, impropers are specified by the IMPR statement followed by sets of
four atoms, with the central atom to which the other three are bonded typically
listed first.
IMPR N -C CA HN C CA +N O

The CMAP correction terms should also be listed explicitly since they are
only applied to the backbone dihedrals, indicated following the CMAP state-
ment.
CMAP -C N CA C N CA C +N

Explicit hydrogen bond terms are no longer present in the CHARMM force
field and are therefore not calculated by NAMD. The DONOR and ACCEPTOR
statements, shown below, specify pairs of atoms eligible to form hydrogen bonds.
The psfgen module in VMD ignores these statements and does not incorporate
hydrogen bonding information into the PSF file.
DONOR HN N
ACCEPTOR O C

Finally in the residue definition are the internal coordinate IC statements.

For each set of four atoms 1 2 3 4, the IC specifies in order the bond length 1-2,
the angle 1-2-3, the dihedral 1-2-3-4, the angle 2-3-4, and the bond length 3-4.
With this set of data, the position of atom 1 may be determined based on the
positions of atoms 2–4, and the position of atom 4 may be determined from the
positions of atoms 1–3, allowing the recursive generation of coordinates for all
atoms in the structure based on a three-atom seed. Improper IC statements are
indicated by a * preceding the third atom, the atom to which the other three
are bonded, as in 1 2 *3 4. The order of atoms in an IC statement is different
from that of an IMPR statement, and values provide the length 1-3, the angle
1-3-2, the dihedral 1-2-3-4, the angle 2-3-4, and the length 3-4.
C TOPOLOGY FILES 97

IC -C CA *N HN 1.3551 126.4900 180.0000 115.4200 0.9996

IC -C N CA C 1.3551 126.4900 180.0000 114.4400 1.5390
IC N CA C +N 1.4592 114.4400 180.0000 116.8400 1.3558
IC +N CA *C O 1.3558 116.8400 180.0000 122.5200 1.2297
IC CA C +N +CA 1.5390 116.8400 180.0000 126.7700 1.4613
IC N C *CA CB 1.4592 114.4400 123.2300 111.0900 1.5461
IC N C *CA HA 1.4592 114.4400 -120.4500 106.3900 1.0840
IC C CA CB HB1 1.5390 111.0900 177.2500 109.6000 1.1109
IC HB1 CA *CB HB2 1.1109 109.6000 119.1300 111.0500 1.1119
IC HB1 CA *CB HB3 1.1109 109.6000 -119.5800 111.6100 1.1114

It was mentioned above that special treatment is required for the first or last
residue in a chain, and that this is implemented as a patch. The following is the
default first-residue patch for proteins, NTER. The syntax is almost identical
to a normal residue, but the ATOM statements may refer to either add a new
atom (HT1, HT2, HT3) or modify the type and charge of an existing atom of
the given name (N, CA, HA). The DELETE statement operates as one would
expect, removing the atom (HN) and any bonds that include it.

PRES NTER 1.00 ! standard N-terminus

GROUP ! use in generate statement
ATOM N NH3 -0.30 !
ATOM HT1 HC 0.33 ! HT1
ATOM HT2 HC 0.33 ! (+)/
ATOM HT3 HC 0.33 ! --CA--N--HT2
ATOM CA CT1 0.21 ! | \
ATOM HA HB 0.10 ! HA HT3
DELETE ATOM HN
BOND HT1 N HT2 N HT3 N
DONOR HT1 N
DONOR HT2 N
DONOR HT3 N
IC HT1 N CA C 0.0000 0.0000 180.0000 0.0000 0.0000
IC HT2 CA *N HT1 0.0000 0.0000 120.0000 0.0000 0.0000
IC HT3 CA *N HT2 0.0000 0.0000 120.0000 0.0000 0.0000

The comment “use in generate statement” indicates that the NTER patch is
used during segment generation, and is applied before angles and dihedrals are
generated. An example of the other type of patch, those applied after segment
generation, is the LINK patch given below. The patch statement requires, in
this case, two residues as arguments, the first being the last residue of a segment
generated without the CTER patch and the second being the first residue of a
segment generated without the NTER patch (the comments have the N and C
termini reversed). The numbers 1 and 2 preceding atom names indicate which
argument residue the named atom belongs to. Since angles and dihedrals will
not be regenerated, they are enumerated.

PRES LINK 0.00 ! linkage for IMAGES or for joining segments

! 1 refers to previous (N terminal)
C TOPOLOGY FILES 98

! 2 refers to next (C terminal)

! use in a patch statement
! follow with AUTOgenerate ANGLes DIHEdrals command
BOND 1C 2N
!the need for the explicit specification of angles and dihedrals in
!patches linking images has not been tested
!ANGLE 1C 2N 2CA 1CA 1C 2N
!ANGLE 1O 1C 2N 1C 2N 2HN
!DIHE 1C 2N 2CA 2C 1C 2N 2CA 2HA 1C 2N 2CA 2CB
!DIHE 1HA 1CA 1C 2N 1N 1CA 1C 2N 1CB 1CA 1C 2N
!DIHE 1CA 1C 2N 2HN 1CA 1C 2N 2CA
!DIHE 1O 1C 2N 2HN 1O 1C 2N 2CA
IMPR 2N 1C 2CA 2HN 1C 1CA 2N 1O
IC 1N 1CA 1C 2N 0.0000 0.0000 180.0000 0.0000 0.0000
IC 2N 1CA *1C 1O 0.0000 0.0000 180.0000 0.0000 0.0000
IC 1CA 1C 2N 2CA 0.0000 0.0000 180.0000 0.0000 0.0000
IC 1C 2N 2CA 2C 0.0000 0.0000 180.0000 0.0000 0.0000
IC 1C 2CA *2N 2HN 0.0000 0.0000 180.0000 0.0000 0.0000

These types of patches are used to alter protonation states (ASPP, GLUP,
HS2), create disulphide bonds (DISU), attach HEME groups (PHEM) and
their ligands (PLO2, PLIG), and even remove unwanted autogenerated angles
(FHEM).
The following is the complete residue definition for GLY, the smallest of the
amino acids. Compare GLY to the ALA residue dissected above.
RESI GLY 0.00
GROUP
ATOM N NH1 -0.47 ! |
ATOM HN H 0.31 ! N-H
ATOM CA CT2 -0.02 ! |
ATOM HA1 HB 0.09 ! |
ATOM HA2 HB 0.09 ! HA1-CA-HA2
GROUP ! |
ATOM C C 0.51 ! |
ATOM O O -0.51 ! C=O
! |
BOND N HN N CA C CA
BOND C +N CA HA1 CA HA2
DOUBLE O C
IMPR N -C CA HN C CA +N O
CMAP -C N CA C N CA C +N
DONOR HN N
ACCEPTOR O C
IC -C CA *N HN 1.3475 122.8200 180.0000 115.6200 0.9992
IC -C N CA C 1.3475 122.8200 180.0000 108.9400 1.4971
IC N CA C +N 1.4553 108.9400 180.0000 117.6000 1.3479
IC +N CA *C O 1.3479 117.6000 180.0000 120.8500 1.2289
IC CA C +N +CA 1.4971 117.6000 180.0000 124.0800 1.4560
IC N C *CA HA1 1.4553 108.9400 117.8600 108.0300 1.0814
C TOPOLOGY FILES 99

IC N C *CA HA2 1.4553 108.9400 -118.1200 107.9500 1.0817

PATCHING FIRS GLYP

Unlike ALA and most protein residues, in which CA is bonded to HA and

CB, in GLY it is bonded to a pair of hydrogens, HA1 and HA2. Also, the
hydrogen bonded to N is named H rather than HN. For these reasons, the
default NTER patch cannot be applied to GLY and the PATCHING statement
is used to change the default first-residue patch for GLY to GLYP. Similarly, the
LINK patch above cannot be used for GLY residues, so the additional patches
LIG1, LIG2, and LIG3 are provided to link GLY to non-GLY, non-GLY to GLY,
and GLY to GLY.
Water is also defined as a residue, as shown below, but some care is re-
quired for its use. A third bond, not required by NAMD, is included to allow
CHARMM to make the water molecule rigid. This ring structure would confuse
angle and dihedral autogeneration, so segments of water must be generated with
autogeneration disabled, and therefore an explicit angle is also included. If this
third bond is removed from the water topology, then autogeneration must still
be disabled to avoid duplicating the angle, unless the angle is also removed.

RESI TIP3 0.000 ! tip3p water model, generate using noangle nodihedral
GROUP
ATOM OH2 OT -0.834
ATOM H1 HT 0.417
ATOM H2 HT 0.417
BOND OH2 H1 OH2 H2 H1 H2 ! the last bond is needed for shake
ANGLE H1 OH2 H2 ! required
ACCEPTOR OH2
PATCHING FIRS NONE LAST NONE

The residue definitions for the ions below are exceedingly simple.

RESI SOD 1.00 ! Sodium Ion

GROUP
ATOM SOD SOD 1.00
PATCHING FIRST NONE LAST NONE

RESI CLA -1.00 ! Chloride Anion

GROUP
ATOM CLA CLA -1.00
PATCHING FIRST NONE LAST NONE

We have discussed only those parts of the topology files associated with
proteins and solvent. There is much additional information regarding proteins,
not to mention lipids and nucleic acids, included in the comments in the topology
files themselves.
D PARAMETER FILES 100

D Parameter Files
A CHARMM forcefield parameter file contains all of the numerical constants
needed to evaluate forces and energies, given a PSF structure file and atomic
coordinates. The parameter file is closely tied to the topology file that was used
to generate the PSF file, and the two are typically distributed together and
given matching names.
The current versions of the CHARMM forcefield are CHARMM22 for pro-
teins and CHARMM27 for lipids and nucleic acids including CMAP correction to
proteins. The individual parameter files are named, respectively, par_all22_prot_cmap.inp,
par_all27_lipid.prm, and par_all27_na.prm. To enable hybrid systems,
combinations are also provided, named par_all27_na_lipid.prm, par_all27_prot_lipid.prm,
and par_all27_prot_na.prm which can all be found in the CHARMM31 re-
lease. While the tools used with NAMD allow multiple topology and parameter
files to be used simultaneously, it is preferable to use these pre-combined files.
The CHARMM parameters are available for download from the MacKerell web
site:
http://mackerell.umaryland.edu/charmm_ff.shtml

We will examine par_all27_prot_lipid.prm; the other files are similar. At

the beginning of the file is the header, indicated by lines beginning with *s:

*>CHARMM22 All-Hydrogen Parameter File for Proteins and Lipids <<

*>>>>> Includes phi, psi cross term map (CMAP) correction <<<<<<<
*>>>>>>>>>>>>>>>>>>>>>> July, 2003 <<<<<<<<<<<<<<<<<<<<<<<<<<
* All comments to ADM jr. via the CHARMM web site: www.charmm.org
* parameter set discussion forum
*

Comments in the parameter file are indicated by ! anywhere on a line. The

next part of the file is a long set of comments containing references to the papers
and other sources of the parameters in the file:

! references
!
!PROTEINS
!
!MacKerell, A.D., Jr,. Feig, M., Brooks, C.L., III, Extending the
!treatment of backbone energetics in protein force fields: limitations
!of gas-phase quantum mechanics in reproducing protein conformational
!distributions in molecular dynamics simulations, Journal of
!Computational Chemistry, 25: 1400-1415, 2004.
!
!MacKerell, Jr., A. D.; Bashford, D.; Bellott, M.; Dunbrack Jr., R.L.;
!Evanseck, J.D.; Field, M.J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.;
!Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F.T.K.; Mattos,
!C.; Michnick, S.; Ngo, T.; Nguyen, D.T.; Prodhom, B.; Reiher, III,
!W.E.; Roux, B.; Schlenkrich, M.; Smith, J.C.; Stote, R.; Straub, J.;
D PARAMETER FILES 101

Figure 24: Internal coordinates for bonded interactons: r governs bond stretch-
ing; θ represents the bond angle term; φ gives the dihedral angle; the small
out-of-plane angle α is governed by the so-called “improper" dihedral angle ψ.

!Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. All-atom

!empirical potential for molecular modeling and dynamics Studies of
!proteins. Journal of Physical Chemistry B, 1998, 102, 3586-3616.
!

The first set of entries in the parameter file are those for bonds, indicated
by the BONDS keyword. Each entry consists of a pair of atom types, a spring
constant, and an equilibrium length. The bond potential function is Kb (b−b0 )2 ,
where b is the bond length in Angstroms. Bonds are a stiff degree of freedom
in biomolecules, so the energy function is only accurate for values near the
equilibrium length. Entries are present for every type of bond present in the
topology file. Fig. 24 illustrates how bond length, bond angle, dihedral angle
and improper angle are defined. The beginning of the bonds section is shown
below:
BONDS
!
!V(bond) = Kb(b - b0)**2
!
!Kb: kcal/mole/A**2
!b0: A
!
!atom type Kb b0
!
!Carbon Dioxide
CST OST 937.96 1.1600 ! JES
!Heme to Sulfate (PSUL) link
SS FE 250.0 2.3200 !force constant a guess
!equilbrium bond length optimized to reproduce
!CSD survey values of
!2.341pm0.01 (mean, standard error)
!adm jr., 7/01
C C 600.000 1.3350 ! ALLOW ARO HEM
D PARAMETER FILES 102

! Heme vinyl substituent (KK, from propene (JCS))

CA CA 305.000 1.3750 ! ALLOW ARO
! benzene, JES 8/25/89
CE1 CE1 440.000 1.3400 !
! for butene; from propene, yin/adm jr., 12/95
CE1 CE2 500.000 1.3420 !
! for propene, yin/adm jr., 12/95
CE1 CT2 365.000 1.5020 !
! for butene; from propene, yin/adm jr., 12/95
CE1 CT3 383.000 1.5040 !
! for butene, yin/adm jr., 12/95
CE2 CE2 510.000 1.3300 !
! for ethene, yin/adm jr., 12/95

The CHARMM potential function is designed to confuse physicists, as the

form is Kb (b − b0 )2 , rather than the traditional 12 kb (b − b0 )2 , and therefore
the Kb given in the parameter files is half the value of a traditional spring
constant. Also, while the units kcal/mol for energy, Angstroms for length,
atomic masses, electron charges, and either fs or ps for time may be convenient
for input and output, tortuous unit conversions are required to express the
equations of motion.
The next section gives parameters for every type of angle present in the
topology file, indicated by the ANGLES keyword. The angle potential function
is Kθ (θ − θ0 )2 , where θ is the measure of the angle in degrees. Angles are a
stiff degree of freedom in biomolecules as well, so the energy function is only
accurate for values near the equilibrium angle. Since angles are formed from
combinations of bonds, there are many more types of angles than types of bonds.
Each entry consists of three atom types, a spring constant, and an equilibrium
angle. A small minority of entries also contant Urey-Bradley parameters, which
are a spring constant and equilibrium length for a bond-like term between the
first and third atoms in the angle. The beginning of the angles section is shown
below, with a Urey-Bradley term in the first entry only:
ANGLES
!
!V(angle) = Ktheta(Theta - Theta0)**2
!
!V(Urey-Bradley) = Kub(S - S0)**2
!
!Ktheta: kcal/mole/rad**2
!Theta0: degrees
!Kub: kcal/mole/A**2 (Urey-Bradley)
!S0: A
!
!atom types Ktheta Theta0 Kub S0
!
!Carbon Dioxide, JES
OST CST OST 3000.00 180.0000 ! CO2, JES
!Heme to Sulfate (PSUL) link
D PARAMETER FILES 103

CS SS FE 50.0 100.6 !force constant a guess

!equilibrium angle optimized to reproduce
!CSD survey values
!107.5pm0.6 (mean, standard error)
!adm jr., 7/01
SS FE NPH 100.0 90.0 !force constant a guess
!adm jr., 7/01
!
CA CA CA 40.000 120.00 35.00 2.41620 ! ALLOW ARO
! JES 8/25/89
CE1 CE1 CT2 48.00 123.50 !
! for 2-butene, yin/adm jr., 12/95
CE1 CE1 CT3 48.00 123.50 !
! for 2-butene, yin/adm jr., 12/95
CE1 CT2 CT3 32.00 112.20 !
! for 1-butene; from propene, yin/adm jr., 12/95
CE2 CE1 CT2 48.00 126.00 !
! for 1-butene; from propene, yin/adm jr., 12/95
CE2 CE1 CT3 47.00 125.20 !
! for propene, yin/adm jr., 12/95

The next section gives parameters for every type of dihedral present in the
topology file; there are even more dihedrals than there are angles. Since di-
hedrals represent the energy of rotation around a covalent bond, which is the
source of most conformational flexibility in biomolecules, they must provide a
smooth energy for 360 degrees. This is done in most cases with a single sinu-
soid, Kχ (1 + cos(n(χ − δ)) where χ is the angle between the plane containing
the first three atoms in the dihedral and the plane containing the last three.
The “multiplicity” n is typically 1, 2, or 3, although for a small number of cases
two or three terms with different values of n are provided for the same atom
types. You may can observe in the excerpts below that the dihedral spring con-
stants are one to two orders of magnitude lower than for angles, with an order
of magnitude difference between flexible and inflexible dihedrals.

DIHEDRALS
!
!V(dihedral) = Kchi(1 + cos(n(chi) - delta))
!
!Kchi: kcal/mole
!n: multiplicity
!delta: degrees
!
!atom types Kchi n delta
!
!Heme to Sulfate (PSUL) link
X FE SS X 0.0000 4 0.00 ! guess
!adm jr., 7/01
X CS SS X 0.0000 3 0.20 ! guess
!from methanethiol, HS S CT3 HA
D PARAMETER FILES 104

!adm jr., 7/01

C CT1 NH1 C 0.2000 1 180.00 ! ALLOW PEP

! ala dipeptide update for new C VDW Rmin, adm jr., 3/3/93c
C CT2 NH1 C 0.2000 1 180.00 ! ALLOW PEP
! ala dipeptide update for new C VDW Rmin, adm jr., 3/3/93c
C N CP1 C 0.8000 3 0.00 ! ALLOW PRO PEP
! 6-31g* AcProNH2, ProNH2, 6-31g*//3-21g AcProNHCH3 RLD 4/23/93
CA CA CA CA 3.1000 2 180.00 ! ALLOW ARO
! JES 8/25/89
CA CPT CPT CA 3.1000 2 180.00 ! ALLOW ARO
! JWK 05/14/91 fit to indole

Because of the large numbers of dihedral terms required to describe a com-

plete protein, the wildcard atom type X is occasionally used. These parameters
will be used in NAMD if a more specific match is not found elsewhere in the
parameter file.

!X C C X 4.0000 2 180.00 ! ALLOW HEM

! Heme (6-liganded): substituents (KK 05/13/91)
X C NC2 X 2.2500 2 180.00 ! ALLOW PEP POL ARO
! 9.0->2.25 GUANIDINIUM (KK)
X CD OH1 X 2.0500 2 180.00 ! ALLOW PEP POL ARO ALC
! adm jr, 10/17/90, acetic acid C-Oh rotation barrier
X CD OS X 2.0500 2 180.00 ! ALLOW PEP POL
! adm jr. 3/19/92, from lipid methyl acetate
X CE1 CE1 X 0.1500 1 0.00
! 2-butene, adm jr., 2/00 update
X CE1 CE1 X 8.5000 2 180.00
! 2-butene, adm jr., 2/00 update
X CE2 CE2 X 4.9000 2 180.00 !
! for ethene, yin/adm jr., 12/95

The final bond-like terms in the parameter file are impropers, which are
used exclusively and explicitly in the molecular topology to maintain planarity.
As such, the harmonic form Kψ (ψ − ψ0 )2 with a large spring constant and ψ0
typically zero is used to restrain deformations among an atom and three atoms
bonded to it. As with dihedrals, ψ is angle between the plane containing the
first three atoms and the plane containing the last three. Notice below that
wildcard atom types occur in the second and third positions, rather than the
first and fourth as in dihedrals.

IMPROPER
!
!V(improper) = Kpsi(psi - psi0)**2
!
!Kpsi: kcal/mole/rad**2
!psi0: degrees
!note that the second column of numbers (0) is ignored
D PARAMETER FILES 105

!
!atom types Kpsi psi0
!
CPB CPA NPH CPA 20.8000 0 0.0000 ! ALLOW HEM
! Heme (6-liganded): porphyrin macrocycle (KK 05/13/91)
CPB X X CE1 90.0000 0 0.0000 ! ALLOW HEM
! Heme (6-liganded): substituents (KK 05/13/91)
CT2 X X CPB 90.0000 0 0.0000 ! ALLOW HEM
! Heme (6-liganded): substituents (KK 05/13/91)
CT3 X X CPB 90.0000 0 0.0000 ! ALLOW HEM
! Heme (6-liganded): substituents (KK 05/13/91)
!HA C C HA 20.0000 0 0.0000 ! ALLOW PEP POL ARO

The next section gives the interpolation values for CMAP, which is a cor-
rection map to the backbone dihedral energy. Each φ and ψ dihedral angle,
ranging from -180 to 180, is divided into 24 grid points; the energy correction
at each point is given. A continuous function can be constructed using these
values along with an interpolation formula, which can be found in the references
below.
CMAP
! 2D grid correction data. The following surfaces are the correction
! to the CHARMM22 phi, psi alanine, proline and glycine dipeptide surfaces.
! Use of CMAP requires generation with the topology file containing the
! CMAP specifications along with version 31 or later of CHARMM. Note that
! use of "skip CMAP" yields the charmm22 energy surfaces.
!
! references
!MacKerell, A.D., Jr., Feig, M., Brooks, C.L., III, Accurate Treatment of
!Protein Backbone Conformational Energetics in Empirical Force Fields, Submitted
!for publication.

!MacKerell, A.D., Jr., Feig, M., Brooks, C.L., III, Improved Treatment of the
!Protein Backbone in Empirical Force Fields, Journal of the American Chemical
!Society, In Press.

! alanine map
C NH1 CT1 C NH1 CT1 C NH1 24

!-180
0.126790 0.768700 0.971260 1.250970 2.121010
2.695430 2.064440 1.764790 0.755870 -0.713470
0.976130 -2.475520 -5.455650 -5.096450 -5.305850
-3.975630 -3.088580 -2.784200 -2.677120 -2.646060
-2.335350 -2.010440 -1.608040 -0.482250

!-165
-0.802290 1.377090 1.577020 1.872290 2.398990
2.461630 2.333840 1.904070 1.061460 0.518400
D PARAMETER FILES 106

-0.116320 -3.575440 -5.284480 -5.160310 -4.196010

-3.276210 -2.715340 -1.806200 -1.101780 -1.210320
-1.008810 -0.637100 -1.603360 -1.776870

Finally we come to the nonbonded interaction parameters. The NON-

BONDED statement includes a list of parameters which are used as defaults
by the program CHARMM, but are ignored by NAMD. Those shown below
correspond to the NAMD settings exclude scaled1-4, switching on, pairlistdist
14.0, cutoff 12.0, switchdist 10.0, dielectric 1.0, and 1-4scaling 1.0:
NONBONDED nbxmod 5 atom cdiel shift vatom vdistance vswitch -
cutnb 14.0 ctofnb 12.0 ctonnb 10.0 eps 1.0 e14fac 1.0 wmin 1.5
!adm jr., 5/08/91, suggested cutoff scheme

Recall that the partial charge of each atom is specified in the topology and
PSF files and is independent of the atom type. Therefore the only type-based
parameters are for the van der Waals interactions, which are represented by
the classic Lennard-Jones potential (expressed in the somewhat unconventional
form) [(rmin /r)12 − 2(rmin /r)6 ]. Observe that at r = rmin the force is zero
and the energy is −. Rather than providing a different value of epsilon could
be provided for every possible combination of atom types, only one value is
provided per type and inter-type interactions are calculated using the sum of
the radii rmin /2 and the geometric mean of the well-depths . By convention,
the  values are negative in the parameter file, as seen here:
!
!V(Lennard-Jones) = Eps,i,j[(Rmin,i,j/ri,j)**12 - 2(Rmin,i,j/ri,j)**6]
!
!epsilon: kcal/mole, Eps,i,j = sqrt(eps,i * eps,j)
!Rmin/2: A, Rmin,i,j = Rmin/2,i + Rmin/2,j
!
!atom ignored epsilon Rmin/2 ignored eps,1-4 Rmin/2,1-4
!
C 0.000000 -0.110000 2.000000 ! ALLOW PEP POL ARO
! NMA pure solvent, adm jr., 3/3/93
CA 0.000000 -0.070000 1.992400 ! ALLOW ARO
! benzene (JES)
CC 0.000000 -0.070000 2.000000 ! ALLOW PEP POL ARO
! adm jr. 3/3/92, acetic acid heat of solvation
CD 0.000000 -0.070000 2.000000 ! ALLOW POL
! adm jr. 3/19/92, acetate a.i. and dH of solvation
CE1 0.000000 -0.068000 2.090000 !
! for propene, yin/adm jr., 12/95
CE2 0.000000 -0.064000 2.080000 !
! for ethene, yin/adm jr., 12/95
CM 0.000000 -0.110000 2.100000 ! ALLOW HEM
! Heme (6-liganded): CO ligand carbon (KK 05/13/91)

When the scaled1-4 exclusion policy is used (as it should with the CHARMM
force field) nonbonded interactions of atoms separated by three bonds (i.e.,
D PARAMETER FILES 107

atoms 1 and 4 in the chain 1-2-3-4) are modified. Even if the scaling factor for
electrostatics is 1.0 (as it should be for modern CHARMM force fields), special
modified van der Waals parameters are used for 1-4 pairs of atoms for which
they are specified, as in the examples below.

!atom ignored epsilon Rmin/2 ignored eps,1-4 Rmin/2,1-4

!
CP1 0.000000 -0.020000 2.275000 0.000000 -0.010000 1.900000 ! ALLOW ALI
! alkane update, adm jr., 3/2/92
CP2 0.000000 -0.055000 2.175000 0.000000 -0.010000 1.900000 ! ALLOW ALI
! alkane update, adm jr., 3/2/92
CP3 0.000000 -0.055000 2.175000 0.000000 -0.010000 1.900000 ! ALLOW ALI
! alkane update, adm jr., 3/2/92

The parameter file ends with a reference to parameters for explicit hydrogen
bond energy terms. These are obsolete, no longer present in the CHARMM
force field, and therefore not implemented by NAMD.

HBOND CUTHB 0.5 ! If you want to do hbond analysis (only), then use
! READ PARAM APPEND CARD
! to append hbond parameters from the file: par_hbond.inp

END

For information on how the parameters have been derived, you must consult
the corresponding publications referenced in the parameter files themselves or
listed on the MacKerell web site http://mackerell.umaryland.edu/charmm_ff.shtml.
E NAMD CONFIGURATION FILES 108

E NAMD Configuration Files

The NAMD configuration file (also called a config file, .conf file, or .namd file)
is given to NAMD on the command line and specifies virtually everything about
the simulation to be done. The only exceptions are details relating to the parallel
execution environment, which vary between platforms. Therefore, the config file
should be portable between machines, platforms, or numbers of processors in a
run, as long as the referenced input files are available.
As a convenience, on startup NAMD will switch to the directory that con-
tains the config file, so that all file paths are relative to that directory. NAMD
also accepts multiple config files on the command line, but changes directories
before parsing each file, so it is best to keep everything in the same directory
when using multiple config files.
NAMD parses its configuration file using the Tcl scripting language, which
you should be familiar with if you have done any serious work using VMD.
However, to remain compatible with config files from earlier version of NAMD
which did not include Tcl, several compromises have been made. Tcl is a case
sensitive language, so the $temp and $TEMP are different variables. NAMD’s
config file options are implemented as case-insensitive Tcl commands, however,
so temperature 300, Temperature 300, or even tEmPerATURe 300 will all have
the same effect.
NAMD config files were historically order-independent, and the whole file
was parsed before any files were read or calculations done. Now that NAMD
incorporates more advanced scripting, the config file is parsed up to the point
where one of the case-sensitive “action” commands such as minimize or run are
encountered; at this point the input files are read and real calculation begins.
Beyond this point, the config file is order-dependent and only these action com-
mands and a subset of configuration options may be given. Also, at this point
any comments must follow the rules of Tcl, using a semicolon to end a command
before using # to start a comment on the same line; it is good practice to follow
these rules everywhere in the file.
We will now walk through a NAMD configuration file template that contains
options for common simulation methods in NAMD. A version of this file should
be available in the tutorial materials as sample.conf.
First we specify the files that contain the molecular structure and initial
conditions. A coordinates PDB file is always required, even if the actual
coordinates will come from a binary coordinates file.

structure mypsf.psf
coordinates mypdb.pdb

Setting the Tcl variable $temperature makes it easy to change the target
temperature for many options.

set temperature 310 ;# target temperature used several times below

E NAMD CONFIGURATION FILES 109

If we are starting from scratch, we’ll use the coordinates from the PDB
file and take velocities randomly from a Boltzmann distribution, using the
$temperature variable.
# starting from scratch
temperature $temperature ;# initialize velocities randomly

Otherwise we’ll read in the binary output files of the previous run. We
use the Tcl variable $inputname to avoid errors in typing the input file names,
since we will end up copying and modifying this file several times during the
course of a long simulation. The extendedSystem file holds the periodic cell
dimensions that are needed to continue after a constant pressure simulation.
In order to start numbering timesteps with the final step of the previous run,
we use firsttimestep to specify the timestep on which the input coordinates
and velocities were written. The number of steps calculated in this run will be
numsteps - firsttimestep.
# continuing a run
set inputname myinput ;# only need to edit this in one place!
binCoordinates $inputname.coor ;# coordinates from last run (binary)
binVelocities $inputname.vel ;# velocities from last run (binary)
extendedSystem $inputname.xsc ;# cell dimensions from last run
firsttimestep 50000 ;# last step of previous run
numsteps 100000 ;# run stops when this step is reached

The outputName prefix will be used to create all of the trajectory (.dcd, .xst),
output (.coor, vel, .xsc), and restart (.restart.coor, .restart.vel, .restart.xsc) files
generated by this run. Be careful that this is different from $myinput, or the
job will overwrite its input files!
outputName myoutput ;# base name for output from this run

We will write restart files, coordinate trajectory .dcd files, and extended
system (periodic cell) trajectory .xst files at regular intervals.
restartfreq 500 ;# 500 steps = every 1ps
dcdfreq 500
xstFreq 500

The default is to print energies every steps, but that makes for a very long
file, so we’ll cut it down to every 100 steps. We’ll also enable the printing of
performance information every 1000 steps.
outputEnergies 100 ;# 100 steps = every 0.2 ps
outputTiming 1000 ;# shows time per step and time to completion

Next are the parameter file itself and the options that control the nonbonded
potential functions. These are mostly specified by the CHARMM force field,
but the cutoff and other distances may be shortened when full electrostatics are
used; Just be sure that you adjust all of the distances together. The pairlistdist
can be made longer if you see warnings in the NAMD output, but making it
bigger than necessary will reduce performance.
E NAMD CONFIGURATION FILES 110

# Force-Field Parameters
paraTypeCharmm on
parameters par_all27_prot_lipid.inp

# These are specified by CHARMM

exclude scaled1-4
1-4scaling 1.0
switching on

# You have some freedom choosing the cutoff

cutoff 12. ;# may use smaller, maybe 10., with PME
switchdist 10. ;# cutoff - 2.

# Promise that atom won’t move more than 2A in a cycle

pairlistdist 14. ;# cutoff + 2.

The chosen stepspercycle is related to pairlistdist above, as pairlists

are generated at the beginning of each cycle and are assumed to contain every
atom that passes within the cutoff distance during entire cycle. Warnings will
appear in the output if pairlistdist is too short (or stepspercycle is too
large). The length of the simulation must be an integer number of cycles.

stepspercycle 10 ;# redo pairlists every ten steps

The integration timestep is normally limited to 1 fs. This can be extended

to 2 fs by fixing the length of all bonds in the molecule involving hydrogen
atoms via rigidBonds all, which also makes water molecules completely rigid.
To only make water rigid, use rigidBonds water and a 1 fs timestep. For
any simulations involving water molecules, one should make water rigid since
water molecules have been parametrized as rigid water. Nonbonded forces must
be calculated at least every 2 fs, which in this example is every step. Full
electrostatics forces (from particle mesh Ewald, discussed below) are evaluated
every other step in this example. nonbondedFreq and fullElectFrequency
must evenly divide stepspercycle.

# Integrator Parameters
timestep 2.0 ;# 2fs/step
rigidBonds all ;# needed for 2fs steps
nonbondedFreq 1 ;# nonbonded forces every step
fullElectFrequency 2 ;# PME only every other step

Langevin dynamics balances friction with random noise to drive each atom
in the system towards a target temperature. The following parameters are good
for equilibration, but a langevinDamping as low as 1. would be sufficient for
simulation. Higher damping may be necessary if, for example, work done on the
system by steering forces is driving up the temperature.

# Constant Temperature Control

langevin on ;# langevin dynamics
E NAMD CONFIGURATION FILES 111

langevinDamping 1. ;# damping coefficient of 1/ps

langevinTemp $temperature ;# random noise at this level
langevinHydrogen no ;# don’t couple bath to hydrogens

If the simulation will use periodic boundary conditions, they are specified
as shown below. Like the temperature option, these should only be given
when starting a simulation from scratch, since the basis vectors will fluctuate
during constant pressure simulation and updated values need to be read via
extendedSystem from a .xsc file.

# Periodic Boundary conditions

cellBasisVector1 31.2 0. 0. ;# vector to the next image
cellBasisVector2 0. 44.8 0.
cellBasisVector3 0. 0 51.3
cellOrigin 0. 0. 0. ;# the *center* of the cell

The cell origin is the one coordinate that does not change during constant
pressure simulation. This point plays a role in calculating the pressure contri-
bution due to fixed atoms, harmonic restraints, and other “external” forces. To
minimize pressure fluctuations, this point should be close to the center of the
macromolecule of interest.
The origin also defines the center of the periodic cell to which coordinates will
be wrapped on output if the following options below are used. With wrapping,
some molecules will jump between sides of the cell in the trajectory file to yield
the periodic image nearest to the origin. Without wrapping, molecules will
have smooth trajectories, but water in the trajectory may appear to explode as
individual molecules diffuse. Wrapping only affects output, not the correctness
of the simulation.

wrapWater on ;# wrap water to central cell

wrapAll on ;# wrap other molecules too
wrapNearest off ;# use for non-rectangular cells

Particle mesh Ewald (PME) full electrostatics are more accurate and less
expensive than larger cutoffs, and are recommended for most work. PME is
only applicable to periodic simulations, and the user must specify a grid corre-
sponding to the size of the periodic cell. PME grid dimensions should have small
integer factors only and be greater than or equal to length of the basis vector.
To manually define the grid size instead of letting NAMD choose the dimensions
for you according to these guidelines, replace the GridSpacing command with
explicit GridSize commands instead.

#PME (for full-system periodic electrostatics)

PME yes

# let NAMD determine grid

PMEGridSpacing 1.0
E NAMD CONFIGURATION FILES 112

# manually specify grid

#PMEGridSizeX 32 ;# 2^5, close to 31.2
#PMEGridSizeY 45 ;# 3^2 * 5, close to 44.8
#PMEGridSizeZ 54 ;# 2 * 3^3, close to 51.3

Constant pressure is recommended for periodic simulations. Using group-

based pressure to control the periodic cell fluctuations is desirable because the
atom-based pressure has more high-frequency noise. useFlexibleCell is useful
for anisotropic systems such as membranes, allowing the height, length, and
width of the cell to vary independently, possibly even fixing the lipid cross-
sectional (x-y plane) area with useConstantArea. For a protein surrounded by
water there is nothing to prevent the cell from becoming highly extended in one
dimension, so it is better to choose useFlexibleCell no in this case.

# Constant Pressure Control (variable volume)

useGroupPressure yes ;# needed for rigid bonds
useFlexibleCell no ;# no for water box, yes for membrane
useConstantArea no ;# no for water box, maybe for membrane

The actual parameters of the Nosé-Hoover Langevin piston method control

the target pressure and the dynamical properties of the barostat. A “piston”
with a longer period (i.e., larger mass) will better damp out fluctuations in
the instantaneous pressure. Langevin dynamics is applied to the piston itself
coupling it to a heat bath with a damping constant of 1/langevinPistonDecay.
We set langevinPistonDecay smaller that langevinPistonPeriod to ensure
that harmonic oscillations in the periodic cell are overdamped.

langevinPiston on
langevinPistonTarget 1.01325 ;# pressure in bar -> 1 atm
langevinPistonPeriod 100. ;# oscillation period around 100 fs
langevinPistonDecay 50. ;# oscillation decay time of 50 fs
langevinPistonTemp $temperature ;# coupled to heat bath

When initially assembling a system it is sometimes useful to fix the protein

while equilibrating water or lipids around it. These options read a PDB file
containing flags for the atoms to fix. The number and order of atoms in the
PDB file must exactly match that of the PSF and other input files.

fixedAtoms on
fixedAtomsFile myfixedatoms.pdb ;# flags are in this file
fixedAtomsCol B ;# set beta non-zero to fix an atom

The interactive MD features of NAMD and VMD allow you to connect to

a running simulation to apply steering forces manually. These options affect
performance, and should therefore not be used unless you are actually steering
the simulation.
E NAMD CONFIGURATION FILES 113

IMDon on
IMDport 3000 ;# port number (enter it in VMD)
IMDfreq 1 ;# send every 1 frame
IMDwait no ;# wait for VMD to connect before running?

Now we minimize the system to eliminate bad initial contacts, reinitialize the
velocities to the desired target temperature (since minimization sets velocities
to zero), and run for 100 ps. We could accomplish the same thing with two
different NAMD runs using the numsteps and minimization options. Scripting
commands such as those below override numsteps.

minimize 1000 ;# lower potential energy for 1000 steps

reinitvels $temperature ;# since minimization zeros velocities
run 50000 ;# 100ps

Naturally, these are not all of the configuration options accepted by NAMD,
but only a rapid introduction to those you are most likely to encounter. Docu-
mentation for these and many other options can be found in the Users Guide.
F NAMD STANDARD OUTPUT 114

F NAMD Standard Output

When NAMD runs, important information on the progress of the simulation is
written to standard output, appearing on the console unless output from NAMD
is redirected to a file. This output is designed to be easy for both humans and
programs to interpret.
At the beginning of the file, the version of NAMD being used is presented,
along with other useful information, some of which is reported back to the
developers over the network to help us track NAMD’s popularity.

Info: NAMD 2.5b2ss03 for Linux-i686-Clustermatic

Info:
Info: Please visit http://www.ks.uiuc.edu/Research/namd/
Info: and send feedback or bug reports to [email protected]
Info:
Info: Please cite James C. Phillips, Rosemary Braun, Wei Wang, James Gumbart,
Info: Emad Tajkhorshid, Elizabeth Villa, Christophe Chipot, Robert D. Skeel,
Info: Laxmikant Kale, and Klaus Schulten. Scalable molecular dynamics with
Info: NAMD. Journal of Computational Chemistry, 26:1781-1802, 2005.
Info: in all publications reporting results obtained with NAMD.
Info:
Info: Built Fri May 30 13:09:06 CDT 2003 by jim on umbriel
Info: Sending usage information to NAMD developers via UDP.
Info: Sent data is: 1 NAMD 2.5b2ss03 Linux-i686-Clustermatic 47 umbriel jim
Info: Running on 47 processors.

This is followed by a long list of the various configuration options and sum-
maries of information extracted from the parameter and structure files. This is
mostly self-explanatory and just restates the config file options, or the defaults
for options not specified.
NAMD goes through several phases to setup data structures for the actual
simulation, reporting memory usage (for the first process only!). Patches are
the cells that NAMD uses to spread atoms across a parallel simulation, although
the size of the patch grid is independent of the number of processors being used.

Info: Entering startup phase 0 with 6668 kB of memory in use.

Info: Entering startup phase 1 with 6668 kB of memory in use.
Info: Entering startup phase 2 with 10212 kB of memory in use.
Info: Entering startup phase 3 with 10212 kB of memory in use.
Info: PATCH GRID IS 4 (PERIODIC) BY 4 (PERIODIC) BY 4 (PERIODIC)
Info: REMOVING COM VELOCITY -0.0765505 0.00822415 -0.00180344
Info: LARGEST PATCH (31) HAS 408 ATOMS
Info: Entering startup phase 4 with 13540 kB of memory in use.
Info: Entering startup phase 5 with 13540 kB of memory in use.
Info: Entering startup phase 6 with 13540 kB of memory in use.
Info: Entering startup phase 7 with 13540 kB of memory in use.
Info: COULOMB TABLE R-SQUARED SPACING: 0.0625
Info: COULOMB TABLE SIZE: 705 POINTS
Info: NONZERO IMPRECISION IN COULOMB TABLE: 4.03897e-28 (692) 1.00974e-27 (692)
F NAMD STANDARD OUTPUT 115

Info: NONZERO IMPRECISION IN COULOMB TABLE: 2.5411e-21 (701) 5.92923e-21 (701)

Info: Entering startup phase 8 with 13540 kB of memory in use.
Info: Finished startup with 13540 kB of memory in use.

Energies, along with temperatures and pressures, are printed as a single line
with a unique prefix for easy processing with utilities such as grep and awk. An
ETITLE line is printed once for every ten ENERGY lines. The spacing of the
fields, however, is designed to be easy to scan on an 80-column display, as seen
below:

ETITLE: TS BOND ANGLE DIHED IMPRP

ELECT VDW BOUNDARY MISC KINETIC
TOTAL TEMP TOTAL2 TOTAL3 TEMPAVG
PRESSURE GPRESSURE VOLUME PRESSAVG GPRESSAVG

ENERGY: 1000 0.0000 0.0000 0.0000 0.0000

-97022.1848 9595.3175 0.0000 0.0000 14319.5268
-73107.3405 300.2464 -73076.6148 -73084.1411 297.7598
-626.5205 -636.6638 240716.1374 -616.5673 -616.6619

Energy values are in kcal/mol. This example is from a simulation of pure,

rigid water, so the BOND, ANGLE, DIHED, and IMPRP terms are all zero.
BOUNDARY energy is from spherical boundary conditions and harmonic re-
straints, while MISC energy is from external electric fields and various steering
forces. TOTAL is the sum of the various potential energies, and the KINETIC
energy. TOTAL2 uses a slightly different kinetic energy that is better conserved
during equilibration in a constant energy ensemble. TOTAL3 is another varia-
tion with much smaller short-time fluctuations that is also adjusted to have the
same running average as TOTAL2. Defects in constant energy simulations are
much easier to spot in TOTAL3 than in TOTAL or TOTAL2.
Pressure values are in bar. Temperature values are in Kelvin. PRESSURE
is the pressure calculated based on individual atoms, while GPRESSURE in-
corporates hydrogen atoms into the heavier atoms to which they are bonded,
producing smaller fluctuations. The TEMPAVG, PRESSAVG, and GPRES-
SAVG are the average of temperature and pressure values since the previous
ENERGY output; for the first step in the simulation they will be identical to
TEMP, PRESSURE, and GPRESSURE.
To estimate NAMD’s performance for a long simulation, look for the “Bench-
mark time” lines printed in the first several hundred steps. These are a measure
of NAMD’s performance after startup and initial load balancing have been com-
pleted; this number should be very close to the long-time average.

Info: Benchmark time: 47 CPUs 0.0475851 s/step 0.275377 days/ns 13540 kB memory

Periodic performance output, if enabled, shows the CPU and wallclock time
used by the first processor (not the aggregate of all processors in the simulation).
The hours remaining are estimated based on performance since the last TIMING
output.
F NAMD STANDARD OUTPUT 116

TIMING: 1000 CPU: 18.35, 0.01831/step Wall: 50.1581, 0.0499508/step, 6.92374

hours remaining, 14244 kB of memory in use.

Lines such as these are printed whenever file output occurs.

OPENING COORDINATE DCD FILE

WRITING COORDINATES TO DCD FILE AT STEP 1000

It is important to search the NAMD output for Warning and ERROR lines,
which report unusual and possibly erroneous conditions. In the example below,
the pairlist distance is too short or the cycle length is too long, possibly leading
to reduced performance.

Warning: Pairlistdist is too small for 1 patches during timestep 17.

Warning: Pairlists partially disabled; reduced performance likely.
Warning: 20 pairlist warnings since previous energy output.

It is foolish to ignore warnings you do not understand!

G WATER SPHERE TCL SCRIPT 117

G Water Sphere tcl Script

### Script to immerse ubiquitin in a sphere of water just large enough
### to cover it

set molname ubq

mol new ${molname}.psf

mol addfile ${molname}.pdb

### Determine the center of mass of the molecule and store the coordinates
set cen [measure center [atomselect top all] weight mass]
set x1 [lindex $cen 0]
set y1 [lindex $cen 1]
set z1 [lindex $cen 2]
set max 0

### Determine the distance of the farthest atom from the center of mass
foreach atom [[atomselect top all] get index] {
set pos [lindex [[atomselect top "index $atom"] get {x y z}] 0]
set x2 [lindex $pos 0]
set y2 [lindex $pos 1]
set z2 [lindex $pos 2]
set dist [expr pow(($x2-$x1)*($x2-$x1) + ($y2-$y1)*($y2-$y1) +
($z2-$z1)*($z2-$z1),0.5)]
if {$dist > $max} {set max $dist}
}

mol delete top

### Solvate the molecule in a water box with enough padding (15 A).
### One could alternatively align the molecule such that the vector
### from the center of mass to the farthest atom is aligned with an axis,
### and then use no padding
package require solvate
solvate ${molname}.psf ${molname}.pdb -t 15 -o del_water

resetpsf
package require psfgen
mol new del_water.psf
mol addfile del_water.pdb
readpsf del_water.psf
coordpdb del_water.pdb

### Determine which water molecules need to be deleted and use a for loop
### to delete them
set wat [atomselect top "same residue as {water and ((x-$x1)*(x-$x1) +
(y-$y1)*(y-$y1) + (z-$z1)*(z-$z1))<($max*$max)}"]
set del [atomselect top "water and not same residue as {water and
((x-$x1)*(x-$x1) + (y-$y1)*(y-$y1) + (z-$z1)*(z-$z1))<($max*$max)}"]
G WATER SPHERE TCL SCRIPT 118

set seg [$del get segid]

set res [$del get resid]
set name [$del get name]
for {set i 0} {$i < [llength $seg]} {incr i} {
delatom [lindex $seg $i] [lindex $res $i] [lindex $name $i]
}
writepsf ${molname}_ws.psf
writepdb ${molname}_ws.pdb

mol delete top

mol new ${molname}_ws.psf

mol addfile ${molname}_ws.pdb
puts "CENTER OF MASS OF SPHERE IS: [measure center [atomselect top all] weight mass]"
puts "RADIUS OF SPHERE IS: $max"
mol delete top