LDB problem

LDB: LOAD: AVG 4.34903 MAX 4.4356 MSGS: TOTAL 643 MAXC 13 MAXP 8 Refine

(http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/2669.html)

Those LDB messages are part of every parallel run and are quite normal.
It's just reporting how the load balancer is doing. In this case, your
maximum load is only slightly more than your average load, and there are
171 total messages being sent between processors, which looks pretty good.

The older message from Brain Bennion was a hang during load balancing,
probably due to a compiler bug (a workaround is in 2.6b1).

Align molecule

(http://www.ks.uiuc.edu/Research/namd/tutorial/NCSA2002/hands-on/)
mol delete all

mol load psf pope.psf pdb pope_move.pdb
global popemol
set popemol [molinfo top]
mol load psf protein.psf pdb protein.pdb

display projection orthographic

mol modselect 0 $popemol "name P"
mol modstyle 0 $popemol VDW

puts {
     1. Switch to Move Molecule mode by moving the mouse into the graphics
        window and pressing the 8 key.  The cursor should now have the form
        of a cross-hair.

     2. While holding down the shift key, click on one of the protein atoms
        and drag the mouse.  You should see just the protein rotating
        around the atom you picked.  Try to align the axis of the protein
        with an imaginary line coming out of the screen, so that it's
        perpendicular to the membrane.  If you accidentally rotate the
        whole scene instead of just the protein, type "display resetview"
        in the text console to get back to the original view.

     3. Type "rotate x by -90" in the text console.  Now you want to center
        the protein between the phosphate atoms of the bilayer.  This time,
        click and hold on one of the protein atoms _without_ holding down
        shift.  Move the protein up and down until it's centered.

     4. Switch back to the orginal view with "display resetview".  If the
        protein looks good, you can go on to the next step; otherwise repeat
        steps 2 and 3 above.
}

Color molecule using electrostatic potential

1: install apbs:
download the source code.
./configure --prefix=/use/local
make all
sudo make install

2: install pdb2pqr:(http://pdb2pqr.sourceforge.net/userguide.html#introduction)
(wasn't able to find numpy, so pdb2pka is disabled)
./configure --disable-pdb2pka --prefix=/usr/local
make
sudo make install

3: following the instructions on the tutorial
http://137.189.50.96/kbwong/teaching/pymol/pymol_tutorial.html

a. Use PDB2PQR to convert the PDB format to PQR format
> pdb2pqr.py --ff=amber --apbs-in 1w2i_nowat.pdb pymol.pqr
The PQR file will be output to pymol.pqr.

b. Use psize.py to determine the grid dimensions for APBS calculation
> psize.py pymol.pqr
You should be able to see the following results:

Center = 37.468 x 31.798 x 12.177 A
:
:
Coarse grid dims = 53.011 x 58.568 x 65.807 A
Fine grid dims = 51.183 x 54.452 x 58.710 A
Num. fine grid pts. = 97 x 97 x 97

Take a note on these parameters.

c. Edit the apbs.in
You need to enter the following parameters:

cgcent 37.468 31.798 12.177 # Grid Center
fgcent 37.468 31.798 12.177 # Grid Center

cglen 53.011 58.568 65.807 # coarse mesh lengths (A)
fglen 51.183 54.452 58.710 # fine mesh lengths (A)

dime 97 97 97 # Grid Points

d. Run APBS
> apbs apbs.in
After a while, it will create an electrostatics map called "pymol.dx".

just use vmd to import both .pdb and .dx file. Then choose drawing method to be surf, and coloring method to be volume. Then change the color scale data range in Trajectory bar to be like -8.00 to 8.00.

Beautiful plot appeared.!

blue surface : positively charged
red surface : negatively charged
white surface : neutral

easy way of compiling code?

http://www.deisa.eu/science/benchmarking

wget \
http://www.deisa.eu/science/benchmarking/files/benchmark_NAMD.tar.gz
gunzip benchmark_NAMD.tar.gz
tar xf benchmark_NAMD.tar
#--- Download NAMD source from http://www.ks.uiuc.edu/Research/namd/
mv NAMD_2.6_Source.tar.gz DEISA_BENCH/applications/NAMD/src/
cd DEISA_BENCH/applications/NAMD/
perl ../../bench/jube -debug bench-Cray-XT4-HECToR.xml

Lipid Tail Selection

(From dechandl@ks.uiuc.edu)
"
I just wanted to point out that selections like "lipids" in VMD are really
just pre-defined atomselect macros. You can create your own macros for more complex selections like this. For example, here is my macro for dealing with this for POPE, POPG and POPC - you'll have to change the selections if you have different lipid preferences:

atomselect macro lipidtails {(resname POPE and not name N HN1 HN2 HN3 C12 H12A H12B C11 H11A H11B P O11 O12 O13 O14 C1 HA HB C2 HS O21 O22 C21 C3 HX HY O31 O32
C31) or (resname POPC and not name N C12 C13 C14 H21 H22 H23 H31 H32 H33 H41 H42 H43 C11 H11 H12 C15 H51 H52 P1 O1 O2 O3 O4 C1 HA HB O21 O22 C21 C3 HX HY O31 O32 C31) or
(resname POPG and not name C12 O12 H31 H21 H22 C11 H11 O11 H12 C15 H51 H52 P1 O1 O2 O3 O4 C1 HA HB C2 HS O21 O22 C21 C3 HX HY O31 O32 C31)}

I keep these in my .vmdrc file so that they're defined at startup. Then selections like "lipidtails" will work from then on. This is useful if you know you'll want to re-use a selection often.
"

Take a look at the paper : bilayer.pdf(http://www.ks.uiuc.edu/Training/CaseStudies/pdfs/) to figure out how to choose the lipid head.

VMD movie plugin for MAC

convert command comes from : ImageMagick
http://www.imagemagick.org/script/binary-releases.php
sudo port install imagemagick(from darwin port)

ppmtompeg comes from : Netpbm
http://netpbm.sourceforge.net/
sudo port install netpbm(from darwin port)

problem with the solvation

I have the opposite z direction with Todd, so the box size in z direction should be inversed.

The deletion of lipid didn't work so well. So we should first minimize the system, to diminish the overlap of lipid and protein. Then from this structure, we generate the restrained structure.

write pdb file from dcd trajectory

mol load psf yourprotein.psf dcd yourprotein.dcd

set nf [molinfo top get numframes]

for { set i 0 } {$i < $nf } { incr i } {
set sel [atomselect top protein frame $i]
$sel writepdb $i.pdb
}

/* ============================== */
want to convert NAMD binary out to pdb file?
just wrong 0 step namd simulation. :-)