dPCA not running

[Fulbabu]

I am unable to run dPCA in carm. There is an error,

carma v.1.7____________________________________________________________________

PSF with the EXT extension detected.
Less than two " N    CA   C   " atoms found for segid " A   ". Abort.
 my  .psf file is below, 

[Nicholas M. Glykos]

> PSF with the EXT extension detected.
> Less than two " N CA C " atoms found for segid " A ". Abort.

Firstly, the SEGID in your PSF file is "SYS" and not "A" (as you told
carma).

Secondly, this type of PSF with a header "PSF CHEQ EXT XPLOR" will probably
not work with carma (the format and alignment of columns is new to me).

Try preparing the PSF using VMD as was shown in a previous message
(starting from the .prmtop and .crd files). After you prepared your new
PSF, it should look similar to this :

______________________________________________________________________

PSF

1 !NTITLE
REMARKS VMD-generated NAMD/X-Plor PSF structure file

7400 !NATOM
1 1 VAL N N3 0.057700 14.0100 0
2 1 VAL H1 H 0.227200 1.0080 0
3 1 VAL H2 H 0.227200 1.0080 0
4 1 VAL H3 H 0.227200 1.0080 0

______________________________________________________________________



Then, you edit the PSF and you add SEGIDs to the atoms, so that it will
look like this [notice how the protein has a SEGID of A, which is distinct
for the SEGIDs of ions and water (I and W)].


______________________________________________________________________

PSF

1 !NTITLE
REMARKS VMD-generated NAMD/X-Plor PSF structure file

7400 !NATOM
1 A 1 VAL N N3 0.057700 14.0100 0
2 A 1 VAL H1 H 0.227200 1.0080 0
3 A 1 VAL H2 H 0.227200 1.0080 0
4 A 1 VAL H3 H 0.227200 1.0080 0
....
....
324 A 19 LEU OXT O2 -0.819900 16.0000 0
325 I 20 Cl- Cl- IM -1.000000 35.4500 0
326 I 21 Cl- Cl- IM -1.000000 35.4500 0
327 W 22 WAT O OW -0.834000 16.0000 0
328 W 22 WAT H1 HW 0.417000 1.0080 0
329 W 22 WAT H2 HW 0.417000 1.0080 0
330 W 23 WAT O OW -0.834000 16.0000 0
331 W 23 WAT H1 HW 0.417000 1.0080 0
332 W 23 WAT H2 HW 0.417000 1.0080 0
....
....

______________________________________________________________________







--


Nicholas M. Glykos, Department of Molecular Biology
and Genetics, Democritus University of Thrace, University Campus,
Dragana, 68100 Alexandroupolis, Greece, Tel/Fax (office) +302551030620,
Ext.77620, Tel (lab) +302551030615, https://utopia.duth.gr/glykos/

[Fulbabu]

Thanks for the correcttion.

I have another question, Sir;

I have 20 .mdcdrd (or .dcd) trajectories files, each trajectory contrain 500 frame (Total=20*500=10000 (100 ns)).

I could not understand how to use these 20 .dcd files for calculating dPCA in Carma?

Can I merge all .dcd files into a single .dcd file or there is any other way to use all those files? 

[Nicholas M. Glykos]

> I have 20 .mdcdrd (or .dcd) trajectories files, each trajectory contrain
> 500 frame (Total=20*500=10000 (100 ns)). I could not understand how to
> use these 20 .dcd files for calculating dPCA in Carma? Can I merge all
> .dcd files into a single .dcd file or there is any other way to use all
> those files?

PCA is time-agnostic, you'll get the same results irrespectively of the
order of frames. So, you can merge the trajectory files and feed carma just
one DCD (plus its corresponding PSF). The only issue that can arise is if
these 20 trajectory files are repetitions of the same run and they all
start from the same initial structure. This will erroneously bias the PCA
analysis because the same set of structures will be present multiple times
for reasons that have nothing to do with how stable are these structures.
If the 20 files are simple successive restarts from an otherwise continuous
trajectory, you are OK.

[Fulbabu]

Thank you for the clarification.

On Sunday, February 10, 2019 at 12:51:04 PM UTC+5:30, Fulbabu wrote:

[Fulbabu]

Sir, Again I have problem with dPCA,

--------------------------------------------------

carma -verbose -write -color -segid A -dPCA 5 3 300 md2.dcd ionized.psf

carma v.1.7____________________________________________________________________

Less than two " N    CA   C   " atoms found for segid " A   ". Abort.

 My psf file is below,

On Sunday, February 10, 2019 at 12:51:04 PM UTC+5:30, Fulbabu wrote:

[Nicholas M. Glykos]

> --------------------------------------------------
> carma -verbose -write -color -segid A -dPCA 5 3 300 md2.dcd ionized.psf
>
> carma
> v.1.7____________________________________________________________________
>
> Less than two " N CA C " atoms found for segid " A ". Abort.
> My psf file is below,

When you edited the PSF file, you inserted additional spaces, and so you
translated all atom names to the wrong place (PSF is a fixed format file).

I attach with this message a corrected version of your PSF file (compressed
with bzip2).

[Nicholas M. Glykos]

... and to possibly reduce these problems, next time try running VMD using
the following variation :


vmd -dispdev text test.prmtop test.crd
set solvent [atomselect top solvent]
$solvent set segname W
set protein [atomselect top protein]
$protein set segname A
set ions [atomselect top ions]
$ions set segname I
set all [atomselect top all]
$all writepsf ionized.psf
quit


which will attempt to add SEGIDs automatically. The trouble this approach is
that you'll have to carefully examine what VMD did because (a) special groups
may not be recognized, (b) some ions may not be recognized, etc.

[Nicholas M. Glykos]

> If I want to calculate dPCA for a specific loop in the structure, then how
> to do that? Is any tutorial for that?


You assign a new distinct SEGID to all atoms you want to use (say "G"), and
then you run carma with the 'segid G' flag. Even better : you use grcarma
which allows you to select specific residues on specific chains.




> Another question is my dPCA.ps plots are not continuous, so how set the
> size of the bins for getting a smooth plot?


I assume that you mean that the images are 'pixelated'. This is a
consequence of having a relatively small number of frames. Two solutions :


- The first is to use the corresponding ".dat" files with your favorite
plotting program whichever this may be.


- The second goes like this :

cd <directory where carma files are>
wget http://utopia.duth.gr/glykos/progs/plot
chmod 755 ./plot
./plot -cc < carma.dPCA.DG_01_02.dat

... and then : press the 'p' key to save the plot (file plot.ppm), and then
'q' to quit.

[Nicholas M. Glykos]

> I have the plot but how do I get the color bar (CB) for the plot?

If you want to prepare plots with color bars and everything like, for
example, this one :

http://nefeli.mbg.duth.gr/cgi-bin/wiki.pl/download/variance_covariance_color_map_from_carma_and_gnuplot_sigmoidal_weights_2

then you'll have to use gnuplot (using possibly 'set pm3d' + 'set dgrid3d' ...).

You can of course prepare beautiful plots with R.



> In my carma_dPCA_DG_01_02.dat have 8 columns, I do not understand what that
> means?

https://utopia.duth.gr/glykos/carma.html#section_57

https://utopia.duth.gr/glykos/carma.html#section_80