From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Thu Dec 03 2009 - 02:57:24 CST

On Wed, Dec 2, 2009 at 9:03 PM, BIN ZHANG <zhngbn_at_gmail.com> wrote:
> Dear francesco:
>
> I think the procedure to build a rbcg model should be basically the same as
> an atomistic one.
> 1 : obtain the coordinates for your membrane, or protein. This can be done
> by coarse graining all-atom model using cgtools plugin.
> 2 : solvate the system using cgsolvate.tcl
> 3 : add ions to the system.
> About your "cgsolvate.tcl" problem, I think the reason u get the error:
> conflicting versions provided for package "solvate": 1.3, then 1.2.
> is not because the conflict with VMD 1.8.7, but you already type "package
> require solvate" before the command "source ${script_dir}/cgsolvate.tcl".
> Try to open a new fresh VMD console, and type "source
> ${script_dir}/cgsolvate.tcl".

Dear Bin:

Unless I misunderstood your suggestions, Anton's cgsolvate.tcl
requires version 1.2 of "solvate", while VMD 1.8.7 has version 1.3. It
seems to me that (as Anton suggested) the only chance to use that
cgsolvate.tcl is installing an older version of VMD (see below).

Please check what I did in order to rule out any mistake from my side:

(1) Launch vmd 1.8.7

(2) Load psf and pdb of my protein

(3) Exten ... Tkcons ...source ${working_directory}cgsolvate.tcl
conflicting versions provided for package "solvate": 1.3; then 1.2

I had to place cgtools in the same dir as the psf pdb of my molecule.
If it in another dir (specified as you indicated in the .tcl):

can't read "/home/francesco/scripts_param/vmd/Anton" no such variable.
Typing "package require solvate" before sourcing cgsolvate.tcl: same
error.

Looking at Anton's cgsolvate.tcl:

if { $num != 0 } {
    puts "Found $num water atoms near the solute! Please report this bug to"
    puts "vmd_at_ks.uiuc.edu, including, if possible, your psf and pdb file."
    error "Solvate 1.2 failed."
  }
  puts "Solvate 1.2 completed successfully."
  return [list $min $max]
}

is it possible to simply replace "Solvate 1.2" with "Solvate 1.3"? I
guess NO, otherwise Anton would have already suggested that.

Please notice that I am using VMD version 1.8.7 for amd64; this means
(as relevant or unrelevant as it might be) that (under Debian Linux
amd64) there are two tcl directories

/usr/local/lib/vmd/plugins/LINUXAMD64/tcl/autopsf1.2 etc.

/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0

/usr/local/lib/vmd/plugins/noarch/tcl/solvate1.3

I have recently upgraded to vmd 1.8.7, saving previous installed files
of vmd 1.8.6 by renaming "local" "original_local):

/usr/original_local/lib/vmd/plugins/noarch/tcl/solvate1.2
/usr/original_local/lib/vmd/plugins/noarch/tcl/solvate1.2/pkgIndex.tcl
/usr/original_local/lib/vmd/plugins/noarch/tcl/solvate1.2/solvate.tcl
/usr/original_local/lib/vmd/plugins/noarch/tcl/solvate1.2/wat.pdb
/usr/original_local/lib/vmd/plugins/noarch/tcl/solvate1.2/wat.psf
/usr/original_local/lib/vmd/plugins/noarch/tcl/solvate1.2/wat.top

/usr/original_local/lib/vmd/plugins/noarch/tcl/cgtools1.0
/usr/original_local/lib/vmd/plugins/noarch/tcl/cgtools1.0/cggui.tcl
/usr/original_local/lib/vmd/plugins/noarch/tcl/cgtools1.0/cgnetworking.tcl
/usr/original_local/lib/vmd/plugins/noarch/tcl/cgtools1.0/cgtools.tcl
/usr/original_local/lib/vmd/plugins/noarch/tcl/cgtools1.0/pkgIndex.tcl
/usr/original_local/lib/vmd/plugins/noarch/tcl/cgtools1.0/protein.cgc
/usr/original_local/lib/vmd/plugins/noarch/tcl/cgtools1.0/water.cgc

Comparing the above content of cgtools1.0 (vmd 1.8.6) with that for vmd 1.8.7

/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/cggui.tcl
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/cgnetworking.tcl
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/cgtools.tcl
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/pkgIndex.tcl
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/protein.cgc
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/rbcg-2007.par
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/rbcg-2007.top
/usr/local/lib/vmd/plugins/noarch/tcl/cgtools1.0/water.cgc

do you think there is any chance to carry out the whole job with VMD
1.8.6 (or can partly 1.8.6, partly 1.8.7 be used)? Clearly, the ideal
would be to update Anton's scripts (cgsolvate.tcl and others if
needed) to VMD 1.8.7. I would do that if I were a capable developer,
which, unfortunately, I am not. Is anyone needing those scripts
willing to do that? I can't imagine how much time it will take.

thanks

francesco

> But you do need to change the following lines in the cgsolvate.tcl file to
> provide correct directory:
>   154   readpsf /Projects/anton/ReducedModels/CG/dat/CGSOLV/cgwat.psf
>   155   coordpdb /Projects/anton/ReducedModels/CG/dat/CGSOLV/cgwat.pdb
>   194   topology /home/amyshih/CGSOLV/water.top
> Hope this helps,
> Bin
> On Dec 2, 2009, at 10:13 AM, Francesco Pietra wrote:
>
> Hi;
> After I posted the subject problem as secondary matter, I carried out
> an accurate search on the web, unable to find an equilibrated rbcg
> lipid bilayer for vmd 1.8.7/namd martini implementation (or a plain
> procedure to get it from scratch). As for sbcg, I guess that direct
> conversion from an equilibrated all-atoms lipid bilayer will not work.
> Thanks for info.
> francesco pietra
>
>