Transition from all27_prot_lipid to all36_lipid

Hi CHARMM users,

I am currently moving one of my simulations to from all27_prot_lipid and toppar_all27_lipid_cholesterol ff to all36 ff.

I just was wondering if there are some tools to make the transition more convenient.
(except convert_lipid_c27_c36_xopc.str)

So did anyone automate all the required steps?
1) Reading old psf/coor with old ff, renaming alpha chain atoms and create PDB.
2) Split PDB into single segmented files (with grep).
3) generate new psf/coor/pdb file from the separated pdb files with the new ff.

I made the steps via PDB to omit trouble with the "MASS" statement.

Is there a more sleekly way while I am already asking.

Cheers
Bjoern

It is fairly simple if you use CHARMM coord files instead of PDB files; there's no need to split the file. Also, CHARMM can read a multiple segment PDB file, with a little work; there's an example in the Script Archive forum. For the conversion via COOR CARD format:

1) read old psf/coor, rename alpha chain, write new coor file
2) make and write c36 psf; read new coor and calc energy as a check

From that point on, you can simply use the new psf/coor file pair.

Finally, one can always continue to use the old alpha chain naming convention for DOPC and POPC, via the other stream file supplied with the C36 lipid parameters.

*digging through the Script Archive*

Thanks for the hint with the psf/coor files. I new I missed something.

Thanks a lot.

Originally Posted By: rmv

For the conversion via COOR CARD format:

1) read old psf/coor, rename alpha chain, write new coor file
2) make and write c36 psf; read new coor and calc energy as a check

How do I create a PSF from the coor (card) file? I was searching the io.doc but without success.

[EDIT]
Might be not the smartest way, but I just regenerated the setup of my PSF and read the coordinates from the crd (card) file:

Code:

open read card unit 10 name ../top_all36_lipid.rtf
read  rtf card unit 10

open read card unit 20 name ../par_all36_lipid.prm
read para card unit 20

stream ../toppar_all36_lipid_cholesterol.str

read sequ CHL1 43
read sequ POPC 86
read sequ POPS 43
read sequ CHL1 43
read sequ POPC 86
read sequ POPS 43
generate MEMB setup warn first none last none

read sequ TIP3 9232
generate TIP3 noangle nodihe

read sequ POT 88
generate POT noang nodihe

read sequ CLA 2
generate CLA noang nodihe

open read card unit 10 name memb30_c36_step1.crd
read coor card unit 10 resid

coor stat
energy

! WRITE CONVERTED FILES
open write card unit 10 name memb30_c36_step2.pdb
write coor pdb  unit 10

open write card unit 10 name memb30_c36_step2.crd
write coor unit 10 card

open write unit 10 card name memb30_c36_step2.psf
write psf  unit 10 card

open write unit 10 card name memb30_c36_step2.xplor.psf
write psf  xplo unit 10 card

stop

[/EDIT}

Thanks
Bjoern

That's usually what I do, and more or less what I meant by making a new PSF. The issue is that READ SEQU and GENER are needed for each segment; specifying the residue count is often simpler than reading the sequence from a coordinate file. That's only one of several ways to establish a sequence.

Just out of curiosity:
Is it possible to read from and up to a certain "line" for from a crd/(card) file?

I am thinking of multimer proteins where it would be nasty to type each AA sequence ;-)

something like:

Code:

open read card unit 10 name test.crd
read line 1 100 unit 10 resid
generate X1 setup warn first none last none

read line 101 201 unit 10 resid
generate X2 setup warn first none last none

This is not possible in current versions, but a similar (chain/segid) functionality will be available in the next release.
Instead of typing the sequences you can use an editor or shell scripts to separate your multimer file into separate files for each unit. The mmtsb toolset and charmming.or or charmm-gui.org can also do this for you.

rmv, just a note:

in "toppar_all36_lipid_oldxopc.str" you wrote:

Quote:

!Parent files that have to be read prior to streaming this file
!top_all36_lipid.inp
!par_all36_lipid.inp

I guess it should be:

Quote:

!Parent files that have to be read prior to streaming this file
!top_all36_lipid.rtf
!par_all36_lipid.prm

Converting the old psf/coor charmm files to the new naming scheme worked, but it is rather inconvenient to manipulate NAMD pdb files to be accepted by CHARMM, so I decided to stick with the old naming scheme, generate a new xplor.psf for NAMD and then just continue the simulation instead of starting a new simulation from the last coordinates I would have read from the NAMD PDB/COOR file.

When streaming your toppar_all36_lipid_oldxopc.str

Code:

open read card unit 10 name ../../top_all36_lipid.rtf
read  rtf card unit 10

open read card unit 20 name ../../par_all36_lipid.prm
read para card unit 20

stream ../../toppar_all36_lipid_cholesterol.str

stream toppar_all36_lipid_oldxopc.str

I get the following error:

Code:

CHARMM>    stream toppar_all36_lipid_oldxopc.str
 VOPEN> Attempting to open::toppar_all36_lipid_oldxopc.str::
 OPNLGU> Unit 99 opened for READONLY access to toppar_all36_lipid_oldxopc.str

                    INPUT STREAM SWITCHING TO UNIT    99
 RDTITL> * CHARMM36 ALL-HYDROGEN LIPID FORCE FIELD TOPPAR STREAM FILE
 RDTITL> * DOPC, POPC RESIDUES USING OLD C27 ATOM NOMENCLATURE, WHICH
 RDTITL> * HAS AN INCONSISTENT ALPHA CHAIN ATOM NAMING SCHEME
 RDTITL> * REPLACES C36 DOPC, POPC WITH STANDARDIZED NAMING (SAME AS DMPC, DPPC)
 RDTITL> *
 Parameter: IN1 <- "" <empty>
  
 CHARMM>     
  
 CHARMM>    !Parent files that have to be read prior to streaming this file
 CHARMM>    !top_all36_lipid.inp
 CHARMM>    !par_all36_lipid.inp
 CHARMM>     
  
 CHARMM>    read rtf card append
 MAINIO> Residue topology file being read from unit  99.
 RDTITL> * REPLACEMENT DOPC, POPC; OLD C27 ALPHA CHAIN ATOM NAMES
 RDTITL> *
*** WARNING **** residue DOPC already exists (old one deleted)
          RESI DOPC        0.00
 **** ERROR in RTFRDR ****  The atom type code, OSP    is unknown. The atom will be ignored.
          ATOM O1   OSP    -0.57
 **** ERROR in RTFRDR ****  The atom type code, OSP    is unknown. The atom will be ignored.
          ATOM O2   OSP    -0.57
 **** WARNING from RTFRDR **** The total charge of the residue, DOPC,   1.1400000,
 does not equal the expected charge,   0.0000000.
*** WARNING **** residue POPC already exists (old one deleted)
          RESI POPC         0.00
 **** ERROR in RTFRDR ****  The atom type code, OSP    is unknown. The atom will be ignored.
          ATOM O1   OSP    -0.57
 **** ERROR in RTFRDR ****  The atom type code, OSP    is unknown. The atom will be ignored.
          ATOM O2   OSP    -0.57
 **** WARNING from RTFRDR **** The total charge of the residue, POPC,   1.1400000,
 does not equal the expected charge,   0.0000000.
 There were   4 warning(s) from RTFRDR.
 There were    4 error(s) from RTFRDR.
 Execution will be terminated.

Cheers
Bjoern

Originally Posted By: lennart

This is not possible in current versions, but a similar (chain/segid) functionality will be available in the next release.
Instead of typing the sequences you can use an editor or shell scripts to separate your multimer file into separate files for each unit. The mmtsb toolset and charmming.or or charmm-gui.org can also do this for you.

Okay, without some bash/python scripting it does not work, thats what I was curios about. Thanks

With next release you mean c37a1?

Thanks
Bjoern

Did you get the files from the MacKerell web site? I thought we'd fixed that OSP bug. Just change the OSP to OSLP for those 2 phosphate O atoms in the old definitions of POPC and DOPC in the stream file.

The convenience of NAMD usage was not a consideration.

Originally Posted By: rmv

Did you get the files from the MacKerell web site? I thought we'd fixed that OSP bug.

Yes, downloaded it today again cause I accidentally deleted the "conversion scripts" so I am sure it is from the MacKerell web site.

Originally Posted By: rmv

Just change the OSP to OSLP for those 2 phosphate O atoms in the old definitions of POPC and DOPC in the stream file.

Yep, I changed them already, but thought it would be of interest. But since I was tinkering with NAMD I was not shure if it was my fault or really a "BUG"

Originally Posted By: rmv

The convenience of NAMD usage was not a consideration.

Cheers
Bjoern

If general interest to future readers was a concern, this probably should have been posted to the Parameter Set forum.

Originally Posted By: rmv

If general interest to future readers was a concern, this probably should have been posted to the Parameter Set forum.

Right. But I guess you'll fix the downloadable files so it will no longer be an issue and since I ran into this issue within the context of this thread, I didn't think of creating a new thread in the appropriate forum. Sorry about that.

Cheers
Bjoern

To be honest, with a topic of "Transition from all27_prot_lipid to all36_lipid" and an initial post about a script provided with the distributed all36_lipid parameters, I tend to think the whole thread probably belongs in the Parameter Set forum.

I can have it moved ...