faq_1.3.htm
FREQUENTLY ASKED QUESTIONS:

(AmberFFC version 1.3, written: May 2003, released: July 2003)




  1) Why AmberFFC ?
  2) On which Operating System (OS) does AmberFFC work ?
  3) What is needed to execute AmberFFC ?
  4) How to use AmberFFC under Windows (Win9x, Win2000 or WinNT) ?
  5) How does AmberFFC work ?
  6) Where to get information about the format of the AMBER, GLYCAM & Accelrys FF files ?
  7) Force field (FF) files obtained under Windows can not be used under UNIX. Why ?
  8) The AMBER software is installed but the AMBER FF files cannot be found. WHY ?
  9) "binff" and "binff3" Accelrys scripts are only launched by AmberFFC under IRIX or AIX. WHY ?
10) When "binff" or "binff3" are executed, errors dealing with SGI "libblas" libraries are displayed. WHY ?
11) When "binff3" is executed, errors dealing with "duplicate version number" are displayed. WHY ?
12) The Weiner et al. FF with "United Atoms" is not converted by AmberFFC. WHY ?
13) Dihedral parameters are not found when Discover (or NMR-Refine) is used. WHY ?
14) The Cys and Met residues are not recognized by AMBER FF converted with AmberFFC. WHY ?
15) When using Discover and amber91/91X (Weiner et al. FF), some FF parameters are missing. WHY ?
16) Using Discover and FF converted by AmberFFC, NEUTRAL termini amino acids are not recognized. WHY ?
17) The LP atom type is missing in amber99. WHY ?
18) How to model small molecules such as organic molecules with FF converted by AmberFFC ?
19) Why to use the USER FRAGMENT library to build glycoconjugates in InsightII ?
20) How to use this new USER FRAGMENT library with a GLYCAM FF ?
21) Where to put the three GLYCAM FF in the file system tree ?
22) When GLYCAM FF are converted to the Accelrys format by AmberFFC, Missing_glycamXX files are generated. Why ?
23) How are the GLYCAM RESIDUE libraries built by AmberFFC ?
24) How are the monosaccharide units differentiated in a glycam_XX.rlb file ?
25) Monosaccharide UNIT names of the GLYCAM FF (AMBER software) are slightly different from the ones of the Accelrys version. Why ?
26) Here are some typos to correct in the GLYCAM ".prep" files
27) Why did some carbon atom names change in the Neu5Ac units ?
28) Why "binff" and "binff3" are not executed by AmberFFC for the GAFF conversion ?
29) Why no topology database is created by AmberFFC for the GAFF conversion ?
30) Why the new ff02 and ff02EP FF are not converted by AmberFFC ?
31) Implementation differences of the FF in the AMBER and Accelrys software. Implication the $BC variable.
32) What is the meaning of the $XTERM variable in the AmberFFC code ?
33) Differences between the AmberFFC versions 1.0, 1.1, 1.11, 1.2 and 1.3
34) How to run AmberFFC (versions 1.0 => 1.3) in a network ?
35) The references of the AMBER and GLYCAM FF
36) Who wrote AmberFFC and how to cite AmberFFC ?



1) Why AmberFFC ?

ANSWER
Among the diFFerent versions of the AMBER force fields (FF) available in the public domain, Accelrys Inc. (formerly "Molecular Simulation" or "Biosym", San Diego, http://www.accelrys.com/) only provides the "old" Weiner et al. FF. This "Accelrys Weiner et al." FF presents the following differences with the original version:

1- Out of plan (oop) FF parameters for the NMe, Tyr and Phe residues,

2- Partial charges for the NH3+ and CO2- termini of all the amino acids,

3- Some amino acids are missing in the topology database:
=> PRON (NH2+Pro), Lys, Glu and Asp (neutral side chain for these three residues),

4- Some FF parameters are different:
=> See the C-CT-N3, HC-CT-N and HC-CT-HC angle parameters, for instance.

=> We decided to re-adapt this old Weiner et al. and to create the Cornell et al. and the Wang et al. AMBER FF for the Accelrys software (AmberFFC version 1.0). Thus, AmberFFC converts the AMBER topology database and the FF parameter files to the Accelrys format. Moreover, a pseudo-atom library (".plb" file) is created for each FF (except for the GAFF FF).

With AmberFFC version 1.2, three GLYCAM FF (Woods et al., http://glycam.ccrc.uga.edu/) are converted for the Accelrys molecular mechanics modules.

With AmberFFC version 1.3, a new version of parm99 (Simmerling et al.) and the new GAFF (General AMBER force field; for organic molecules) FF are also converted to the Accelrys format.

These AMBER & GLYCAM FF are usable with the "Discover" and "Discover_3" Accelrys modules (other names: "FDiscover" & "CDiscover").



2) On which Operating System (OS) does AmberFFC work ?

ANSWER
We tested AmberFFC using LINUX-PC (RedHat 6.1, English version or Mandrake 7.0, French version), SOLARIS 7.0 (SUN), IRIX 6.3/6.5 (SGI), AIX 4.2 (IBM RS/6000) and HP-UX 10.2 (Hewlett-Packard). AmberFFC should work on all UNIX platforms having perl5 installed. We also tested AmberFFC under Windows (for using AmberFFC under Windows, read also faq 4).
We did NOT test AmberFFC under Macintosh (for perl under Macintosh, see http://www.macperl.com)



3) What is needed to execute AmberFFC ?

ANSWER
You need to have:
1- perl5 (not perl4 !) and the "locale" module installed:
perl (Larry Wall) is an interpreted programming language to transform/modify/extract data from ASCII (or text) files. Once perl5 is installed on your system, the "locale" module should be present by default. This module is useful to avoid errors with the decimal point between the French (",") and English (".") systems.
For more information about perl5, have a look at the CPAN site http://www.cpan.org or at the O'Reilly site http://www.perl.com:
- Use the perl Search Engine to find a module http://search.cpan.org/
- Read "What To Do Once You've Downloaded A Module From The CPAN http://www.cpan.org/modules/INSTALL.html if you need to install a module.

We tested AmberFFC with the following perl5 versions: 5.005_03, 5.005_02 and 5.004_04. Use "
perl -V" and "perl -v" to check which perl version you have on your machine.

2- Force field files:
- AMBER FF files:
The AMBER FF belong to the public domain: you can get these files from the AMBER web site (http://amber.scripps.edu/dbase.html).

* To build the Weiner et al. FF (without "United Atoms"): "PLEP" version:
=> parm91.dat, all.in, allnt.in and allct.in files are used.

* To build the Weiner et al. FF (without "United Atoms"): "LEAP" version:
=> parm91X.dat, all.in, allnt.in and allct.in files are used.
(Read http://amber.scripps.edu/Questions/improp.html to understand the diFFerence between "PLEP" and "LEAP" versions).

* To build the Cornell et al. FF (version 1994):
=> parm94.dat, all_amino94.in, all_aminont94.in, all_aminoct94.in and all_nuc94.in files are used.

* To build the first modification of the Cornell et al. FF (version 1996):
=> parm96.dat, all_amino94.in, all_aminont94.in, all_aminoct94.in and all_nuc94.in files are used.

* To build the second modification of the Cornell et al. FF (version 1998):
=> parm98.dat, all_amino94.in, all_aminont94.in, all_aminoct94.in and all_nuc94.in files are used.

* To build the Wang et al. FF (version 1999):
=> parm99.dat, all_amino94.in, all_aminont94.in, all_aminoct94.in and all_nuc94.in files are used.

* To build the Simmerling et al. FF (version 2002):
=> parm99mod.dat, all_amino94.in, all_aminont94.in, all_aminoct94.in and all_nuc94.in files are used.

* To build the GAFF FF (version 2002):
=> gaff.dat is used (no topology database is provided for GAFF).
[If you use the "gaff.dat.1.1" or "gaff.1.2" (gaff.dat.X.Y) file simply rename it into "gaff.dat"]

- GLYCAM FF files:
With AmberFFC version 1.2, three GLYCAM FF are converted. They are all constituted by a "glycam_XX.dat" file (FF parameter file) and a set of numerous ".prep" files (database of monosaccharide and miscellaneous units with topology rules).

- Accelrys FF files:
=> cvff.rlb, amber.rlb and cvffa.plb files are simply used as "format donor" input files.

These files belong to Accelrys, so you need to have a Discover/Discover_3 (FDiscover/CDiscover) license (see http://www.accelrys.com/contact/index.html). As AmberFFC generates AMBER and GLYCAM FF for Discover/Discover_3, I would not understand why you would need AmberFFC without having a Discover/Discover_3 license...



4) How to use AmberFFC under Windows (Win9x, Win2000 or WinNT) ?

ANSWER
1- Install Active perl for Windows, a perl5 package available for Win32 platforms.
See http://www.ActiveState.com/ActivePerl to download and install it.

2- Run AmberFFC using the "non-automatic way" or "Second way" (see also
faq 5).

3- To run AmberFFC under Windows, simply run "
perl AmberFFC" in a "dos" window/terminal.
Remark:
- The colors are not recognized under winNT 4.0: We decided to deactivate the colors for all the Windows OS (win9X, win2000 et winNT).
- As Accelrys software does not run under Windows OS (see also
faq 9), you will have to transfer by ftp your newly obtained FF files to a SGI or an IBM RS/6000 workstation (see also faq 7 for more information).

However, using AmberFFC under Windows is only interesting to read and check the FF files but you cannot use them !



5) How does AmberFFC work ?

ANSWER
You do not need to compile the script. Just launch "perl AmberFFC" or "perl5 AmberFFC" (if you have perl4 and perl5 installed on your workstation) in a UNIX terminal or X-window terminal. You can run AmberFFC using two ways:

1- "First way": automatic way or default way
=> AMBER and Accelrys software have to be installed on the same workstation.
We tested AmberFFC on IRIX 6.3 (perl5 was installed from the SGI freeware site: http://freeware.sgi.com as only perl4 is provided by default in IRIX 6.3) and on IRIX 6.5.

AmberFFC looks for:
- The $BIOSYM_LIBRARY variable in your environment to find the three Accelrys FF input files needed.
- The AMBER directory in your $PATH variable to find all the AMBER and GLYCAM FF files.

Just run AmberFFC and answer to the question to select which AMBER FF you want to build: Here is what you get once you have executed "
perl AmberFFC" (version 1.3):
____________________________________________________________________________________

Which force field do you want to build for F&CDiscover ? (X to exit)

(A)  Amber   91  (PLEP)
(B)  Amber   91X (LEAP)
(C)  Amber   94
(D)  Amber   96
(E)  Amber   98
(F)  Amber   99
(G)  Amber   99mod
(H)  Gaff    2002
(I)  Glycam  93
(J)  Glycam  2000
(K)  Glycam  2010

(L)  All


Your choice:
___________________________________________________________________________________

2- "Second way":
If Accelrys and AMBER software are not installed you can get AmberFFC FF files, anyway. Just modify the beginning of the "MAIN PROGRAM" section: You have to specify where the Accelrys, AMBER and GLYCAM FF files are located in the file system tree. Here is an example we used:

Line 1451 in the code:
__________________________________________________________________________________________________________

# If you don't have the Accelrys and/or AMBER software and you'd like to launch "AmberFFC", create the following tree
# with the corresponding Accelrys, AMBER and GLYCAM FF files:
#
# /usr/people/guest/FF_Files/
#   - accelrys   cvffa.plb, cvff.rlb, amber.rlb   (Accelrys FF files).
#
#   - dat   parm91.dat, parm91X.dat, parm94.dat, parm96.dat, parm98.dat, parm99.dat, parm99mod.dat, gaff.dat
#           all.in, allct.in, allnt.in,
#           all_amino94.in, all_aminoct94.in, all_aminont94.in, all_nuc94.in   (AMBER FF files).
#
#           - glycam/glycam_93       ".prep" files with a FF parameter file   (GLYCAM FF files)
#           - glycam/glycam_2000     ".prep" files with a FF parameter file   (GLYCAM FF files)
#           - glycam/glycam_2010     ".prep" files with a FF parameter file   (GLYCAM FF files)
#
# Next, follow the instructions which are below.

if ($ENV{'BIOSYM_LIBRARY'}) {
     $BL="$ENV{'BIOSYM_LIBRARY'}";
     if (substr($BL,length($BL)-1,1) ne "/") { $BL=$BL."/"; }
}
else {
#   PrintColor("RED","Are you sure InsightII is installed ?\n");
#   PrintColor("RED","The \$BIOSYM_LIBRARY variable is not in your environment !\n\n");
#   PrintColor("RED","Read the \"faq\" file for more information.\n");
#   if (!$ENV{"Windir"}) { ‹STDIN›; }
#   exit(-1);
    $BL="/usr/people/guest/FF_Files/accelrys/";

}

@temp=split(':',$ENV{'PATH'});

$found=$i=0;
while (!$found and $i<=$#temp) {
     if (uc($temp[$i])=~/AMBER/) {
         $posi=index(uc($temp[$i]),"AMBER");
         if (($posi=index($temp[$i],"/",$posi))!=-1) { $AMBER=substr($temp[$i],0,$posi); }
         else { $AMBER=$temp[$i]; }
         $AMBER=$AMBER."/";
         $found=1;
     }
     $i++;
}

if (!$found) {
#   PrintColor("RED","Are you sure Amber is installed ?\n");
#   PrintColor("RED","The Amber directory is not in your PATH !\n");
#   PrintColor("RED","Read the \"faq\" file for more information.\n");
#   if (!$ENV{"Windir"}) { ‹STDIN›; }
#   exit(-1);
    $AMBER="/usr/people/guest/FF_Files/";
}
__________________________________________________________________________________________________________

3- "Third way":
The Accelrys package is installed but you do not have the AMBER software. So, get the AMBER and GLYCAM FF files, and modify AmberFFC as it is described below:
__________________________________________________________________________________________________________

if ($ENV{'BIOSYM_LIBRARY'}) {
     $BL="$ENV{'BIOSYM_LIBRARY'}";
     if (substr($BL,length($BL)-1,1) ne "/") { $BL=$BL."/"; }
}
else {
     PrintColor("RED","Are you sure InsightII is installed ?\n");
     PrintColor("RED","The \$BIOSYM_LIBRARY variable is not in your environment !\n\n");
     PrintColor("RED","Read the \"faq\" file for more information.\n");
     if (!$ENV{"Windir"}) { ‹STDIN›; }
     exit(-1);
     #$BL="Full_Path";

}

@temp=split(':',$ENV{'PATH'});

$found=$i=0;
while (!$found and $i<=$#temp) {
     if (uc($temp[$i])=~/AMBER/) {
         $posi=index(uc($temp[$i]),"AMBER");
         if (($posi=index($temp[$i],"/",$posi))!=-1) { $AMBER=substr($temp[$i],0,$posi); }
         else { $AMBER=$temp[$i]; }
         $AMBER=$AMBER."/";
         $found=1;
     }
     $i++;
}

if (!$found) {
#   PrintColor("RED","Are you sure Amber is installed ?\n");
#   PrintColor("RED","The Amber directory is not in your PATH !\n");
#   PrintColor("RED","Read the \"faq\" file for more information.\n");
#   if (!$ENV{"Windir"}) { ‹STDIN›; }
#   exit(-1);
    $AMBER="/usr/people/guest/FF_Files/";
}
__________________________________________________________________________________________________________



6) Where to get information about the format of the AMBER, GLYCAM & Accelrys FF files ?

ANSWER
1- Get information about the AMBER FF:
http://amber.scripps.edu/eqn.txt
http://www.accelrys.com/doc/life/insight2K/FFbs/2_Forcefields.html#568080

2- Get information about the format of the AMBER files:
=> GLYCAM FF have the same format than the AMBER FF
http://amber.scripps.edu/dbase.html

3- Get information about the format of the Accelrys FF files:
- For the ".frc" file format: the FF parameter file
http://www.accelrys.com/doc/life/insight2K/cdiscover/D_Files.html#925808
- For the ".rlb" file format: the topology database file or residue library file
http://www.accelrys.com/doc/life/insight2K/formats980/File_Formats_1998.html#144453
- For the ".plb" file format: the pseudo-atom library
http://www.accelrys.com/doc/life/insight2K/formats980/File_Formats_1998.html#779673
- For the "template.dat" file:
http://www.accelrys.com/doc/life/insight2K/insight/a_PTR.html#309226

4- Get information about the GLYCAM FF files:
Ask to Dr R. J. Woods. Email: rwoods@ccrc.uga.edu
Here is the official GLYCAM web site: http://glycam.ccrc.uga.edu/index.html



7) Force field (FF) files obtained under Windows can not be used under UNIX. Why ?

ANSWER
You have some problems with the CR/LF character (Carriage Return / Line Feed). Windows uses the CR/LF characters (ASCII codes: 13 and 10) at the end of each line (Unix uses only the LF one). We wrote a CRLF function in AmberFFC that you could probably adapt for this purpose or you can use the following perl5 script: We named it "DOS_UNIX" and we use it by running "perl DOS_UNIX files_to_be_transformed".
__________________________________________________________________________________________________________

sub CRLF {
     open(FILE,$_[0]) or return(0);
     if (!open(FILE2,"> dos_unix.tmp")) { close FILE; return(0); }
     while (defined($line=‹FILE›)) {
         $line=~s/\r\n/\n/g;
         print FILE2 $line;
     }
     close FILE;
     close FILE2;
     unlink($_[0]);
     rename("dos_unix.tmp",$_[0]);
     return(1);
}

if ($ARGV[0] eq "") { print "You have to specify at least one file !\n"; exit(-1); }

$NbErrors=0;     $NbFiles=0;
for($i=0;$i<=$#ARGV;$i++) {
     if (substr($ARGV[$i],length($ARGV[$i])-1,1) ne "*") {
         $NbFiles++;
         if (!(CRLF($ARGV[$i]))) { $NbErrors++; }
     }
     else {
         $j=-1;     $k=0;
         while ($k < length($ARGV[$i])) {
             if (substr($ARGV[$i],$k,1) eq "/") {$j = $k}
             $k++;
         }
         if ($j!=-1) { $rep=substr($ARGV[$i],0,$j+1); }
         else { $rep="./"; }
         if ($j!=length($ARGV[$i])-2) { $file=substr($ARGV[$i],$j+1,length($ARGV[$i])-$j-2); }
         else { $file=""; }
         if (!opendir(REP,$rep)) { $NbErrors++; }
         else {
             while ($fichier=readdir(REP)) {
                 if (-f "$rep$fichier") {
                     if (!$file or $fichier=~/^$file/) {
                         $NbFiles++;
                         CRLF("$rep$fichier");
                     }
                 }
             }
             close REP;
         }
     }
}

print $NbFiles-$NbErrors." file(s) has/have been modified.\n";
__________________________________________________________________________________________________________



8) The AMBER software is installed but the AMBER FF files cannot be found. WHY ?

ANSWER
We tested AmberFFC with AMBER5, AMBER6 and AMBER7. For instance, the place of the "parm91X.dat" file is different in these three versions (parm91X.dat is in the ~/amber5/Leap/lib directory, in the ~/amber6/dat directory and in the ~/amber7/dat/leap/parm directory). AmberFFC only looks for the AMBER FF files in these three former directories. As we do not have AMBER4, we do not know where the FF files in this version are. If some AMBER FF files are not found by AmberFFC, copy them in the amberX/dat directory (or make a link).


9) "binff" and "binff3" Accelrys scripts are only launched by AmberFFC under IRIX or AIX. WHY ?

ANSWER
Accelrys products mainly work on SGI and IBM RS/6000 workstations. Thus, running "binff"/"binff3" Accelrys scripts under LINUX/SOLARIS/HP-UX or Windows Operating System (OS) would have no interest. AmberFFC detects which OS is installed on your workstation: If AmberFFC finds that your OS is either IRIX or AIX, it automatically runs "binff" and "binff3" and build the two Accelrys binary FF files (".bin" and ".xfrc" files of the AMBER and GLYCAM FF). On the contrary, if it finds another OS ONLY the ASCII or text files (".frc," ".rlb", ".plb" and empty_template.dat) are generated. About Accelrys and hardware, see http://www.accelrys.com/doc/life/insight2K/sysguide/requirements.html#911584

In fact, using AmberFFC under LINUX, SOLARIS, HP-UX or Windows OS (OS different than AIX or IRIX) is only interesting for reading/checking the FF files.



10) When "binff" or "binff3" are executed, errors dealing with SGI "libblas" libraries are displayed. WHY ?

For instance, here are two typical errors:

iris 1% binff amber99
Creating new file amber91.bin
10424:binffe: rld: Fatal Error: Cannot Successfully map soname 'libblas_mp.so'
under any of the filenames /usr/msi/License_Pack/irix6m3/libs/libblas_mp.so:
/usr/msi/980/irix6m3/biosymso/libblas_mp.so:/usr/lib32/libblas_mp.so:
/usr/lib32/internal/libblas_mp.so:/lib32/libblas_mp.so:/opt/lib32/libblas_mp.so:
/usr/msi/License_Pack/irix6m3/libs/libblas_mp.so.1:/usr/msi/980/irix6m3/biosymso/libblas_mp.so.1:
/usr/lib32/libblas_mp.so.1:/usr/lib32/internal/libblas_mp.so.1:/lib32/libblas_mp.so.1:/opt/lib32/libblas_mp.so.1:

iris 2% binff3 amber99
10430:discovery: rld: Fatal Error: Cannot Successfully map soname 'libblas.so'
under any of the filenames /usr/msi/License_Pack/irix6m3/libs/libblas.so:
/usr/msi/980/irix6m3/biosymso/libblas.so:/usr/lib32/libblas.so:
/usr/lib32/internal/libblas.so:/lib32/libblas.so:/opt/lib32/libblas.so:
/usr/msi/License_Pack/irix6m3/libs/libblas.so.1:
/usr/msi/980/irix6m3/biosymso/libblas.so.1:/usr/lib32/libblas.so.1:
/usr/lib32/internal/libblas.so.1:/lib32/libblas.so.1:/opt/lib32/libblas.so.1:

ANSWER
Just install the SGI "libblas" libraries. I think they are NOT installed by default (at least, on O2 SGI workstations). You can find these libraries by using the following commands:
cd /CDROM/dist/dist6.X      and
showfiles -f . | grep libblas

These libraries are in the SGI "IRIX development foundation" CDROM for IRIX 6.3 or in the "IRIX Foundation 1" CDROM for IRIX 6.5 (I think you will also need the IRIX Overlay [2 of 2] CDROM). The following subsystems: "ftn_eoe.sw.libblas" and "ftn_eoe.sw64.libblas" need to be installed.


11) When "binff3" is executed, errors dealing with "duplicate version number" are displayed. WHY ?

Here are typical errors found when "binff3" is executed:

Error: Forcefield: Encountered the error - duplicate version number, on line #172
  1.0    1      OW     HW     0.9572     553.0000
      invoked from within
"readFile forcefield file = amber99"

ANSWER
A FF parameter is present in duplicate in the amberXX.frc file. Just remove one of these two lines. This problem occurs only with "binff3" and not with "binff". (This problem is encountered within the "parm99.dat" and "gaff.dat" files, for instance)


12) The Weiner et al. FF with "United Atoms" is not converted by AmberFFC. WHY ?

ANSWER
The Weiner et al. FF uses either "United Atoms" or "Non-hydrogen containing carbons". As these FF versions start to be somewhat old, we decided to convert ONLY the Weiner et al. FF with "Non-hydrogen containing carbons" to the Accelrys format. But I think we could easily modify AmberFFC to allow to get the Weiner et al. FF with "United Atoms".


13) Dihedral parameters are not found when Discover (or NMR-Refine) is used. WHY ?

Here are typical errors found in Discover ouputs:

The following table summarizes the number of interactions for which parameters were either found explicitly in the force field file, automatically generated from bond order information, or set to zero if not available. Each of the major classes of interactions is reported separately. A detailed listing of all the interactions requiring automatically generated parameters has been printed in a separate file. The default name of the file is "fname".prm (logical unit 1).  +---------------------------------------------------------------------------------+
 |                            BOND      VALENCE    TORSION    OUT-OF-   ANGLE*ANGLE|
 |                           LENGTHS     ANGLES     ANGLES    PLANES    CROSS TERMS|
 |                                                                                 |
 | Available explicitly          393        704          0        66          0    |
 | Assigned automatically          0          0          0         0          0    |
 | Set to zero                     0          0       1037         0          0    |
 +---------------------------------------------------------------------------------+
This run must stop because of unfound parameters.


ANSWER
For instance, in the AMBER Cornell et al. and Wang et al. FF, the dihedral potential is:

> E = SUM(n=1,4) { (Vn/m) * [ 1 + cos(n*Phi - Phi0(n)) ] }
> m = multiplicity or total number of torsions centered on the same bond.

"Discover" (FDiscover) can NOT handle Vn with n = 4 but only Vn with n = 3: you have to use "Discover_3" (CDiscover).
As "NMR-Refine" uses Discover, FF with V4 dihedral terms can not be used with NMR-Refine. If you want to do Simulated Annealing studies with AmberFFC FF you have to use a Discover_3 script. (see $BIOSYM/gifts/discover/btcl)



14) The Cys and Met residues are not recognized by the AMBER FF converted with AmberFFC. WHY ?

Here are typical errors:

Command:
(Executing: Potentials Forcefield Fix -Print_Potentials Fix -Print_Part_Chargs Fix -Print_Form_Chargs TEST)
Using template file 'AmberFFC/empty_templates.dat' for potential assignment for residue TEST:5
Could not assign potential type to TEST:5:N
Could not assign potential type to TEST:5:CA
Could not assign potential type to TEST:5:HN
Could not assign potential type to TEST:5:HA
Could not assign potential type to TEST:5:C
Could not assign potential type to TEST:5:O
Could not assign potential type to TEST:5:CB
Could not assign potential type to TEST:5:HB1
Could not assign potential type to TEST:5:HB2
Could not assign potential type to TEST:5:SG
Could not assign potential type to TEST:5:LG1
Could not assign potential type to TEST:5:LG2
Could not assign potential type to TEST:5:HG3
Using template file 'AmberFFC/empty_templates.dat' for potential assignment for residue TEST:13
Could not assign potential type to TEST:13:N
Could not assign potential type to TEST:13:CA
Could not assign potential type to TEST:13:HN
Could not assign potential type to TEST:13:HA
Could not assign potential type to TEST:13:C
Could not assign potential type to TEST:13:O
Could not assign potential type to TEST:13:CB
Could not assign potential type to TEST:13:HB1
Could not assign potential type to TEST:13:HB2
Could not assign potential type to TEST:13:CG
Could not assign potential type to TEST:13:HG1
Could not assign potential type to TEST:13:HG2
Could not assign potential type to TEST:13:SD
Could not assign potential type to TEST:13:LD1
Could not assign potential type to TEST:13:LD2
Could not assign potential type to TEST:13:CE
Could not assign potential type to TEST:13:HE1
Could not assign potential type to TEST:13:HE2
Could not assign potential type to TEST:13:HE3

ANSWER
You have to add sulfur lone pairs (LP) for the Cys and Met residues for the Weiner et al. FF. On the contrary, remove these sulfur lone pairs for the Cornell et al. type FF. To add them, selected the sulfur atoms and use the following InsightII command:

(Executing: Hydrogens TEST Lone_Pairs Charged -Assign_IUPAC_Names)



15) When using Discover and amber91/91X (Weiner et al. FF), some FF parameters are missing. WHY ?

Here are typical errors found in Discover ouputs:

Error: Energy expression: cannot find parameters for angle: CB(CT)-CG(C)-OD2(OH)
Error: Energy expression: cannot find parameters for angle: OD1(O)-CG(C)-OD2(OH)
or
Error: Energy expression: cannot find parameters for angle: CG(CT)-CD(C)-OE2(OH)
Error: Energy expression: cannot find parameters for angle: OE1(O)-CD(C)-OE2(OH)
or
Error: Energy expression: cannot find parameters for bond: CE(CT)-NZ(NT)

ANSWER
Some FF parameters are missing for three amino acids: Asp, Glu and Lys (neutral side chain for these three amino acids) in the Weiner et al. FF:
You have two possibilities:
- Take the missing FF parameters from the Cornell et al. FF and add them in the Weiner et al. FF file.
- Use the Cornell et al. FF, for instance (a newer FF).



16) Using Discover and FF converted by AmberFFC, NEUTRAL termini amino acids are not recognized. WHY ?

Here are typical errors:

(Executing: Potentials Forcefield Fix -Print_Potentials Fix -Print_Part_Chargs Fix -Print_Form_Chargs TEST)
Using template file '/usr/people/guest/AmberFFC-1/empty_templates.dat' for potential assignment for residue TEST:1
Could not assign potential type to TEST:1:N
Could not assign potential type to TEST:1:HN1
Could not assign potential type to TEST:1:HN2
Could not assign potential type to TEST:1:CA
Could not assign potential type to TEST:1:HA
Could not assign potential type to TEST:1:C
Could not assign potential type to TEST:1:O
Could not assign potential type to TEST:1:CB
Could not assign potential type to TEST:1:HB1
Could not assign potential type to TEST:1:HB2
Could not assign potential type to TEST:1:HB3
Warning: total formal charge does not equal total partial charge.


ANSWER
With InsightII, you can select which termini you want for your peptide/protein:
- Either neutral (NH2) or charged (NH2+) amino group
- Either neutral (CO2H) or charged (CO2-) acid group
In AMBER FF, ONLY the charged termini are present in the amino acid topology database.
One exception: The terminus of the neutral acid group (in fact, the OH group) is present in the Weiner et al. amino acid topology database. You have two possibilities:
- Use charged termini for your two terminal amino acids or
- Use another Accelrys FF: the CVFF FF, for instance.



17) The LP atom type is missing in amber99. WHY ?

ANSWER
The "Can_Write" AmberFFC function (lines 409-429 in the code) is used to remove unwanted/unused atom types in the AMBER (and GLYCAM) FF parameter files. Thus, all the united atoms are removed from the amber91 and amber91X versions and the LP atom type is suppressed in amber99. If you want to put it back, just modify the line 414 in the code.


18) How to model small molecules such as organic molecules with FF converted by AmberFFC ?

ANSWER
AmberFFC transforms the AMBER and GLYCAM FF to the Accelrys format. If you want to use these FF to model some other molecules than sugars, peptides or oligonucleotides, you could build by yourself an amberXX_templates.dat or glycam_XX_templates.dat file. The role of the templates.dat file is to set up the atom types (or potentials for Accelrys) for the FF if the studied molecule does NOT have a residue library table. You can read the amber_templates.dat or cvFF_templates.dat files (in the $BIOSYM_LIBRARY directory) as examples and read the following web site to get the rules for setting up the full file: http://www.accelrys.com/doc/life/insight2K/insight/a_PTR.html#309226. You could also build your own residue library (".rlb" file) for your molecule.

The GAFF FF has been created for this particular purpose! (see also
faq 28 and faq 29). Thus, the simpliest way might be to create the topology information needed for the studied molecule in the "gaff.rlb" file; see http://www.accelrys.com/doc/life/insight2K/formats980/File_Formats_1998.html#144453).


19) Why to use the USER FRAGMENT library to build glycoconjugates in InsightII ?

ANSWER
1- In Accelrys software package, a relation exists between the FRAGMENT libraries (used for building molecules such as oligopeptides or oligonucleotides with the "Biopolymer" module) and the RESIDUE library belonging to a FF (for example, with the CVFF force field, the "cvff.rlb" file is used to match the residue names, the atom names, the partial charges and atom types of the studied molecule). Indeed, the atom names, residue names, rules of linkage between different units are set up in the RESIDUE and FRAGMENT libraries in a way allowing a perfect fit between these two libraries. Thus, an oligopeptide built with the "Biopolymer" module will be automatically recognized by the CVFF force field.
2- The AMBER FF topology database of the amino-acids and nucleotides ("all_amino94.in", "all_aminont94.in", "all_aminoct94.in" and "all_nuc94.in" files, for the Cornell et al. FF, for instance) shares similar rules with the corresponding Accelrys FF database (RESIDUE libraries provided by Accelrys, "cvff.rlb" and "amber.rlb" files, for instance).
3- Moreover, AmberFFC converts FF for Accelrys molecular mechanics modules keeping the rules found in the AMBER software (FF parameters, topology, linkage etc...) with some minor adaptations.
=> Thus peptides (or oligonucleotides), built using the Accelrys amino-acids FRAGMENT library and "Biopolymer", are compatible with the AMBER FF converted by AmberFFC.

Unfortunately, the monosaccharide topology rules (in particular, the type of linkage between two sugars) selected by Woods et al. in their FF are different than the ones found in the Accelrys sugar FRAGMENT library. Thus, we decided to build a new USER FRAGMENT library that is fully compatible with the Woods et al. FF (the "frags.def" and "FF_frag_user.psv" files are provided with the AmberFFC code; to get information about the "frags.deg" file, see http://www.accelrys.com/doc/life/insight2K/insight/a_FRAG.html#681795; you might also look at http://www.accelrys.com/doc/life/insight2K/formats980/File_Formats_1998.html#144453 to understand the relation between the ".rlb", ".psv" and "frags.def" files). Thus, the three GLYCAM FF converted by AmberFFC (and in particular the glycam_XX.rlb files; XX= 93, 2000 or 2010) automatically recognize oligosaccharides and/or glycopeptides built using this new USER FRAGMENT library. Moreover, new sugar units which are not in the Accelrys sugar FRAGMENT library like the Abequose, the L-Fucose and the Sialic acid are now available...



20) How to use this new USER FRAGMENT library with a GLYCAM FF ?

ANSWER
See the http://www.accelrys.com/support/life/insight/index.html Accelrys web site for better information.
"How can I create a new carbohydrate fragment? ",
"How can I add DNA residues to the fragment library?" and
"How do I add my own fragments to Insight II's fragment libraries?"

Anyway, here is an example of how to use the "GLYCAM_93" FF with the new USER FRAGMENT library provided with AmberFFC:

The login used in this example is: /usr/people/guest

iris 1% mkdir MyFrag
iris 2% cp $BIOSYM/data/insight/frags.def /usr/people/guest/MyFrag
iris 3% cp $BIOSYM/data/insight/FF_*.psv /usr/people/guest/MyFrag

Save the old version of the USER FRAGMENT library provided by Accelrys
iris 4% cd MyFrag;
iris 5% cp frags.def frags.def.orig
iris 6% cp FF_frag_user.psv FF_frag_user.psv.orig


Copy the NEW "frags.def" and "FF_frag_user.psv" provided with AmberFFC in the "MyFrag" directory:
iris 7% cp /usr/people/guest/AmberFFC/lib/frags.def /usr/people/guest/MyFrag
iris 8% cp /usr/people/guest/AmberFFC/lib/FF_frag_user.psv /usr/people/guest/MyFrag


Change the permissions of the "MyFrag" files:
iris 9% chmod 666 /usr/people/guest/MyFrag/*.psv
iris 10% chmod 666 /usr/people/guest/MyFrag/frags.def


For "tcsh" or "csh" shells: edit $HOME/.cshrc and add in the ".cshrc" file
setenv INSIGHT_DATA /usr/people/guest/MyFrag
=> save and quit
iris 11% source $HOME/.cshrc

At this point, you are ready to use the NEW USER FRAGMENT LIBRARY in InsightII:
This library is compatible with the "glycam93.rlb" file (RESIDUE LIBRARY, ".rlb" file) obtained using AmberFFC. If you use the DEFAULT sugar FRAGMENT LIBRARY provided in InsightII, you will NOT be able to use the GLYCAM_93 FF converted by AmberFFC.

To display this new library, use the "Biopolymer/fragment/get" command from the InsightII pulldowns:
In the FRAGMENT_WINDOW, click on the "Fragment Libraries..." button. Toggle "off" the Common Fragments and toggle "on" the User libraries => "OK"
=> The 46 SUGAR units present in the three GLYCAM FF should be displayed.

Important point: the bonds in PURPLE color MUST BE USED to link units together !
They have to disappear in the final glycopeptide or oligosaccharide (after building the global structure) otherwise the Accelrys GLYCAM_93 FF will not recognize the final structure.


Let's take an example: To build a tetrasaccharide "OLIGO":

FA0 -> 6                               L_Fuc-a(1,6)
GA0 -> 4 MBT -> AGL         or         D_Glc-a(1,4)-D_Man-b-OMe
GB0 -> 3                               D_Glc-b(1,3)

Just run InsightII in a "non-graphical" way using the "Oligo.log" file from a X-terminal:
iris 12% echo source Oligo.log | insightII -at > Accelrys.out

Oligo.log file:
# Insight II Version 2000 - Molecular Modeling System
m:Biopolymer
# The new USER FRAGMENT LIBRARY is used
m:Get Fragment To_Modify_Bond MBT OLIGO
m:Bond Create Fragment_Window Single OLIGO:1:H12 AGL$:1:HO -Assimilate
r:Bond Create Fragment_Window Single OLIGO:1:H6 FA0$:1:H12 -Assimilate
r:Bond Create Fragment_Window Single OLIGO:1:HO4 GA0$:1:H11 -Assimilate
r:Bond Create Fragment_Window Single OLIGO:1:HO3 GB0$:1:H12 -Assimilate
# The GLYCAM_93 force field is in the "AmberFFC" directory
m:Select Forcefield Clear_Potentials -Clear_Charges /usr/people/guest/AmberFFC/glycam_93.frc -Make_Copy
m:Potentials Forcefield Fix Print_Potentials Fix -Print_Part_Chargs Fix -Print_Form_Chargs OLIGO
# Atom types and charges are saved in the .car & .mdf files
m:Put Molecule Biosym OLIGO oligo -Transformed -Displayed -PBC_File
m:Quit Confirm

And look at the "Accelrys.out" file obtained, you should have:
Potentials for residue 1 taken from residue library
Potentials for residue 1B taken from residue library
Potentials for residue 1C taken from residue library
Potentials for residue 1D taken from residue library
Potentials for residue 1E taken from residue library
Original total charge on molecule was 0.000
New total charge on molecule is 0.000

=> the GLYCAM_93 FF recognizes the tetrasaccharide OLIGO and the oligo.car and oligo.mdf files are saved with the right atom types and charges.
=> Check the FF parameters you need for your oligosaccharide:
Indeed, some FF parameters are missing in the glycam_93.frc file. You could take the missing ones from an AMBER FF such as "amber91X". A way to check which FF parameters are missing is to use the "force_field/tabulate" command in InsightII (be aware that there are a lot of BUGS in this InsightII command !)

(Executing: Tabulate Forcefield Molecule OLIGO Output_Internals Output_Non_Bonds Missing_Only TEST)

In particular, do not forget to complete the "#hbond_definition section" in each glycam_XX.frc (XX = 93, 2000 or 2010) file as the three GLYCAM FF still use the hydrogen_bond interaction in the AMBER potential. If you modify the ".frc" file, do not forget to re-run "binff" and "binff3" to update your binary FF files.


21) Where to put the three GLYCAM FF in the file system tree ?

ANSWER
AmberFFC has been written to automatically convert GLYCAM FF if they are in the amber/dat directory, in the following subdirectories:
~/amberX/dat/glycam/glycam_93           X = 5, 6 or 7 for "amber5", "amber6" or "amber7", for instance.
~/amberX/dat/glycam/glycam_2000
~/amberX/dat/glycam/glycam_2010

If the AMBER software is installed on your workstation, just add the GLYCAM FF in the file system tree as it is described above. If you do not have the AMBER software, just specify where are the three GLYCAM subdirectories in the AmberFFC code and use the "non-automatic" way (or second way) to run AmberFFC (see the
faq 5 for more information).



22) When GLYCAM FF are converted to the Accelrys format by AmberFFC, Missing_glycamXX files are generated. Why ?

ANSWER
Woods et al. built monosaccharide units for three different GLYCAM FF. Unfortunately, some units where developed for only one of the three GLYCAM FF and are missing in the two other ones. For instance, the "3lna.prep" file is present in the "GLYCAM_2010" FF but is missing in the "GLYCAM_93" and "GLYCAM_2000 " ones. Thus, we decided to write AmberFFC so that, if an unit has been already developed for one of the three GLYCAM FF, AmberFFC is able to convert this unit for the other ones. In fact, all the units present in the USER FRAGMENT library provided with AmberFFC, can be "potentially" converted by AmberFFC for the three GLYCAM FF. Each unit (for instance, "2ma.prep") has very important similarities in the GLYCAM FF. They only differ by the atom type and the charge columns (second and last ones in the ".prep" files, respectively). Thus, if a known unit (i.e. present in the new USER FRAGMENT library and already found in at least one of the three GLYCAM FF) is developed afterwards by Woods et al. or by a GLYCAM user for another GLYCAM FF, everybody will be able to use it in the Accelrys software once it has been converted by AmberFFC. If units not existing in the new user fragment library show up on the web (i. e. totally unknown unit, for instance "4la.prep"), this user fragment library will be updated and AmberFFC should be able to convert this totally unknown unit by incorporating very simple modifications in its code.

As some sugar units were developed for only one or two GLYCAM FF, it means that some oligosaccharides will not be recognized by the three GLYCAM FF. For instance, if we take the tetrasaccharide "OLIGO" (see also
faq 20), it will NOT be recognized (today, The June, 1st, 2001) by the "GLYCAM_2010" because the GA0 and GB0 units are still not developed for this FF. However, new units should pretty soon appear for this last FF.


23) How are the GLYCAM RESIDUE libraries built by AmberFFC ?

ANSWER
To convert the amberXX.rlb files, AmberFFC uses Accelrys "cvff.rlb" and "amber.rlb" files as format donor. These two files are provided in the Accelrys software package. As no sugar RESIDUE library is available in the Accelrys software package, we decided to build GLYCAM RESIDUE libraries using a different strategy, i. e. without Accelrys format donor files. glycam_XX.rlb files (XX = 93, 2000 or 2010) were simply converted directly from the GLYCAM ".prep" files provided in each FF. For instance, the atom name column is the number two in the GLYCAM ".prep" files and the number one in the Accelrys ".rlb" files. Similar differences between the two software also exist for the atom types and partial charges. Some adaptations were also done (atom name, improper torsion, etc ...).


24) How are the monosaccharide units differentiated in a glycam_XX.rlb file ?

ANSWER
InsightII diFFerentiates UNITS in a RESIDUE library mainly based on the number of atoms and on the atom names found in this unit. The Glc-a, Glc-b, Man-a and Man-b monosaccharides have the same number of atoms and should "normally" have the same atom names. If such rules are applied, InsightII is NOT able to differentiate these four sugars. Thus, we decided to differentiate each monosaccharide using different atom name for the endocyclic oxygen present in all the sugars. We decided, to increment the endocyclic atom name "ORX" (X = positive integer) based on the alphabetical order the ".prep" files appear when the "ls -l" UNIX shell command is used. Thus, a unit present or absent in the RESIDUE FRAGMENT library always has the same name "ORX" for its endocyclic oxygen, the name that is found in the USER FRAGMENT library, determined in alphabetical order.

Here are some examples of atom names "ORX" used:


".prep" file names                        Endocyclic oxygen name
AMBER software version          AMBER software version     Accelrys software version
(Woods et al.)                 (Woods et al.)             (Converted by AmberFFC)
23ma.prep                        OR                         OR1
26ma.prep                        OR                         OR2
2ga.prep                         OR                         OR3
2lb.prep                         OR                         OR4
                                            etc...
2xa.prep                         OR                         OR6
2xb.prep                         OR                         OR7
                                            etc...
34nn.prep                        OR                         OR9
34no.prep                        OR                         OR10
                                            etc...
sa.prep                          OR                         OR44
xa.prep                          OR                         OR45
xb.prep                          OR                         OR46

It means that there are 46 sugar units in the USER fragment library provided with AmberFFC.


25) Monosaccharide unit names of the GLYCAM FF (AMBER software) are slightly different from the ones of the Accelrys version. Why ?

ANSWER
The monosaccharide unit names used by Woods et al. (AMBER software version) generally start by a number. Such rule is not accepted in the Accelrys fragment libraries. Thus, we changed the name of each unit based on its ".prep" file name and highly similar unit name were kept in the Accelrys version of the GLYCAM FF:
(i) If a number is found in the first position of a ".prep" file name (AMBER software version), this number is moved to the last position (for instance, "23ma.prep" becomes "MA23").
(ii) if no number is found in the file name, the "zero" number is added at the end of the name (for instance, "fa.prep" becomes "FA0")

Here are some examples:

".prep" file names                     Unit names in the GLYCAM force fields
AMBER software version          AMBER software version     Accelrys software version
(Woods et al.)                 (Woods et al.)             (Converted by AmberFFC)
23ma.prep                        23M                        MA23
346mb.prep                       TMB                        MBT
46nn.prep                        N46                        NN46
4ga.prep                         4GA                        GA4
4sa.prep                         4SA                        SA4
aa.prep                          AA                         AA0
lna.prep                         LNA                        LNA0
                                             etc...
oh.prep                          OH                         OH0
ome.prep                         OME                        AGL

"ca.prep" and "na.prep" units were voluntary not converted.



26) Here are some typos to correct in the GLYCAM ".prep" files

ANSWER
GLYCAM_93 FF version:
1- "fb.prep" should be renamed into "fa.prep". In this file, lines 3 and 4:

B-(AXIAL!)-L-FUCO- .....          =>          A-(AXIAL)-L-FUCO-
fb.dat                            =>          fa.dat

(the "fb.prep" name is wrong, so if you do not change it into "fa.prep", AmberFFC will not be able to convert this unit !) As the configuration of the Fucose is "L", the "a" configuration is equatorial (for a classical 4C1 chair conformation). However, as the L-Fucose is a 6-deoxy-L-sugar, the 4C1 chair conformation is less stable than the 4C1 one. Thus, the L-Fucose takes the 4C1 chair conformation and its "a" bond becomes axial and not equatorial.

2- "46no.prep"
In this file, line 5:

46N                               =>           46O

(This last typo might be corrected for use with the AMBER software; As AmberFFC does not use this line 5, this typo is not important for the Accelrys software)


27) Why did some carbon atom names change in the Neu5Ac units ?

ANSWER
The atom names of the two first carbons (and of the atoms linked to these two carbons) were changed in the Neu5Ac units ("4sa.prep" and "sa.prep" GLYCAM files) as it follows:

GLYCAM atom names          Accelrys atom names
                          (new USER FRAGMENT and RESIDUE libraries)
 C1                         C2
 O1A                        O2A
 O1B                        O2B
 C2                         C1 => used in glycosidic linkage
 H2                         H1

In the new Accelrys USER FRAGMENT and RESIDUE libraries we built, each glycosidic linkage between two monosaccharides is realized between the C1 atom of the first sugar and an oxygen (position 2, 3, 4 or 6, for instance) of the second sugar. In Sialic acid (Neu5Ac) units, the C2 atom is the one that is implied in the glycosidic linkage. In order, to keep homogeneous rules between the diFFerent sugars (Glc, Gal, Man and Neu5Ac) which allow compatibility and recognition between the USER FRAGMENT and RESIDUE libraries, the atom name of the C1 and C2 atoms were interchanged for the "SA4" and "SA0" Accelrys units.


28) Why "binff" and "binff3" are not executed by AmberFFC for the GAFF conversion ?

ANSWER
When "binff" and "binff3" are executed, two errors might be obtained:

Creating new file gaff.bin
 ERROR: TOO MANY PARAMETERS FOR CURRENT DIMENSIONS.
 ONE OR MORE OF THE FOLLOWING DIMENSIONS MUST BE INCREASED
 REQUIRED LIMIT CURRENT LIMIT
            471          1500
           1845          1500
              0          5000
             35          1500

             49          8000

-1- This problem (only related to "binff") is only encountered with the first GAFF version [i. e. with the "gaff.dat" file and not with the newer versions ("gaff.dat.1.1", "gaff.dat.1.2" or ... "gaff.dat.X.Y" files)]. It seems that the number of FF paramters is limited in the Accelrys .frc file... Thus, one should either decrease this number in the "gaff.frc" file (by removing by hand the ones that are not needed, for instance) or use the gaff.dat.1.1 or gaff.dat.1.2 versions to generate the gaff.frc Accelrys file (Do not forget to rename the "gaff.dat.X.Y" file into "gaff.dat" before executing AmberFFC).

-2- When "binff3" is executed, numerous errors dealing with "duplicate version number" are displayed (See also faq 11).
Check carefully the FF parameters which appear in duplicate in the GAFF FF to select the one you want/need to use!

Because of these two problems, we decided to not automatically execute "binff" and "binff3" when the GAFF FF is converted. Thus, one should solve by hand these two problems and then, manually launch "binff" and "binff3", i. e. by executing:

   binff gaff
   binff3 gaff




29) Why no topology database is created by AmberFFC for the GAFF conversion ?

ANSWER
Simply because AMBER does not provide such database. Thus, one have to create by hand the topology information needed for the molecule, one is interested in (by building the corresponding "gaff.rlb" file; see http://www.accelrys.com/doc/life/insight2K/formats980/File_Formats_1998.html#144453)



30) Why the new "ff02" and "ff02EP" FF are not converted by AmberFFC ?

ANSWER
They are not and will not be converted because Discover (FDiscover and CDiscover) does NOT handle polarization. Indeed, "ff02" and "ff02EP" FF are polarizabled FF with "gaz-phase" charges, meaning that polarisation as to be taken into account in the FF [corresponding to the keyword "ipol = 1" in "sander" (See http://amber.scripps.edu/dbase.html)]. On the contrary, the Cornell et al. type FF present charges leading to an exaggerated dipole moment, meaning that implicit polarization is used (i.e. the amount of polarization that would be expected in aqueous solution) and that polarization does not need to be taken into account (corresponding to "ipol = 0" in sander).



31) Implementation differences of the FF in the AMBER and Accelrys software. Implication the $BC variable.

ANSWER
$BC = "rbc"; residue based cutoff (default variable, see line 1514 in the code),
$BC = "gbc"; group based cutoff.

A fundamental difference in the implementation of FF between AMBER and Accelrys programs is in relation to the handling of non-bonded distance cutoff. As calculating all non-bond interactions is computationally expensive for large molecules, molecular mechanics programs use cutoffs to neglect the non-bonded interactions for pairs of atoms separated by distances greater than a certain value. Such cutoff distances are software dependent: AMBER uses residue-based cutoffs (see http://amber.scripps.edu/doc6/index.html, or the "AMBER-sh-10.2.html" file in the AMBER6 documentation), while Accelrys molecular mechanics modules use either atom or group based cutoffs (see http://www.accelrys.com/doc/life/insight2K/ffbs/3_Setup.html#1112957). The divergent treatment of the non-bonded distance cutoffs may lead to some electrostatic energy differences in the two software. Thus, each user should be aware of this problem before using the converted AMBER FF in the Accelrys software.
If $BC = "gbc" is used in an AmberFFC run, each unit (nucleotide or amino-acid for instance) found in the Accelrys residue library file is constituted of several "atom groups" (several "switching atoms" are selected in the same unit) leading to group based cutoff simulation. On the contrary, if $BC = "rbc" each unit contains a unique "atom group" (a unique "switching atom" is present in the middle of the unit) leading to residue based cutoff calculations. Replacing several switching atoms by a single one in the middle of the residue leads to a "residue-like" based cutoff treatment in Accelrys molecular mechanics modules and may allow to get more "AMBER-like" electrostatic energy values. However, using "no cutoff" simulation, might be the best way to get identical non-bonded energy values in the two software.



32) What is the meaning of the $XTERM variable in the AmberFFC code ?

ANSWER
$XTERM = "on" or "off" ("on" is the default variable, see line 1415 in the code)

If $XTERM = "on", the AmberFFC program is launched in a new X-terminal window allowing to automatically set up the size, the font and the colors of this new terminal X. Moreover, it allows to impeach using AmberFFC as "superuser" (It is never a good think to run programs as "root").
If $XTERM = "off", AmberFFC is launched in the same terminal than the one where "perl AmberFFC" is ran. Using $XTERM = "off", might be interesting for running AmberFFC in a Network; particularly if a firewall is present between the two machines and X-window sessions are not authorized.



33) Differences between the AmberFFC versions 1.0, 1.1, 1.11, 1.2 and 1.3.

ANSWER
See also the "HISTORY" file.
1- AmberFFC Version 1.1:
- It is NOT anymore possible to run AmberFFC as "superuser" (root),
- AmberFFC is executed in a new X-terminal window allowing to set up automatically the working window (only for UNIX Operating System (OS), not for Microsoft (MS) OS), see the lines 1415-1445 of the code. We only tested "dtterm" (for AIX, HP-UX and SUNOS), "xterm" (for LINUX) and "winterm" (for IRIX). For other UNIX OS, "xterm" is used by default (see line 1440) and you might have to replace "xterm" by another X-terminal, more appropriate to your personal case. Moreover, if you have perl 4.0 and perl 5.0 installed on your workstation, the "perl" binary probably corresponds to perl 4.0 and the "perl5" one to perl 5.0. In this case, you will have to replace the word "perl" by "perl5" in the line corresponding to your UNIX OS (see lines 1423-1424 for example).

2- AmberFFC Version 1.11:
- Bug n1 is corrected (see the KNOWN BUGS file),
- Bug n2 is corrected (see the KNOWN BUGS file).
- FDiscover can use the amber91 and amber91X FF as these 2 versions only use a Vn (n = 1 to 3 ) dihedral table, and more generally all the FF that do not use V4 dihedral terms (such as GLYCAM_93).

3- AmberFFC Version 1.2:
- A looooot of improvements in the AmberFFC code ($BC, and $XTERM variables, re-writing of the ".frc" section allowing a better flexibility, etc...)
- Conversion to the Accelrys format of three GLYCAM FF.

4- AmberFFC Version 1.3:
- Some improvements in the AmberFFC code...
- Conversion to the Accelrys format of the Simmerling et al. and GAFF FF.



34) How to run AmberFFC (versions 1.0 => 1.3) in a network ?

ANSWER
1- From an UNIX client on an UNIX server:
- For AmberFFC version 1.0, it is very easy by "telnet".
- For AmberFFC version 1.1-1.3: Do "xhost IP_address" (or "xhost +") in the X-terminal of the client BEFORE doing "telnet" and it should be OK. Indeed, an X-window access needs to be allowed. It should be enough if you do NOT have a firewall between your two UNIX machines (see also
faq 32).

2- From a Windows OS (winNT, win9X, win2K) client on an UNIX server
- For AmberFFC version 1.0, it is very easy by "telnet".
- For AmberFFC version 1.1-1.3: A good way would be to install an X-window EMULATOR under your Windows machine (for instance see Exceed: http://www.dcsi.com/Hummingbird/)



35) The references of the AMBER and GLYCAM FF

ANSWER
Weiner et al. FF:
- Weiner, S. J.; Kollman, P. A.; Case, D. A; Singh, U. C.; Ghio, C.; Alagona, G.; Profeta, S., Jr and Weiner P. A New Force Field for Molecular Mechanical Simulation of Nucleic Acids and Proteins. J. Am. Chem. Soc. 1984, 106, 765-784.
- Weiner, S. J.; Kollman, P. A.; Nguyen, D. T. and Case, D. A. An All Atom Force Field for Simulations of Proteins and Nucleic Acids. J. Comput. Chem. 1986, 7, 230-252.

Cornell et al. FF:
- Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Mertz, K. M., Jr.; Ferguson, D. M.; Spellmeyer, C. D.; Fox, T.; Caldwell, J. W. and Kollman, P. A. A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids, and Organic Molecules. J. Am. Chem. Soc. 1995, 117, 5179-5197.
- Kollman, P.; Dixon, R.; Cornell, W.; Fox, T.; Chipot C. and Pohorille, A. The development/application of a 'minimalist' organic/biochemical molecular mechanic force field using a combination of ab initio calculations and experimental data. In Computer Simulation of Biomolecular Systems, Wilkinson, P. Weiner, W. Van Gunsteren, Ed., Elsevier, 1997, Vol. 3 A, p 83-96.
- Cheatham, T.E. III; Cieplak, P. and Kollman, P. A. A modified version of the Cornell et al. force field with improved sugar pucker phases and helical repeat. J. Biomol. Struct. Dyn. 1999 16, 845-861.

Wang et al. FF:
- Wang, J.; Cieplak, P. and Kollman, P. A. How Well Does a Retrained Electrostatic Potential (RESP) Model Perform in Calculating the Conformational Energies of Organic and Biological Molecules? J. Comput. Chem. 2000, 21, 1049-1074.

Simmerling et al. FF:
- Simmerling, C.; Strockbine, B. and Roitberg, A. E. All-Atom Structure Prediction and Folding Simulations of a Stable protein. J. Am. Chem. Soc. 2002, 124, 11258-11259.

GLYCAM FF:
- Woods, R. J.; Dwek, R.A.; Edge C.J. and B. Fraser-Reid, B. Molecular Mechanical and Molecular Dynamical Simulations of Glycoproteins and Oligosaccharides. 1. GLYCAM_93 Parameter Development. J. Phys. Chem. 1995, 99, 3832-3846.

- Pathiaseril, A. and Woods, R. J. Relative Energies of Binding for Antibody-Carbohydrate-Antigen Complexes Computed from free-Energy Simulations. J. Am. Chem. Soc. 2000, 122, 331-338.



36) Who wrote AmberFFC and how to cite AmberFFC ?

ANSWER
- The Weiner, Cornell, Kollman, Cheatham and Wang et al. Force Fields were converted by:
A. Dejoux (1), P. Cieplak (2), G. Moyna (3) & F.-Y. Dupradeau (1)*

- The GLYCAM Force Fields were converted by:
N. Hannick (1), P. Cieplak (2), G. Moyna (3) & F.-Y. Dupradeau (1)

- The Simmerling et al. and GAFF Force Fields were converted by:
F.-Y. Dupradeau (1), P. Cieplak (2) & G. Moyna (3)

  (1) Faculte de Pharmacie, Amiens, France
  (2) Dept. Chem., University of Warsaw, Poland
  (3) Dept. Chem. & Biochem., University of Sciences in Philadelphia, USA


Please, cite the following reference when using AmberFFC:
A. Dejoux, P. Cieplak, N. Hannick, G. Moyna & F.-Y. Dupradeau, AmberFFC, a flexible program to convert AMBER and GLYCAM force fields for use with commercial molecular modeling packages, J. Mol. Model. 2001, 7, 422-432.


Your suggestions and questions are welcome.

F.-Y. Dupradeau      * email: fyd@q4md-forcefieldtools.org



Last update of this AmberFFC page: March, 19th, 2007