APPENDICES

 

  1. Format of GOLPE/SIMCA .dat files
  2. Importing data from SYBYL
  3. Importing data from GRID
  4. Importing data from SPARTAN
  5. Importing data from AMBER
  6. Importing data from DISCOVER
  7. Running Fractional Factorial Variable Selection in a Server.

 


Format of GOLPE/SIMCA .dat files

 

GOLPE reads ASCII files in GOLPE/SIMCA format. Usually these files have the .dat extension, but this is not strictly required.

 

The GOLPE/SIMCA format is as follows:

 

Golpedat Name of the data file
1500 Number of variables (X´s plus Y´s)
30 Number of objects (molecules)
1 Sequential number of the object #1
Molecule_1 Name of the object #1
-0.03
-0.04
-0.04
-0.4
-0.5 The 1500 grid field values for
-1.3 the first object
-2.3
-5.0
-3.0
-1.0
-0.8
0.00
0.00
0.10
0.20
1.00
5.00
...
2 Sequential number of the object #2
Molecule_2 Name of the object #2
-0.04
-0.08
-0.10
-0.90
-1.00
...

 


 

Importing Data From SYBYL

 

Importing binary field files (.qab files)

The most convenient way to import data from SYBYL is through .qab files. These are temporary files that contains most of the information required by GOLPE in binary format. To generate and store these files follow these steps:

 

  1. Run SYBYL and open your molecular spreadsheet. Then choose from the MSS menu the option QSAR>>>Partial Least Squares. A dialog window like this appears:

 

 

SYBYL import dialog

 

 

 

 

The .qab file will be created immediately but it will be automatically removed by SYBYL in about 10 minutes. Before this you have plenty of time to copy this file to a different directory or to use GOLPE to generate the GOLPE/SIMCA file as described in the File>>>Import Fields>>>Sybyl (.qab files) section.

The .qab file does not contains the name of the objects nor the coordinates of the grid cage generated. We strongly recommend to obtain the Spreadshet ASCII file and a Region file in SYBYL, for importing simultaneously with the .qab file. The Spreadshet ASCII can be generated in SYBYL in the Molecular Spreadsheet (MSS), menu using the command File>>>Write Report, and then typing the name of the export file. The Region file can be recognised by its .rgn externsion and is automatically generated by SYBYL during the CoMFA analysis.

 

Importing ASCII field files (.efe and .efs files)

The support of this kind of files was discontinued in GOLPE 4.5. Use binary import instead.


 

Importing Data From GRID

 

GOLPE can import directly kont files produced by the program GRID. However, when running GRID, the directive LIST should be set equal to 1, if a single Target (molecule) is being studied or to -2, when working with several Targets (molecules). In either case, these directives force GRID to produce binary kont files, which can be red straight into GOLPE.

 

For further details see the GRID manual at the section "THE GRID:GOLPE INTERFACE".

 


 

Importing Data From SPARTAN

 

GOLPE can import a set of SPARTAN generated ASCII output files, containing data describing MEP fields. The following procedure was tested on SPARTAN 4.1 for UNIX:

Start importing all the structures, already superimposed or superimpose them using the tools included in SPARTAN .

Use the command File>>>Group as... to define a group of molecules on which all the subsequent operations will be performed. Give a sensible name to this group.

Under Setup menu, choose the semiempirical or quantum mechanics computation and then, in the same menu, Select Setup>>>Volumes.

 

SPARTAN Volume dialog

 

In this dialog select the option Global. Choose the property to represent in the Volume control. Then, click on Expert and edit the line. At the end, add writev=fname[ascii] (no space between the filename and "[ascii]"). For example, if you want to represent the electrostatic potential and you have a line like this

volume=elpot x=-4.18~4.57 y=-4.37~4.37 z=-4.32~4.43

add the above text to obtain:

volume=elpot x=-4.18~4.57 y=-4.37~4.37 z=-4.32~4.43 writev=c[ascii]

Click on Replace and then on Save buttons. Submitt the job using Setup>>>Submit.

After a while, within the directory with the name of the group you will find many directories with the names of each molecule. Within each directory there should be a file named "c". Collect these files in the same directory and give them different names. For example, c_mol1.file, c_mol2.file, etc... This set of files can be directly imported into GOLPE using File>>>Import fields>>>SPARTAN command

 

Please notice that SPARTAN assigns a very small grid spacing. We recommend to use always the lower resolution, but even so, the resulting fields can be huge.  


 

Importing Data From AMBER

 

GOLPE can import a set of outa files produced by the module ANAL of AMBER programs, in order to perform a COMBINE analysis, see C.Perez, M.Pastor, A.R.Ortiz, F. Gago, J.Med.Chem. 41, 836, (1998) and references therein for further reference.

In the ANAL comand file, the per-residues splicing of the interaction energies is obtained by entering a -1 in the residues definition. Which follows is an example for a protein of 135 residues.

example.pdb
    1    0    0    0  135    1
    0  0.0  0.0  0.0  0.0
    1    0    1    0   50    0
   15.     2.0    1.2    4.0
    1   20.0  20.0  20.0   20.0  20.0  20.0  20.0  20.0 20.0
ENERGY
Receptor+Ligand
RES -1 135
END
END
STOP
    

 


 

Importing Data From DISCOVER

 

GOLPE can import a set of out files produced by MSI DISCOVER, in order to perform a COMBINE analysis, see C.Perez, M.Pastor, A.R.Ortiz, F. Gago, J.Med.Chem. 41, 836, (1998) and references therein for further reference.

In order to generate the ligand-receptor energies of interaction it is possible to use the DSL command enclose. This command produces the printout of the nonbond energies between the residues in the active site of a enzyme of receptor. Refer to your DISCOVER documentation to obtain more information.

Which follows is an example of a csh macro which runs DISCOVER in stand-alone mode, using DSL commands, to carry out a mild energy minimization and the nonbond energy analysis for a whole set of ligand-receptor complexes.

#!/bin/csh -f
# Basic settings, adapt to your own system configuration 

setenv DISCOVER /usr/msi/972/bin/fdiscover/discover
setenv FIELD /usr/msi/972/irix6m3/biosym_lib/cvff.bin
setenv NICE 20
setenv NUMBER_CPUS 1 

# Name of the compound to analyze, at least the files
# .car .mdf should exist
foreach BASENAME (`cat list.txt`)

# Remove files from previous runs
unalias rm
rm -f $BASENAME.out
rm -f $BASENAME.prm
rm -f $BASENAME.cor
rm -f $BASENAME.log

# Generate input
cat << EOF > $BASENAME.inp
      overlap = 0.01

      begin simulation
     * add-automatic bond torsion valence out-of-plane
      reduce
! 
      set dielectric  = 4.000000*r
! 
      Fixed atom list generation
     * add main
     * molecule 1
!
      tethered atoms list generation
     * add side atoms
     * molecule 1
     * add all
     * molecules 3 to 4
! 
      template force for 500 cycles 
     * using conjugate gradient with
     * no morse functions and no cross terms
     * and a forcing constant of 10.0 kcal/A
     * until the maximum derivative is less than 0.01 kcal/A 
!
      enclose molecule 2 residue LIG 999H with
     * a surface of radius 10.0 A
!
      end

EOF
# This line launches discover

$DISCOVER $BASENAME $FIELD $NICE y $NUMBER_CPUS
end

IMPORTANT: Please notice that the above macro is only an example to illustrate a possible way to obtain the energies required for a COMBINE analysis. The conditions of the minimization are not guaranteed to be optimal nor suitable for your complexes.


 

Running Fractional Factorial Variable Selection in a Server.

 

It is possible to run FFD variable selections in a different computer. This can be advantageous if you have access to powerful SGI servers (i.e. SGI Origin series). This task does not require to install GOLPE in the server and therefore no new license keys need to be generated.

 

Follow these steps:

 

Installing FFD_Selection.

1. Create in your server a directory with exactly the same path as the GOLPE installation directory in your computer.

2. Transfer the file FFD_Selection from the GOLPE installation directory to the same directory in your server.

 

Running FFD_Selection

1. Execute in your computer the F. Factorial selection as usual but selecting execution in an independent window. When the window appears close it, this will stop the local job.

 

2. Transfer to your server the following files:

filename.dat

filename.dat.VarType

filename.dat.Pretreat

filename.dat.VarGroups (only if you are using groups of variables)

DATA.XY

FFD.csh

 

3. Execute in the server the script FFD.csh. If a error message appears check the execution privileges of FFD.csh. Make this file executable with the command:

chmod +x FFD.csh

 

4. When the selection of variables is finished transfer back to your local workstation the files:

filename.dat.EffectOnPred

filename.dat.FFDselList

 

Continue working in the same way as usual.

 

Manuel Pastor, 1999