

    PPPPPPP      EEEEEEE    RRRRRRR       CCCCCC     HH     HH
    PP    PP     EE         RR    RR     CC    CC    HH     HH
    PP    PP     EE         RR    RR    CC           HH     HH
    PPPPPPP      EEEEEE     RRRRRRR     CC           HHHHHHHHH
    PP           EE         RR  RR      CC           HH     HH
    PP           EE         RR   RR      CC    CC    HH     HH
    PP           EEEEEEE    RR    RR      CCCCCC     HH     HH

                                       
                                PEak reseaRCH
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
           An integrated software for analysis of NMR spectra on PC
                                       
                                       
                                       
                                       
                                      by
                    Reino Laatikainen and Matthias Niemitz
                     Kuopio University NMR Research Group
                           Department of Chemistry
                      University of Kuopio, P.O.BOX 1627
                               SF-70211 Kuopio
                                   FINLAND










                          Version January 15th, 1994
          1994 PERCH Project, University of Kuopio, Kuopio, Finland.
                             All rights reserved.
CONTENT

1.   INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   3

2.   SETUP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   5

3.   THE FLOW OF THE ANALYSIS. . . . . . . . . . . . . . . . . . . . . . . .   6
     Step 1: Open the case . . . . . . . . . . . . . . . . . . . . . . . . .   6
     Step 2: Import and manipulate . . . . . . . . . . . . . . . . . . . . .   6
     Step 3: Create input files. . . . . . . . . . . . . . . . . . . . . . .   6
     Step 4: Simulate the spectrum . . . . . . . . . . . . . . . . . . . . .   7
     Step 5: Assign spectral lines . . . . . . . . . . . . . . . . . . . . .   7
     Step 6: Optimize the spectral parameters. . . . . . . . . . . . . . . .   7
     Step 7: Compare the observed and calculated spectra . . . . . . . . . .   7

4.   GUIDED TOURS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   8
     4.1  Peak-top-fitting . . . . . . . . . . . . . . . . . . . . . . . . .   8
     4.2  Integral-Transform fitting . . . . . . . . . . . . . . . . . . . .   9

5.   EXAMPLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  10
     Example 1:     A basic course . . . . . . . . . . . . . . . . . . . . .  10
     Example 2:     Test another coupling sign combination.. . . . . . . . .  12
     Example 3:     Packing and base-line correction of a 13C-
                    coupled 1H NMR spectrum of benzene . . . . . . . . . . .  13
     Example 4:     A basic course using the IT approach . . . . . . . . . .  14
     STRATEGIES for the IT-analysis: . . . . . . . . . . . . . . . . . . . .  15
     Example 5:     A graduate course. . . . . . . . . . . . . . . . . . . .  16
     Example 6.     TLS-fitting (deconvolution) of a peptide
                    NH-region. . . . . . . . . . . . . . . . . . . . . . . .  17

APPENDIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  18
     A1.  Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  18
     A2.  The PERCH files created for a case . . . . . . . . . . . . . . . .  20
     A3.  Program PERCH. . . . . . . . . . . . . . . . . . . . . . . . . . .  21
     A4.  Program IMP. . . . . . . . . . . . . . . . . . . . . . . . . . . .  21
     A5.  Program APP. . . . . . . . . . . . . . . . . . . . . . . . . . . .  22
     A6.  Program PAC. . . . . . . . . . . . . . . . . . . . . . . . . . . .  22
     A7.  Program PIC. . . . . . . . . . . . . . . . . . . . . . . . . . . .  24
     A8.  Program TLS. . . . . . . . . . . . . . . . . . . . . . . . . . . .  25
     A9.  Program LS . . . . . . . . . . . . . . . . . . . . . . . . . . . .  26
     A10. Program DP . . . . . . . . . . . . . . . . . . . . . . . . . . . .  27

Information in this document is subject to change without notice. No
part of this document may be reproduced or transmitted in any form or
by any means without permission by the authors. 

Microsoft, MS, MS-DOS are registered trademarks, Windows, FORTRAN
Powerstation and Visual Basic are trademarks of Microsoft Corporation.
IBM is registered trademark of International Business Machines
Corporation. PKZIP and PKUNZIP are trademarks of PKWARE Incorporation.
1.   INTRODUCTION

PERCH (PEak reseaRCH) is an integrated NMR-software package for
research
and education. The spectral analysis and simulation up to 10 spins is
based on the program MLDC (see Laatikainen, R., J.Magn. Reson., 92, 1
(1991); Laatikainen, R., QCMP100. MLDC8: Automated Analysis of NMR
Spectra, QCPE Bull., 12, 23 (1992)). The package provides some
powerful tools for preparing the input data for MLDC and for
examination of 1D-NMR  spectra in general, like:

-    accurate peak-picking and peak-top interpolation
-    line shape analysis and simulation
-    multi-term baseline correction
-    deconvolution with total-line-shape fitting
-    graphical assignment procedure for calculated and observed
     spectra
-    effective packing utility
-    several iterative modes, including the traditional LAOCOON-type,
     peak-top-fitting for accurate analysis in presence of
     line-overlap and a total-line-shape mode (no assignments
     required)

PERCH programs take advantage of 32-bit code, are running on
IBM-compatible PC and can be accessed from a graphical user
interface under Windows 3.1 or command driven directly under
MS-DOS. This manual mainly refers to the usage of the package
under MS-DOS. The PERCH shell for Windows 3.1 provides an
online help which covers most of this manual except the
chapters 4 and 5.

Regardless under which environment you are going to use PERCH
you should run the examples in chapter 5 first.

Hardware requirements:

IBM-compatible PC with an 80386 processor or higher, a hard
disk, a 3 1/2" floppy drive, a VGA or compatible display and
at least 4 MB RAM. A disk cache very well improves the
performance and the memory demands under Windows should be
fulfilled by installing a swap file.

Software requirements:

MS-DOS 5.0 or later. Windows 3.1 when using the PERCH shell
(WPERCH.EXE). The commands DAT, INF, PMS, SAV and TOP need the
MS-DOS 5.0 Editor (EDIT.COM) in the path. The PERCH shell is
using Windows NOTEPAD.EXE as default editor. PERCH programs
need the DOS-extender DOSXMSF.EXE to be in the path. The setup
program installs DOSXMSF.EXE in the same directory as PERCH.
The present version contains the following features
(programs):

     PERCH     Opens a case.
     IMP       IMPorts the spectra into the PERCH format.
     PAC       PACks, manipulates and displays the observed
               spectrum.
     PIC       PICks up the peak-top frequencies and
               intensities.
     TLS       Total-Line-Shape analysis (deconvolution).
     MLD       starts the spectral calculation with MLDC.
     PER       starts the spectral calculation with PERCHIT.
     DOR       starts the double resonance simulation with
               DORES.           
     LS        simulates the Line Shape.
     DP        DisPlays computed and/or observed spectra on
               the screen. DP contains also an assignment
               procedure for the observed lines.
     L [type]  prints out and displays files of [type]-type.
               [type] can be CAL, DAT, DOC,INF, OBS, OUT, PAR,
               PMS, SAV or TOP. L DOC or L MLDOC print out the
               documents files; for the last PERCH procedure
               and MLDOC for the last MLD calculation.
     D [type]  empties files of [type]-type. See L.
     BEST      Upgrades the parameter file [name].PMS file
               using the  optimized spectral  parameters in
               [name].SAV.
     ASC       converts the binary spectral files  into ASCII.
               
     APP       APPends spectra.

Some batch-files are provided to customize the usage under
MS-DOS. The command
                         [type]
starts the MS-DOS editor (EDIT.COM) for a [type]-type
file.When a  graphical procedure is interrupted by error or
CTRL+C, the default video mode can be returned by the command
SCR. The command MLB starts the iterative procedure for which
the control parameters are stored in [name].PAR. The command
DI produces the directory listing for the open case.

If you have a licence for PKZIP/PKUNZIP (PKWARE Inc.) archive
utility you can use  the command CLOSE [name] to combine and
compress the files of the case. Extraction of the archive is
performed with the command OPEN [name].
     
The programs IMP (for conversion of the data from a
spectrometer), APP, ASC, BEST, L and D are not included in the
DEMO-release. The conversion programs are customized on
request. Arrays of binary or ASCII reals or integers (the
standard types easily available from modern spectrometers) can
be read routinely. Also WIN-NMR ASCII data and Jeol-data
obtained through HP-plotter output can be read.

The PERCH shell under Windows allows plotting of the spectra on all
Windows printers. The OBS and CAL type digital spectra can be converted
into  ASCII code by the command ASC [type] and then plotted with any
plotting program.
2.   SETUP

PERCH is distributed on two 3 1/2" disks containing compressed files. You
have to run one of the following setup methods to install PERCH on your
hard disk.

For DOS (if you intend to use PERCH under DOS only):

Create a directory for PERCH on the hard disk. Run DOSSETUP.BAT from disk
1.  DOSSETUP.BAT requires the paths from where to where you want  to
install PERCH as parameter. For example installing PERCH from disk drive
a: to directory PERCH on drive c: is done by:
          dossetup a: c:\perch          NOTE: no closing backlash (\)

For Windows 3.1 (to set up the complete package):

Start SETUP.EXE from disk 1 (under Windows 3.1) and follow the given
instructions. Setup installs PERCH in the selected directory and copies
following files to your Windows system directory:

     - WPERCH.DLL        (run time library for WPERCH.EXE)
     - VBRUN300.DLL      (run time library for WPERCH.EXE)
     - CMDIALOG.VBX      (common dialog module for WPERCH.EXE)
     - THREED.VBX        (3D-button module for WPERCH.EXE)
     - SETPERCH.EXE      (setup program for PERCH, in Windows root)
     - SETUPKIT.DLL      (run time library for SETPERCH.EXE)

The last two are needed to setup PERCH only.

Setup makes no modification in the configuration files except  in Windows
SYSTEM.INI.
Setup conditional (you have to confirm the change and a copy of the
previous SYSTEM.INI is made) installs the driver (DOSXNT.386) for the DOS-
extender under Windows.

     NOTE:     PERCH programs will not run under Windows 3.1 without this
               driver declared in the SYSTEM.INI.
3.   THE FLOW OF THE ANALYSIS

The following outline describes some general steps which might be varied
according to the problem. Some examples are given in chapter 5 and the
single PERCH programs and their call syntax is described in the appendix.

Step 1: Open the case

COMMAND:       PERCH [name]

where [name] = a 1-8 letters name for the case (default = the last case).
PERCH creates the necessary files, if not existing, for the case.

     NOTE:     If the case [name] has been created before, it can be
               opened simply with the command: OPEN [name].

Step 2: Import and manipulate data

COMMANDS: IMP, PAC, APP

First, transfer your spectrum from the spectrometer to the file [name].TXT
(see APPENDIX A2). Use IMP then to import the data into [name].OBS, which
is a binary type file and can be listed only using L OBS. You may also use
the command

               IMP  filename  (incl. full path, if path is different from
                              PERCH)
to import any file into [name].obs.

It is often practical to remove peakless parts of the spectrum in
[name].OBS. The removing and also some other manipulations like baseline
correction (essential for the D mode of MLDC) and smoothing are done under
the PAC command (see example 3). If the total spectrum is composed from,
for example, proton, carbon and fluorine spectra, they can be appended by
the command APP. The intensity scaling of the parts can be done also under
DP.
                    
Step 3: Create input files

COMMANDS: PMS, PIC, DAT

The input file for the spectral parameters [name].PMS can be prepared most
conveniently by editing an old PMS-type file. The command PMS starts the
DOS-editor for the [name].PMS file. The iterative programs read the
peak-top frequencies from the file [name].DAT. The observed frequencies
and their intensities can be fed to [name].DAT directly from the spectrum
using the PIC command or via editor using the DAT command.

The current [name].PMS and [name].DAT can be printed by L PMS and L DAT.Step 4: Simulate the spectrum

COMMANDS:  MLD, PER,DOR, LS, DP, L OUT

If going to use the integral transform iteration modes (I or D) of MLDC,
go to the step 6.

Simulation of the spectrum is started using the MLD command. The simulated
frequencies are written into the file [name].OUT and one may add a line-
shape using LS. The computed and observed spectra are compared on the
screen using the DP B command. The spectral parameters in the file
[name].PMS are modified starting the editor with PMS. The current list of
theoretical transition frequencies ([name].OUT) is printed by L OUT.


Step 5: Assign spectral lines

COMMANDS: DP, DAT

Display the observed and calculated spectra simultaneously on the screen
using DP B and assign the spectral lines. If [name].OBS or [name].CAL is
empty, DP displays the stick spectrum from [name].DAT or [name].OUT. If
want to modify [name].DAT, the editor is invoked by the DAT command.


Step 6: Optimize the spectral parameters

COMMAND: MLD,PER

Start the iterative procedure using the MLD command, choose the proper
mode and options and the number of iteration cycles. The detailed
description of the procedure is given in the manual of the program MLDC.

If many weak lines have been observed, the option 11 is recommended
(smaller weights are given for the weak lines).  The option 4 (smaller
weights are given for the lines of larger observed-calculated differences)
is useful in the beginning of the iterative procedure especially if there
seem to be many incorrect assignments.


Step 7: Compare the observed and calculated spectra

COMMANDS: LS, DP, L OUT

The result can be examined by displaying the spectrum to the screen by the
command DP B. If desired, the line-shape is added to the computed spectrum
by the command LS. To inspect the result of the iteration, one may also
list [name].OUT by L OUT, try to find the reasons for the largest
observed-calculated deviations and repeat steps 5-7 to refine the result.4.   GUIDED TOURS


For the available commands and their syntax, see Appendix A1.

4.1  Peak-top-fitting

   1.     Use PERCH [name] to open the case. If a similar old case exists,
          it is  practical to make a copy of the old input files by

                    COPY  [oldcase].*  [name].*

   2.     Transfer the digital experimental spectrum to [name].TXT and use
          IMP (or IMP filename) to  import the spectrum to the PERCH-type
          file [name].OBS.

     If the digital spectrum is not available, the peak-list must fed into
     [name].DAT manually (by DAT). The procedure is continued then from 
     item 5.

   3.     Use PAC to start the packing of the file. If the baseline of the
          spectrum is clearly below the zero intensity, use Z to get  the
          baseline to zero. The correction of the baseline by Fourier-type 
          fitting (recommended for the D mode of MLD) demands peak-picking
          under  the P mode. The spectrum can be then packed using the P
          mode and/or  the parts of the spectrum containing no signals can
          be removed by using the F9 or nA mode.

  4. Start the peak-picking procedure using the PIC command. The procedure
     contains the following steps:
          A.   Choose a well-separated line and estimate the line-shape
               parameters. The line-width is used as a parameter for the
               peak-picking. If no such a line exists, a fair estimate is
               obtained, for example, by measuring a splitting of two 
               peaks for which the peak-to-valley ratio is bigger than 90
               % (a good estimate can be obtained by using the program
               TLS).
          B.   Estimate the noise of the spectrum (normally not
               necessary).
          C    Choose the peak and shoulder criteria, so that all the
               observable lines are detected.
          D.   Interpolate the peak-top frequencies and exit with X and
               OK. The exit with OK produces the file [name].DAT, which
               contains the  interpolated peak-top frequencies and the
               intensities of the lines.

5.        Start editing the trial parameters with PMS. To display or to
          list the file use L PMS.

6.        Start the program MLDC with the command MLD and choose the mode
          S to start the simulation.
7.        Use DP B to display the computed and observed spectra. Use the
          option A to start the assignment procedure. The assignment
          procedure is described in details in chapter 5, example 1.

8.        Use MLD to start the iterative procedure. Give no options and
          then 1-5 iteration cycles. If the rms-value remains large, use
          the C mode to remove  the assignments of large
          observed-calculated difference. If the rms-value  converges well
          so that it is far smaller than the line-width, start  the
          peak-top-fitting choosing the P mode. If the number of the 
          rejected assignments tends to grow too large, rise up the
          rejection  criterium. If the rms-value converges well again,
          stop or try other line-widths and line-shapes in order to check
          the sensitivity of the result  to them.

9.        Use DP B to display the observed spectrum and the computed stick 
          spectrum on the screen. If the fit is good, use LS to add the
          line  shape to the computed spectrum and then DP B to display
          the spectra again on screen.

10.       Use L OUT to list the file [name].OUT. If some large
          computed-observed deviations are seen, try to explain and remove
          them considering, for  example, different sign combinations of
          the coupling constants.


4.2  Integral-Transform fitting

1-5. As 1-5 in 4.1..

6.        If strong windowing has been used in preparing the digital
          spectrum of a XAB-type spectrum, for example, it may be
          necessary to ensure  that ratio of the intensities (in
          [name].DAT and [name].OBS) of the X  and AB parts is 1:2. is
          done by comparing the observed and computed  spectrum under DP B
          and using the S command (see Appendix A10).

7.        Use MLD and the mode D (or mode I if [name].OBS is large or if
          there are not many overlapping lines in the spectrum) to start
          the IT-fitting. Use 1-2 'grand cycles' to optimize the spectral
          parameters. Exit with X and OK to store the last result. Examine
          the result on the screen (using DP B) and either modify the
          trial parameters (in the case of poor rms) or start the final
          refinement of the result with MLD  using the mode P and choose
          the option 1 to continue with the  parameters in [name].SAV.

8.   As 7-10 in 4.1..

     NOTE:     A successful IT-procedure demands a good-quality observed
               spectrum, good baseline correction and absence of strong
               impurity signals.5.   EXAMPLES


The distribution discs contain a few examples. To go through the analysis,
write the given command in DOS (with <enter>) or click the corresponding
button when you have started the PERCH shell (WPERCH.EXE) under Windows.
Additional instructions for the PERCH shell are given in brackets[ ].


Example 1:     A basic course

This is a simple case that can be used as tutorial to the PERCH system. If
a spectrum contains, for example, no AA'BB'-type symmetry and not much
stronger overlap of the multiplets of different nuclei than in the present
case, the analysis can be started with rather poor trial parameters and
the following procedures should work well.

1.   COPY EX1.* BASIC.*  [Open case EX1 and save as basic]
     -    The original digital spectrum file (EX1.OBS) and the input
          parameter file EX1.PMS) are copied into BASIC.*

2.   PERCH BASIC         [Click PERCH]
     -    Press <enter> to accept all defaults and to exit the program.

3.   PIC                 [Click PIC]
     -    Select the single line at about 40 Hz with cursors (right
          button), expand with <SPACE>so that the range is 0.5-1.0 Hz,
          centre the peak by T, fit the line-shape by <enter> and use X to
          skip to the next phase.
     -    Skip with X or estimate the noise using the default range, or
          choose one by yourself (type H for help), and use X to skip to
          the RED page.
     -    Choose peak-picking criteria so that all the well-observable
          signals are included. Giving, say, 0.5P sets the peak criterium
          (P) to 0.5%, plain 0.5 sets P to 0.5% and the shoulder criterium
          (S) to 1.5*P = 0.75%. Skip to the BLUE page with X. 
     -    Press <enter> to get the list of the interpolated peak-tops.
          Exit with X and use OK to store the result.

4.   PMS                 [Click Trial Parameters]
     -    Edit the trial spectral parameters or list them with L PMS. Fair
          trial spectral parameters are already in the file.

5.   MLD                 [Click MLD button]
     -    Choose the S mode to simulate the spectrum. 

     NOTE:     Remember to exit with OK to store the result!6.   DP B                 [Click Assign]
     -    Use A to step into the ASSIGN mode. You may also use O to
          maximize the overlay between the observed and computed spectra.
     -    'Manual': Choose the range by cursors or E (or, say, by 1.2E to
          decrease the expansion factor to 1.2 and then with the L and R
          keys) and move the cursors on the corresponding theoretical and
          observed lines using the cursor keys (CTRL, ALT; help is
          available with H). Assign the lines using A or ALT-A, if there
          are several nearly degenerate transitions under the theoretical
          signal (indicated by colours and seen also from the window up
          left, where the transition numbers, calculated frequencies and
          intensities and the assigned observed frequencies are
          displayed). You may remove assignments with the DEL, F8, F9 and
          F10 keys (see below or Help).
     -    'Automatic': Set the cursors so that only the leftmost six lines
          (the two triplets), for example, are between the cursors, use O
          to maximize the overlay of the spectra and assign them with
          CTRL-A. Use then F9 (the lower cursor defines the assignment to
          be removed ) or F8 (the upper cursor), DEL (the last cursor
          moved)  or F10 (all the assignments between the cursors) to
          remove the questionable assignments.
     -    Change the range  and repeat the above procedure.
     -    When enough lines are assigned, exit with X and use OK to save
          the assignments.

7.   MLD                 [Click MLD button]
     -    Start iteration with no options (0 or <enter>). Give then 5
          iteration cycles.
     -    If you have made a mistake in the assignments (which may appear
          as a large rms), continue the iteration using the C mode to
          remove automatically the poor assignments.
     -    When 5 iterations are done (or rms converged) choose the P mode
          to  perform the peak-top fitting using the default values for
          the line-shape parameters (by pressing <enter>). When converged,
          stop with X and use OK to store the result.

8.   LS                  [Click LS button]
     -    Use no options. Reply to every question with <enter>.

9.   DP B                 [Click Assign]
     -    Compare the result on the screen. If the result is good, exit
          with X and ESC. Otherwise repeat from item 6.


Example 2:     Test another coupling sign combination.

MLDC offers a fast way to test the effect of the sign of a coupling on the
goodness of the solution, after a good solution is found and if the system
is not of very strongly second-order type.

0.   Perform the above procedure (example 1).

1.   SAV                 [Click Refined Parameters]
     -    Change the sign of a coupling (for example, of J(1,3)) in
          BASIC.SAV.

2.   DP B                     [Click Assign]
     -    Use A to get into the assigning procedure.
     -    Use F10 to remove all the assignments.
     -    Skip with X and use OK to store the data.

3.   MLD                 [Click MLD]
     -    CHOOSE THE MODE P AND OPTION 1 (the option 1 to input the
          spectral parameters from BASIC.SAV) and then 5 iterations. The
          modes D and I (with one iteration cycle!) should work as well.
          The mode and options are given as separated  by a comma, the
          mode first:

                         P,1 <enter>

     -    If a mistake is made in giving the options, a letter for the
          number of iterations returns the control to the input of the
          options.
     -    If converged, exit with OK.

5.   DP B                [Click Assign]
     -    Compare the spectra. It is often useful to assign all the lines
          with CTRL-A to visualize a the result.



Example 3:     Packing and base-line correction of a 13C-coupled 1H NMR
               spectrum of benzene

This example shows how to use the program PAC for spectral manipulations
like base-line correction and packing. A proper base-line is essential,
for example, if the TLS-mode of MLDC is to be used.

1. COPY BENZENE.* TEST.* [open case BENZENE and save as TEST]

2. PERCH TEST

3. PAC

Step 1:   To maintain the base-line close to zero, use the command Z.
Step 2:   Find the peak-positions (this is necessary for the base-line
          correction, packing and smoothing procedures):
     -    Use P to step into the peak-picking procedure, give the peak and
          shoulder picking criteria (0.3...0.5 is OK) and the line-width
          (0.03...0.10 Hz is OK) so that all the observable lines are
          detected.
     -    Exit with X.
Step 3:   Find the base-line correction for each multiplet separately:
     -    Choose the range 110,170 (or so) with cursors, use B to step
          into the base-line fitting procedure and then [n]F (with n up to
          20) to find a proper base-line. The last one is accepted with
          <enter>.
     -    Exit with X.
     -    Repeat the procedure with the two other multiplets.

Step 4:   Remove the peakless parts:
     -    Choose range 160,222 and remove it by F9.
     -    Choose range 90,120 and remove it.
     -    Choose range 0,40 and remove it.
     -    Choose range -35,-111 and remove it.

Step 5:   Packing
     -    Use P to step into the peak-picking and packing mode, use
          defaults and then P to perform the packing.
     -    Exit with X.

Exit PAC with X and OK to store the new spectrum. The spectrum can now be
used normally for any PERCH operations. The above procedure  reduces the
number of the data-points from 22 K to about 3 K. If you close the case
with the command CLOSE [name] (demands the  compression utility PKZIP in
the path), the whole case is compressed to  20 K.
Example 4:     A basic course using the IT approach

1.  COPY EX1.* BASIC.*        [Open the case EX1 and save as BASIC]
     -    The original digital spectrum file (EX1.OBS) and the input
          parameter file (EX1.PMS) are copied into BASIC.*

2.  PERCH BASIC
     -    Accept all defaults and exit by pressing <enter>.

3.  PIC
     -    Find the single line at ca. 40 Hz (as above), fit the line-shape
          and use X to skip to the next phase.
     -    Skip with X, or estimate the noise first, to skip to the RED
          page.
     -    Press <enter> to display the spectrum and choose peak-picking
          criteria so that all the observable signals are included. Skip
          to the BLUE page  with X.
     -    Press <enter> to get the list of the interpolated peak-tops and
          exit with X and OK.

4.  PMS  or L PMS             [Click Trial Parameters]
     -    Edit the trial spectral parameters or list them.
     -    To start with very poor trial parameters: COPY EX2.PMS BASIC.PMS
          If curious, use MLD to simulate and then DP B to compare the
          spectra.

5.  MLD
     -    Start iteration with the D (or I) mode and with 1 'grand'
          iteration cycle.
     -    (You may exit MLD with X and OK now and check the result with DP
          B. Restart and continue MLD with mode P and the option 1).  
     -    Start the automatic peak-top-fitting using mode P and 5 iter-
          ations.
     -    Vary the line-width (or Gaussian %) to see the sensitivity of
          the result to the line-width (in the P mode).

6.  LS
     -    Use no options.

7.  DP B                 [Click Assign]
     -    Compare the result on the screen.









STRATEGIES for the IT-analysis:


1) A succesful IT analysis of spectra containing many peak-rich multiplets
demands a proper normalization of the intensities of the multiplets. For
example, if the area of a multiplet arising from a  proton is smaller than
the area of  a multiplet arising from another proton, the IT-procedure
tends to lead to too broad a multiplet for the former and too narrow one
for the latter and as the result the iteration converges poorly. The
situation is common if the spectrum has been treated with a strong
resolution enhancement procedure (the broad signals lose their intensity
in comparison with the sharp ones). The correct intensities of the
observed multiplets can be maintained in DP B with the scaling procedure,
as described in the following example (item 3).

2) Analysis of  large spin-systems demands often many attempts with the
IT-procedure, depending greatly on the goodness of the trial parameters.
However, a fair topological  similarity of the computed and observed
spectra is often found during a couple of IT-runs. If an attempt leads to
a clear improvement in the fit, one can use the command BEST to replace
the original trial parameters in the PMS-file with the last optimized
parameters from the SAV-file.

3) Start optimizing the large couplings first. To make the procedure
sensitive to a certain  coupling, use a SPAN which is 1-2 times the value
of the coupling. The small long-range couplings can be set to zero or kept
fixed. A convenient way to ignore a parameter from the iteration is to set
the index of the parameter (the first number on the card giving the
parameter) negative: the option 2 excludes it out of the iteration.

4) When the trial spectrum is close to the correct one, smaller SPAN than
the default value can be well used. As a rule, the initial SPAN can be set
1-2 times the largest uncertainty in the spectral parameters. 

5) Assignments can be included into the IT-analysis by using the option
33.

6) Parameters can be fixed also by using constraints. For example, one can
set two couplings equal to each others by giving the following type
constrainng equation (see the MLDC manual):
                              1.0*(X23) = 1.0*(X32)
The bigger is the weight of the equation, the more serious is the
constraint; the above weight of 1.0  is rather mild. Sometimes one may
also fix  the sum of two couplings.

Testing of the IT-analysis and a development of strategies for analysis of
strongly windowed spectra of large spin-system is under progress.  
Example 5:     A graduate course

The AA'BB'XX' spectrum of o-difluorobenzene (AABBXX.OBS) is an example of 
symmetrical spin-systems. For such systems fairly good trial spectral
parameters should be used. In the present case some typical problems may
arise from the FF-coupling: the information of the coupling is carried by
the weak 'satellite' lines and almost any value of it (if kept fixed)
gives a small rms in the IT-mode. 

1.   Start with PERCH AABBXX. No manipulation under PAC is needed in this
     case.

2.   Perform the line-shape and noise analysis and peak-picking so that
     even the smallest peaks are picked.

3.   Simulate the spectrum (for example with the standard values of
     couplings given in the AABBXX.PMS) and compare it with the observed
     one. In order to scale (see also A10) the observed spectrum so that
     the intensity of the lower field part is the same as that of the
     upper field part, set the cursors so that only the lower field parts
     of the observed and computed spectra are between the cursors and then
     press CTRL-S. Do the same for the upper field part. 

4.   Start MLD with the mode D and 1 iteration cycle. Use all defaults.
     Store the spectrum and compare the result with DP B. 

5.   If you start the analysis with the parameters given in AABBXX.PMS and
     use the defaults, the iteration tumbles down into a wrong minimum.
     Although the rms is  small, the satellite line patterns are
     incorrect. The  reason is that the default 95% range is too small for
     the parameter 9 (this is seen from MLDOC) and due to the correlation
     between parameters 9 and 10, the parameter 10 (FF-coupling) is
     changed greatly. - The situation can be solved in two different ways:
     (i) starting MLD again and setting the range of the parameters to
     200% or (ii) giving a bigger trial value for the parameter 9.  If the
     iteration is restarted with option 1, the procedure is not able to
     rise out of the local minimum.

     As a rule, if the value of a coupling is uncertain, it seems to be
     better to give too big trial value for it rather than too small one. 

NOTE  1:  Even if the range is set to bigger than 100%, the sign of the
coupling is not allowed to be changed. If the trial value of a parameter
is 1.0 Hz,  the range of 200% allows thus its variation  between 0 to 3.0
Hz. 

NOTE  2:  The run of the iteration depends on the defaults of the program
and may vary from a program version to another. The run is different in
PERCHIT (the trial parameters are given in A2B2X2.PMS), because the
original couplings between chemically equivalent nuclei are replaced with
their sums and differences. 
Example 6.     TLS-fitting (deconvolution) of a peptide NH-region.


1.  COPY NHFIT.OBS NH.* [open case NHFIT and save as NH]

2.  PERCH NH
     -    Set the line-width = 5...6 Hz, the Gaussian % = 0...-25% and the
          default noise = 1%. If the given line-width is much too large or
          too small, there may arise some problems in the peak-picking.

3.  PIC
     -    Use X (twice) to skip to the RED page..
     -    Set the PEAK CRITERIUM and the SHOULDER CRITERIUM to find also
          the poorly resolved lines. Skip to the next page.
     -    Use X to interpolate the positions of the peak-tops and to exit.
          Save the peak-top-list with OK.

4.  TLS
     -    Press <enter> to use the defaults and to skip to the green page.
     -    Use 3A to compute the total-line-shape using the present
          peak-list and line-shape and the 3-terms baseline function. The
          red line is the  optimized baseline, the green line shows the
          difference between the observed and computed spectra.
     -    Add or remove peaks by setting one cursor to the desired
          position and press <Ins> for insert or <Del> for delete. If you
          want to define also the line-width, define the half-heights of
          the lines with the cursors and use CTRL-<INS>.
     -    Improve the fit using the 3Z mode.
     -    Refine the fit using the 3Q mode.
     -    When ready, exit the page using X and use OK to store the list
          of transitions.
     -    List the transitions with L TOP.
          -    The statistics of the fitting is obtained by printing the
               DOC file (with L DOC).


If you have some external information about the spectrum, for example,
that it is composed of NH-proton doublets, the information can be
incorporated into the fitting as constraints. The trial line-widths can be
also given directly in the [name].TOP. This is convenient when the widths
of the lines vary much.
       
     NOTE:     The file NHFIT2.TOP shows an example of input file with
               constraints based on the knowledge that the spectrum is
               composed of a number of NH-doublets.
APPENDIX


A1.  Syntax

Program call syntax

Each program is started from DOS with the command line

                                           [command] s

where [command] is the name of the program and s is either the name of the
case (for the commands IMP, PERCH and APP) or the type of the file (for
the commands DP and LS) to be handled. For example, DP BS displays Both
the observed and computed spectra in the Stick mode, while the plain DP
displays only the calculated theoretical spectrum.

Control parameters

The control parameters and options demanded within the programs can be
given by the general syntax
                                                 xy

where x and y are letters, numbers or function keys. The integers or real
numbers are separated by a comma. Confirmation of the input by pressing
<enter> is needed if numerical characters are given only. For example,

     0.05W     in MLD gives the value of 0.05 Hz for the line-width.

     3.0E or 3E     in the display mode (PAC, PIC and DP) expands the
                    spectrum  with a factor of 3. When once defined the
                    expansion factor  is used also for the other display
                    commands (C,L,R,I,D) until the command W is used. W
                    sets the  expansion factor to the default value (=2).
     10,-10         or -10,10 or 10.0,-10 or 10,-10.0 <enter>  in the
                    display mode defines the range displayed. The order is
                    not significant in giving the spectral range. Single
                    number 10 brings the  midpoint of the spectrum to 10.
     CTRL-A    performs the automatic assignment procedure. The type of
               the letter (low/high case) is important only if i, I, d or
               D is used in dual display (DP B).

One may usually use as well real as integers in giving the control
parameters.


     NOTE:     Press F1 or h for more information.
ZOOMING

There are 4 different methods to scale the spectrum on the screen:
     1.   Enter frequencies:
          -    Give two frequencies (e.g. 10,25 or 10.0,25.0 or 25.,10) to
               define the range and confirm with <enter>.
          -    Give one frequency to centre the spectrum on this
               frequency.

     2.   Use short cuts:
          -    E expands the spectrum by a factor of two, C compresses the
               spectrum. If want to change the scaling factor (SF), give,
               for example, 1.2E to expand the spectrum by a factor of
               1.2. The scaling factor is valid so long as it is redefined
               or W is used.
          -    W displays whole the spectral range.
          -    R moves the spectrum to right, L to left. The default shift
               (corresponding to SF of 2) is 50% of the range. Change the
               shift, for example, 1.1R to move the spectrum by ca. 10%.
          -    In DP CTRL-R adjusts the range so that first observed line
               at right is brought to the screen.
          -    T moves the spectrum so that the highest signal is brought
               to the middle of the frequency scale.
          -    I, i, D, and d change the Y-scaling (see cursors).

     3.   Use cursor key
          -    The cursors are moved with the CURSOR keys: <left> and
               <right> move the LEFT cursor, CTRL-<left> and -<right> move
               the RIGHT one. ALT-<left> and ALT-<right> move both.
          -    The cursor range is expanded with the <space> .
          -    The cursors are  set to the left and right ranges with the
               F5-key.
          -    In the ASSIGN mode of DP there are two types of cursors
               (the frequency cursors and the peak-top cursors). The con-
               trol between these cursors is handled with the F3-key.
          -    The <up>-cursor expands the intensity, the <down>-cursor
               compresses it. The CTRL- and ALT-keys shift the spectrum in
               the vertical direction.

     4.   Use mouse
          -    Click (and keep pressed) the left mouse button to draw a
               box for defining a new view. When releasing the left mouse
               button, the spectrum is rescaled.
          -    Click (and keep pressed) the right mouse button to select
               and move one cursor. Release the button to fix the cursor
               at his current position and repeat the procedure for the
               second cursor. Press <space> or F2 to rescale the spectrum.

     NOTE:     Press W to display the whole spectrum.A2.  The PERCH files created for a case

The following files, except the [name].TXT file, are created for each case
with the PERCH command.

[name].TXT     contains the original digital spectrum (may have a user
               defined name as well).
[name].INF      contains spectral information and the current state of the
                analysis. The file can be listed with L INF.
[name].OBS contains the spectrum prepared from [name].TXT by using the 
           command IMP. The file is binary type but can be listed with 
           L OBS.
[name].DAT contains the peak-top-frequencies and the assignments for 
           MLD. The file is listed with L DAT. The editing of the file
           is started with DAT.
[name].PMS contains the trial spectral parameters for MLDC. It is
           listed  with L PMS and edited with PMS. The best way to
           make a new PMS-file is to edit it from an old PMS-file. The
           description of the parameters are given in the manual of
           MLDC.
[name].SAV contains the last iterated spectral parameters. The
           iteration is continued from these values using the option 1
           in MLDC. The file is listed with L SAV and edited with SAV.
[name].OUT contains the calculated frequencies produced by MLDC.
           Listed  with L OUT.
[name].CAL      contains the computed digital spectrum in the same format
                as [name].OBS. Can be listed with L CAL.
[name].TOP contains result of the deconvolution. (Compare DAT).
DOC        contains the document of the last operation. The document is
           listed with L DOC.
MLDOC   contains the document of the last MLD run. The document is 
        listed with L MLDOC.
VECTORS is a used for storing eigenvectors. With the option 1 it is us
        for prediagonalization of the Hamiltonian: this ensures the
        preservation of the transition indices when spectral  parameters
        are changed from the those used for the assignment  of the
        spectrum.

Following file are created and used by the PERCH shell (WPERCH.EXE) only:
[name].DOC contains the collected documents of  the PERCH programs.
[name].HIS contains history information concerning the document files.

 NOTE:  To clear the history, delete both [name].DOC and [name].HIS
        file!

If you have PKZIP/PKUNZIP (PKWARE Inc.) in the path:
CLOSE [name]    Compresses and archives [name].* to [name].ZIP and deletes
                [name].*.
OPEN [name]     Extracts archive [name].ZIP and opens the case without the
                PERCH procedure.
[name].ZIP           contains all the files of the case in compressed form.A3.   Program PERCH

Syntax:         PERCH [name]

[name] is a 1-8 letters name of the case to be opened. If the case does
not  exist, PERCH opens the necessary files (see A2) and writes [name]
into the file ICASE. If the case has been opened before, PERCH copies
[name].INF into ICASE. If [name] is not given on the command line, PERCH
asks for it till a 1-8 letters [name] is given (default = last case)

If the case has been created before, it can be opened with the command
OPEN [name] too.

Control parameters:

The program asks for rough estimates of the line-width, the Gaussian
contribution (in %) to the line-shape and the spectral noise (in %), which
are then stored into [name].INF to be used as default for example within
line-shape simulation (LS), etc..

 NOTE:  Proper estimates of these parameters are obtained using the PIC
        program. PERCH is also a way to change these values.


A4.   Program IMP

Syntax:  IMP filename

filename = is the name of the file to be imported (incl. full path if path
is different from the PERCH directory), default = [name].TXT.

IMP imports the spectral data into the file [name].OBS. Because the
spectrum can be modified by the PERCH manipulations, it is recommended
that the original spectral file is always saved.

 NOTE:  The [name].OBS and [name].CAL files are in binary PERCH format.

The conversion of the following data types is available in the present
version:

    1)    ASCII data without header.
    2)    ASCII data with header (Win-NMR ASCII).
    3)    Binary data of real*4 numbers (e.g. Win-NMR binary).
    4)    Binary data of integer*4 numbers (e.g. Bruker X32,..).
    5)    24 bit binary data with header (Bruker Aspect).
    6)    Jeol data imported to PC via HP-plotter channel.   
    7)    Custom format for binary data (16/24/32 bit, header size, bit-
swap, value skip). 

NOTE:   Custom format is able to import almost any kind of binary data
(e.g. Varian).   A5.   Program APP

Syntax:         APP [name]

The [name] is a 1-8 letters name of the spectrum created by appending a
non-overlapping set of  spectra. The program appends the [name].OBS and
[name].DAT type files. The spectra must be given in the correct order, so
that the highest frequency range spectrum is given first and so on.

The scaling of the spectrum is possible at this stage. For example, if an
A2X3 spectrum has been sliced into  the A and X parts before bringing to
the PERCH system, the parts must be scaled so that the former has an
intensity of 200 while the second one gets the value of 300. The scaling
is possible also under DP after the theoretical spectrum is created.



A6.   Program PAC

Syntax:  PAC

This program offers a versatile tool for manipulation of the spectrum. The
main purpose of the program is to compress the observed spectrum. There
are three different ways to reduce the number of spectral points: remove,
average and pack. The program offers also procedures for base-line
correction and smoothing. The manipulations (except p and s under the
peak-picking mode) are done on the range defined by the cursors or on the
whole range shown on the screen, when the cursors are not on the screen. 
This rule is not strict in the other programs.

 NOTE:  The base-line correction and smoothing should be done before the
        compression procedures.

Smoothing:           S

There are two procedures:
   1) Choose the spectral range to be smoothed and start smoothing
      with S. Give the number of the points to be  used in the
      polynomial fitting and the intensity level below which  the
      procedure is applied.
   2)   The second procedure is available in the P mode and is
        analogical to the  packing procedure (see below).Base-line correction:          B

Choose the range with CURSORS and start the procedure with B.

The option Z simply shifts the base-line so that negative intensities are
mostly removed. For the n-terms Fourier fitting the peak-picking must be
done first. Try different numbers of Fourier terms (3-20) and accept the
correct one with <enter> or exit with X.



Remove:         F9

Define the spectral range to be removed with the CURSORS and use the F9
key then to remove the part.


Average:             [n]A  (n = the number of the points to be averaged)

Define the range with cursors and use [n]A to average the range. For
example, 2A means that the intensities of points ...5,6,7,8,9,10,.. are
averaged so that I6 = (I5+I6+I7)/3 and that the points 5 and 7 are then
removed; as the result the number of the spectral points is reduced into
half.
Use values between 5-20 to compress areas with no signals. If you want to
compress areas having some weak and noisy signals, try values between 2-3.


Pack:           P

Packing demands that the peak-picking is done and the first 'P' starts the
peak-picking procedure. Choose the peak-picking criteria so that all the
essential signals are found and use then P to start the packing. The
procedure does not contain the peak-top-interpolation procedure (see PIC)
and the peak-top-frequencies are not stored when exiting the program.

The procedure removes spectral points around the peak-tops so that no
points are removed within +0.5*line-width and that the points out of the
range are smoothed so that the averaging range is doubled for each
0.5*line-width. The smoothed intensities are obtained by polynomial
fitting, which makes the procedure rather slow.




A7.   Program PIC

Syntax: PIC

PIC starts the peak-analysis of the observed spectrum. PIC contains the
following features: line-shape analysis, estimation of noise, peak-picking
and peak-top-interpolation.


Line-shape-analysis       (first page, first phase)

      Find a suitable peak (a single line), expand the view so that this is
      the only one on screen and centre it with T. With <enter> an
      iterative fitting is done. The procedure does not optimize the
      asymmetry factor, which may be given by, say, 1.5A (which means that
      the right side of the signal is broader). The last values are kept
      when the phase is skipped with X.


Estimation of noise       (first page, second phase)

      Choose a range without many signals and estimate the noise with
      <enter>. The violet line shows the baseline obtained by polynomial
      fitting.


Peak-picking              (second page)

      The program determines the default peak and shoulder criteria on the
      basis of the above analysis. The program uses also the line-width as
      one peak-picking criterium and too large a value prevents the
      detection of weak shoulders.

      Find a typical part and change the peak and shoulder criteria till
      all the peak-tops are detected and/or add or remove peaks by setting
      one cursor to the desired position and pressing <Ins> for insert or
      <Del> for delete.

      It is not harmful if some noise tops are included into the peak-list,
      because they are normally rejected in the spectral analysis.


Peak-top-interpolation    (third page)

      The default value for the number of points used in peak-top
      interpolation is determined by the line-width. The interpolation is
      based on polynomial fitting of the first derivative of the observed
      signal.

      The third page contains an option also for overlap correction of the
      peak-intensities. This may be useful if the I mode of MLDC is to be
      used.A8.   Program TLS

Syntax: TLS

TLS performs total-line-shape fitting (deconvolution) after a trial
peak-top-frequency list (received by PIC, default = [name].TOP) is given.
The procedure offers a useful tool for many applications of NMR. The
result is stored into [name].TOP.

The iterative algorithm allows refinement of 100 unknown parameters at the
same time. This means that the frequency, intensity and line-width of 32
lines can be refined simultaneously together with a 4 terms
baseline-function. If the line-width is kept the same for all the lines,
47 lines can be fitted. If the relative intensities of the lines are kept
constant (given by spectral simulation, see above), the number of
line-positions to be refined is 95.

The maximum number of points that can be handled simultaneously is
presently 8192. The default range can be changed using mouse, cursors
and/or shortcuts (press F1 or H for more information). The fitting can be
done in any order and the type of fitting is expressed for each line by
colours.

The basic procedure:

   1)   The first page: adjust the defaults if necessary. The trial
        values of the line positions, intensities and line widths are
        read from [name].TOP as the default. The first [name].TOP is
        created by  PIC.


   2)   (Optional) The second page:   shows the given constraints.
        If some lines are supposed to have, for example, an equal
        intensity, the  piece of information can be given as constraint
        line at the end of the [name].TOP file:

                                 10.0 * I15 = 10.0 * I16

       Where I15 means that the Intensity of the line 15 (the index given
      in  [name].TOP (or DAT or OUT) is equal to that of the line 16. The
      number  '10.0' gives the strength of the constraint. Use F (or any)
      for  frequencies and W for line widths.

  3)  The third page: use 3A (3 for three baseline terms) to see how
      well the total-line-shape can reproduced by using the line-shape
      parameters and line-frequencies obtained in the preceding
      procedures. The green line shows the difference between the
      observed and predicted spectra; the peaks of the green line show
      positions of hidden lines or shoulders.

  4)  Peaks can be added and removed by setting one cursor to the
      desired position and pressing <Ins> for insert or <Del> for
      delete. It's also possible to define a line-width for a single
      peak by setting both cursors and pressing CTRL+Ins. The
      TLS-fitting is started again with 3A or <enter>.    5)  The iterative adjustment of the peak-positions and intensities
      is started with [n]H. The programs does automatically at most 10
      iterative cycles  which can be repeated if the rms-fit is not
      converged.

  6 )   In the last stage the line-width or the line-widths for each
        line can be optimized with [n]Z or [n]Q.

Peak can be added and removed at any stage. One can maintain the trial
values by editing [name].TOP by TOP.

If you have some external information about the spectrum, for example that
it is composed of NH-proton doublets, the information can be incorporated
into the fitting as constraints. The trial line-widths can be also given
directly in the [name].TOP. This is convenient when the line-widths of the
lines vary very much.

 NOTE:  Standard errors for each peak are given in the document file
        DOC.


A9.   Program LS

Syntax: LS x.

x = C (for CAL) or D (for DAT) or T (for TOP).

LS adds the line-shape to the peak-list in [name].OUT , [name].DAT  or
[name].TOP and produces a digital spectrum in [name].CAL or, with option
15, in [name].OBS. One can use the command also to prepare synthetic
spectra with noise (option 2). In the basic mode the spectrum is computed
with a minimal number of spectral points. In the mode 1 the number of the
computed points is still reduced.
 
The option 15 allows a convenient way to make synthetic spectra for
educational or testing purposes: the option updates also the [name].INF
correspondingly.

 NOTE:  The LS procedure rescales the peak-intensities in [name].OUT.

 NOTE:  Also the line-widths given in [name].TOP and in [name].CAL
        produced by DORES simulation are included. If the line-width is
        negative, dispersion line-shape is used. If the individual line-
        widths vary much it may be necessary to use option 2.
A10.  Program DP

Syntax: DP xy

[x = C (for CALculated, default), O (for OBServed), B (for BOTH); y = S
for STICK spectra]

DP without options displays the last computed spectrum.

If line-shape simulation has not been done after the last completed
MLDC-calculation (exit with OK), DP displays the stick spectrum only. DP
also allows the scaling of the observed spectrum, for example, after using
APP or when the spectrum has been prepared by using strong resolution
enhancement (which often decreases the intensity of broad lines).

The commands CTRL-L and CTRL-R move the spectrum to the next calculated
line.

The assignment procedure is started with A. 

 -    Set the LOWER cursor ('the main cursor') with the CTRL-cursor keys to
      the theoretical transition you want to assign.
 -    Move the UPPER cursor with the cursor keys and press A to assign. The
      cursors are then moved automatically to the next transition and 
      peak-top.
 -    The ALT-cursor keys move both the cursors.
 -    If there are several nearly degenerate transitions close to each
      others (seen from the window at the upper left corner), use ALT-A to
      assign all of them. 
 -    Use O to maximize the overlap between the spectra. The procedure
      works better if both spectra are of stick-type. Find spectral part
      where the upper and lower parts resemble to each others, use O again
      and them CTRL-A to perform the auto-assignment. CTRL-A assigns
      automatically all the lines on the screen.
 -    F8 removed the assignments defined by the UPPER cursor, F9 removes
      the assignment defined by the LOWER cursor, F10 removes all the
      assignments on the screen. DEL removes the assignment defined the
      cursor that was moved last.

The CTRL-A option is useful also to compare refined and computed spectra
in detail.

 NOTE:  Be careful not to assign weak combination lines hidden under the
        strong main transitions. If a line is composed of several
        transitions, the intensity of the current transition defined by
        the cursor is indicated by magenta colour. The current
        transitions also are highlighted in the upper left window.
Scaling:

Scaling (S or CTRL-S) is useful if the intensities of the observed spectra
are biased due to windowing, different relaxation rates, etc.. Choose the
scale so that the corresponding parts of observed and computed spectra are
on the screen. The command S (or CTRL-S)  scales the intensity of the
lines in [name].DAT (or [name].OBS) so that it is the same as the
intensity of the corresponding lines in [name].OUT.



AN ASSIGNMENT PROCEDURE BASED ON THE USE OF MOUSE IS UNDER TESTING !!!





