MacWorld 1997 September/Macworld (1997-09).dmg/Serious Software/Cherwell Scientific Demos/gNMR□□/gNMR Guided Tour

Text File | 1997-06-17 | 22.2 KB | 483 lines | [TEXT/ttxt]

                  *******************
                   gNMR GUIDED TOUR
                  *******************


Welcome to the gNMR GUIDED TOUR.  We hope you will find it useful
in finding your way around the gNMR demo v 3.6.5.



NMR Simulation for Windows and Macintosh
****************************************

If you need a versatile software package to speed up your analysis of spectral
data, then take a look at gNMR! This short guided tour of the program takes you
through several examples and lets you see how easy it is to use.


Terminology
***********

This guide covers the use of both Macintosh and Windows versions. Where there is
a difference in the command or filename, written, the Macintosh instructions are
in brackets immediately after the Windows instructions.

The use of various symbols has also been made to indicate keyboard presses:

<Tab>     = Tab key
<Ctrl>+     = Hold Control key (Windows) or Command key (Mac) while pressing
additional key


Limitations of the demo
***********************

This demo contains a fully functional version of the gNMR main simulation
program, with only the following two differences:

* You cannot save files

* Printed output, clipboard copies, and sometimes screen displays of spectra will
have the word "DEMO" in big letters over it.

The full gNMR package contains several auxiliary programs for importing and
editing spectra. With it, you will receive a manual that not only documents gNMR
itself but also explains in detail how to use simulation to interpret NMR
spectra.

The size of systems gNMR can handle is determined mainly by the amount of RAM in
your system. A system of 10 inequivalent nuclei is usually the limit. See Step 8
on Simulatating Large Molecules for more details.


NMR simulation - an introduction
********************************

gNMR has been designed to help the simulation process right from the start of
entering information through to the data processing and actual simulation. gNMR
places particular emphasis on the analysis of second-order effects and allows the
direct comparison between experimentally acquired NMR data and simulated spectra.
gNMR can also help distinguish between reaction mechanisms of rearrangements via
inter- and intramolecular exchange processes.

Exact calculation of higher-order spectra for larger molecules can take a lot of
time and memory. gNMR will typically handle systems with up to 10 or 11 nuclei
(chemical-exchange calculations are limited to even smaller systems). In the
presence of symmetry or magnetic equivalence, you may be able to handle slightly
larger systems. If you try to do a simulation of a molecule that is too large for
gNMR, you may see a warning message suggesting that you save your data first (we
recommend that you do so). It is particularly easy to enter too-large systems
when you use structure import (Step 4 of this demo) and Step 8 gives some hints
for doing simulation of larger molecules. Simulation can be a process of trial
and error in an attempt to achieve the best fit between experimental and
simulated data, the process of assignment and full-lineshape iteration (see Step
7) will help match simulated data to your experimental results.

***************************************************************
* Step 1: Simulating a simple first-order spectrum of ethanol *
***************************************************************

Getting started
***************

Double-click on the gNMR program icon in your gNMR working directory to start
gNMR. It will start with the File Options dialog, in which you can set various
parameters including spectrometer frequency. Click on OK to accept the default
settings. The Molecule window will appear, in which you can enter shifts,
coupling constants etc. for a single molecule.


Entering data
*************

Let us first simulate the spectrum of very pure ethanol. Press <Tab>  to move to
the column labelled Group #n (number of nuclei). Enter a "3" (the CH2 group, 3
equivalent hydrogens) and press <enter> to move to the next row. Enter a "2" (the
methylene group), press <enter>, and enter a "1" for the OH group. Now we have
the nuclei we need we will enter their chemical shifts. Press <enter> twice to
move to the top of the next column, and then type "1.3",<enter>, "3.7", <enter>,
"5.0". The next column contains linewidths, which we don't need now. Press
<(enter>, <enter>, <Tab> (to skip to the next column) and enter the coupling
constants: "7.5", <Tab>, "6.0".

These are all the data we need: now click on the Recalculate button to calculate
the spectrum . A new window will appear containing the spectrum.

Once you have a spectrum on the screen, you can change its appearance in a number
of ways.

Clicking the arrow button-bar buttons (or choosing their menu equivalents) lets
you expand or contract the spectrum. Dragging in the empty area immediately above
the spectrum lets you move the spectrum; by Shift-dragging in the same area you
can select a subspectrum. To return to the full spectrum choose Plot|Full
Spectrum.

The Plot|Options dialog gives more precise control over many aspects of your
spectrum display. You can also copy or print the spectrum (it will have the text
DEMO over it).

**********************************************
* Step 2: Simulating a second order spectrum *
**********************************************

Now, change the chemical shift of the CH3 group to "3.3" by clicking on the
number "1.3" in the Molecule window and entering the new value; click on
Recalculate to calculate the new spectrum. Click on the Spectrum window to bring
it to the front. This will look much more complicated: the close proximity of the
CH3 and CH2 groups results in strong second-order effects. Now change the shift
value back to "1.3" , Recalculate and move to the spectrum window.


********************************************
* Step 3: Adding chemical exchange effects *
********************************************

Usually, ethanol contains some acidic impurities that exchange with the hydroxyl
proton and cause loss of the CH2-OH coupling. Let us try to simulate this (this
exchange calculation will take a while on a slow system, it takes 20 seconds on a
Pentium 90, 3 minutes on a Quadra 650 running the FPU version, or 35 seconds on a
PowerMac 8100). Select Molecule|Go To|Molecule 2 to add a second molecule to this
"sample". Set its chemical shift to 9.5 ppm (it is acidic after all). Click on
the Options button of this window, set the concentration to "1e-5" and click OK.

If you recalculate the spectrum now (Plot|Go To Window|Window 1), a larger range
is shown (to include the peak at 9.5 ppm, which is too small to show), but
otherwise the spectrum would be unchanged.
Now select Molecule|Exchange. Enter a rate ("10") and then press <enter>, "=",
<enter>(, "=", <enter>, "2-1", <enter>, "1-3".

What you have entered are the positions that each nucleus moves to in the
exchange reaction. From the ethanol molecule, nuclei 1 and 2 do not move ("="),
but the third one moves to the position of Molecule 2 - nucleus 1 ("2-1") and
vice-versa ("1-3").

All done? Press <Ctrl>+1(on the Mac :command key - 1) (equivalent to Plot|Go To
Window|Window 1) to recalculate the spectrum.

This example has also been prepared: ethanol.dta
(on the Mac : ethanol exchange).

This may take some time: exchange calculations, especially intermolecular ones,
take much longer than normal simulations.

Play a bit with the rate to see which value gives the strongest broadening and
what rate you need to see the fast-exchange limit.

**************************************************
* Step 4: Importing data via chemical structures *
**************************************************

Pasting in structures
*********************

gNMR can import chemical structures drawn with a number of chemical drawing
packages. Among the DEMO files, you will see ChemDraw, ChemIntosh and ISIS/Draw
files for o-chloro-aniline; we will show how to import these in gNMR. Now import
a structure file by choosing File|Import Molecule and selecting either
oca.cd2 (on the Mac : o-Chloro-Aniline (ChemDraw)),
oca.cw2 (on the Mac : o-Chloro-Aniline (ChemIntosh)) or oca.mol file (ISIS/Draw
for Windows only).

Note:     If you have one of these programs, you can also open the file from
within its creator program, copy the molecule to the clipboard, and then paste it
into an empty gNMR Molecule window using Edit|Paste Molecule.

A dialog will appear. Click on OK to use the defaults. gNMR will now take some
time to read in the structure and try to predict shifts and coupling constants.
This procedure uses a fragment list to help predict the shifts and coupling
constants and the predicted values provide a good starting point for the
simulation process.

After this, the Structure window appears, showing the molecular structure and
fields to enter parameters. Click on the aromatic proton to the left of the NH2
group and then enter a value for its linewidth (L.W.) field ("0.5"). Then click
on one of the NH2 protons (the one that results in a display of parameter values)
and enter a much larger linewidth ("10"). Finally click on the Recalculate button
to compute the spectrum. If you want to change parameters, you can do so via the
Structure window and/or the Molecule window used in the previous example.

************************************************
* Step 5: Dealing with other NMR-active nuclei *
************************************************

Other nuclei
************

Chlorine consists of a mixture of 35Cl and 37Cl, both with spin 3/2. Normally,
you don't see any coupling to Cl in NMR because relaxation is rapid. But in a
simulation you can change this. Click on the Cl atom and then on the Add button.
Then Shift-click on the hydrogen next to the chlorine, and enter a value of 40 in
the Jij field. The molecule will now contain an NMR-active chlorine atom with a
coupling constant of 40 Hz between 35Cl and 1H(and approx 33 Hz between 37CL and
1H).  Click on the Recalculate button: the Plot Options dialog appears, allowing
you to choose between 1H, 35Cl and 37Cl nuclei. Press <enter> to accept the
default 1H.


*****************************************************
* Step 6: Chemical exchange and reaction mechanisms *
*****************************************************

NMR can not only be used to understand the static structure of a compound, but
also (in favourable cases) to distinguish between reaction mechanisms of
rearrangements. As an example, we will consider two possible mechanisms for the
fluorine scrambling in Me2NPF4. They are called the "one-pair" and "two-pair"
mechanisms.


Low temperature simulation
**************************

To save you some time, the input files for both systems have already
been prepared for you.

me2npf41.dta (on the Mac: Me2NPF4 one-pair exchange),
me2npf42.dta (on the Mac: Me2NPF4 two-pair exchange)

In each mechanism, there are several equivalent ways in which the nuclei can
move; these will of course have the same rates. You have to tell gNMR about this,
which makes the Exchange windows more complicated than in the ethanol exchange
example of Step 3. For both cases, you can simulate the slow-exchange (static)
spectrum by entering a rate of 0: the results will be identical (triplet of
triplets). At very high rates (say, 10E5), both systems will give identical
quintets (try this out!). But at intermediate exchange rates (around 300) you
will see that the broadening of the centre line compared to the outer lines is
more pronounced in the one-pair mechanism; there are also distinctive differences
in the behaviour of the other lines. The best way to see this is to open both
files simultaneously (gNMR allows you to have several open files) and display the
calculated spectra side by side in separate Plot windows. The differences are
clear enough that a comparison with experimental
data has been used to prove the two-pair exchange mechanism for this compound.


Dealing with other NMR active nuclei
************************************

gNMR can also simulate the 19F spectrum of Me2NPF4 and this can be displayed
using Plot|Options and changing from 31P to 19F and the new spectrum is
displayed. The 19F spectrum is also rate-dependent and shows similar exchange
processes to the 31P spectrum. Why not try this out?


*********************
* Step 7: Iteration *
*********************

Helping spectra assignment
**************************

Calculating a spectrum and playing with the effects of changing shifts and
couplings can be interesting. But often, you have a measured spectrum, and you
want to know whether you can reproduce it by a simulation. This may be to see
whether your idea about the structure of the compound is correct, or it may be to
extract accurate shifts and couplings. In either case, you will not be satisfied
by trial-and-error methods: you want to have some kind of "best fit".

In NMR, there are two ways to do this. You can try to fit the positions of all
peaks in the spectrum by entering positions (and possibly intensities) obtained
from a measured spectrum; this is called "assignment iteration". Or
alternatively, you can try to fit on the whole spectrum; this is called
"full-lineshape iteration". The former procedure is much faster but also requires
more understanding and a better initial guess by the user; the second method
requires a measured spectrum in electronic form. It would take too much space
here to explain how to set up such calculations; this demo contains two
ready-to-go examples for the rather simple example of o-dichlorobenzene.


Assignment iteration
********************

For the assignment iteration example, open the file odcbex1.dta (on the Mac :
Assignment Iteration) in the gNMR working directory. The Molecule window will
show not only shifts and coupling constants, but also a set of "names" ("a",
"aa'", etc.) indicating which spectral parameters are to be optimized; equal
names mean that the parameters will be kept equal. Select Plot|Go To
Window|Window 1 to see the calculated and observed spectra together.

Observed spectra can be read in from external files created by a wide variety of
spectrometers eg Bruker and Varian. A conversion utility (gCVT, see later on) for
these files is provided in the full release of gNMR. Example files are already
prepared for this demo version.

If you want to inspect the list of calculated and observed peak positions, choose
Iterate|Assignments, and select 1H from the Nucleus pop-up in the dialog that
appears. Press OK and the log window will fill with data. To start the iteration,
select Iterate|Go. After a few short cycles, the procedure has converged: the
calculated and observed spectra will now be rather similar. If you again select
Iterate|Assignments|1H,(ensure that the Molecule or Spectrum window is active)
you will see a much closer match between the observed and calculated positions.


Full-lineshape iteration
************************

Full-lineshape iteration does not require the user to enter a list of peak
positions. To start this example, open the file odcbex2.dta (on the Mac:
Full-lineshape Iteration) file. The Molecule window will show the same set of
"names" as in the previous example. Display the calculated and observed spectra
by selecting Plot|Go To Window|Window 1. You will see that the match between
observed and calculated spectra is much worse than in the assignment iteration
example: we have made this example more difficult by choosing poorer starting
values for the shifts and coupling constants. In the Plot|Options dialog section
Experimental, the items Full-lineshape Iteration and Iterate on Linewidth have
been checked to tell gNMR to use this window for iteration. To start this
calculation, again choose Iterate|Go and sit back (this takes about 2 minutes on
a 486 66, 40 seconds on a Pentium 90, 5 minutes on a Quadra 650 and 50 seconds on
a PowerMac 8100). The results illustrate clearly that near
-perfect fits are possible. For a more spectacular example, you might want to run
the example.dta (on the Mac : Large Iteration Example) file. For best results,
open Plot windows 1 and 2 before starting the iteration. This should take up to 5
minutes on a Pentium 90 and 6 minutes on a 8100 Power Mac.


**************************************
* Step 8: Simulating large molecules *
**************************************

The first thing to do when you want to simulate a large molecule is to make sure
that gNMR has enough available memory and disk space. Allocating up to 10 Mb of
both can be useful; if a simulation requires even more, gNMR will run into other
limitations before then or the simulation will take unacceptably long anyway.

Let's suppose you want to simulate a molecule containing about 20 nuclei. Doing a
full and exact calculation on such a molecule is currently impossible on any
computer. The best way to approach this problem is to break it up into pieces. If
this can be done in such a way that there are no couplings between the pieces,
you will still get the correct result. In gNMR, you can put the pieces in
separate "molecule" windows. The file largamin.dta (on the Mac: Large Amine)
shows how this can be done: the full molecule 1, file largamin.cw2 (on the Mac:
Large Amine - ChemIntosh) was pasted in three Molecule windows of the same file,
and two rings in every molecule window were then excluded from the simulation.
Simulation of this 17-spin system then runs without problems. Depending on a
number of settings, gNMR may do such a reduction internally, so the simulation of
the full molecule may also succeed if you do not break it up yourself.

For cases where such a division of the molecule is not possible, gNMR has a new
method ("chunking") for doing approximate calculations. Basically, this
calculates spectra piecemeal by including, for each nucleus, only its relevant
environment in the calculation. As long as there are no nuclei in the system that
couple to nearly every other nucleus, this method can work well. Because it is
still considered experimental, this method is disabled by default, but you can
enable it by setting the Chunking Method in the File/Options dialog (Symmetry
section) to Fine (most accurate variation; recommended) or "Coarse".

If you want to test this (only recommended for 486 or higher and PPC systems),
open the file vrylarge.dta (on the Mac: Very Large System), which contains
molecule 2 (a 12-spin system) with chunking enabled.

Simulation (click on the Recalculate button) is no problem here, although it will
take a fair amount of computer time. Provided you have allocated enough memory to
gNMR, this spectrum can also be calculated exactly (set "Chunking Method" to
"None"). If you do this and compare it with the approximate result, you will see
that the differences are very small. For a more extreme example, open the file
toolarge.dta (on the Mac: Too Large System) corresponding to 16-spin molecule 3.
This still runs because chunking has been enabled, but if you disable it and try
to recalculate the spectrum the program will exit with the message "System Too
Large".


gNMR's other features include:
******************************

* Simulate spectra of mixtures containing up to 10 different compounds

* One-dimensional polymer simulation

* Baseline and phasing parameters

* Anisotropy

* Quadrupoles and more.

There are options for creating PostScript in clipboard copies (Mac) and .PS
files(Windows), customizing the appearance of spectra, and changing the defaults
for nearly all gNMR settings. This demo version supports all the features of the
commercially available full version of gNMR, except the saving of files and
utilities to convert and edit measured NMR spectra.


What you get with gNMR
**********************

The full gNMR package also includes gCVT (conversion of many spectrum formats to
gNMR format), gSPG (editing spectra, baseline correction, etc.) and a utility for
constructing symmetry databases for gNMR. The comprehensive gNMR manual explains
all of these, and also gives a critical overview of the use of simulation for
interpreting NMR spectra.


The gCVT File Conversion Program
********************************

The gCVT file conversion program can be used to import experimental spectra from
a number of 'foreign' file formats: Bruker Win-NMR, Bruker Aspect, Lybrics,
General Electric GE-SUN and Varian VNMR. As the most general-purpose exchange
format, plain ASCII import is also available. In addition, gCVT can be used to
move gNMR data and spectrum files between different versions of gNMR (both
Windows and Macintosh). This is useful if you want to exchange data with
colleagues who use a different version of gNMR.


gSPG Spectral Editor
********************

Experimental spectra are seldom perfect: they show noise, baseline drift,
impurity peaks, imperfect phasing, etc. This can be annoying, especially if you
need a good picture for a paper. And if you want to use the spectrum for a
full-lineshape iteration, such imperfections are often fatal. Obviously, the
first remedy to this problem is to obtain high-quality experimental data.
Sometimes, however, you just have to make do with a given spectrum. Even for
imperfect spectra, it pays to spend time on careful phasing and baseline
corrections; you may also want to try adjusting various weighting functions to
enhance the quality of the spectrum.

If you have done your best to obtain a good spectrum, but the iteration still
doesn't produce reasonable results, the most probable causes are baseline errors
and impurity peaks in the experimental spectrum: you can use the gSPG
spectrum-editing utility to do primitive baseline corrections and remove impurity
peaks.


How to order gNMR
*****************

You can order a full version of gNMR for evaluation on our risk-free 30-day
money-back guarantee. Order today and it could be on your desk in just a few
days. Ask about our special prices for multiple copies. Show gNMR to your
colleagues and we're sure they'll be grateful.

Call, fax or e-mail your nearest Cherwell Scientific office or your local
reseller to place your order, or visit our web site at
http://www.cherwell.com

Cherwell Scientific Publishing
The Magdalen Centre
Oxford Science Park
Oxford OX4 4GA

Tel:  [phone]
Fax: [phone]
e-mail:  [e-mail]

Cherwell Scientific Publishing
c/o CHEM Research GmbH
Hamburger Allee 26-28
D-60486 Frankfurt

Tel: [phone]
Fax: [phone]
e-mail:  [e-mail]

Cherwell Scientific Publishing Inc
744 San Antonio Road
Palo Alto, Ca 94303
USA

Tel:  [phone]
Fax: [phone]
e-mail:  [e-mail]

We look forward to receiving your order.