iNMR news #68

[GB, author of iNMR]

Here are the main news of the last two months:

1) MORE DOSY
As expected, the introduction of DOSY processing didn't generate any
significative feedback. As promised, I added a second tool. Don't know
how much useful it can be: it allows the extraction of the virtual
spectra of the pure components, assuming you already know the
diffusion coefficients. It is something I haven't found elsewhere and
quite an easy (and natural) tool to create. I call it "QR-DOSY", which
says a lot about the computational details. You can't find this term,
however, into iNMR. The internal name for the method is more helpful:
"DOSY > Extract Components".
QR-DOSY has some important limitations: the separation is seldom
complete and for the method to really work as promised the diffusion
coefficients must be distant on the diffusion scale.
You should not be surprised: the good thing about numerical treatments
is that the outcome (pitfalls included) can be anticipated with common
sense.
The two DOSY methods implemented so far (mono-exponential fitting and
QR-DOSY) can be applied to 3-D experiments. I have received a sample 3-
D spectra from a customer, which demonstrates that this 3-D capability
is necessary.

2) MORE SCRIPTS
Towards the end of July it was possible to see the list of published
scripts to grow:
http://www.inmr.net/library/index.html#Lua
the latest additions are valuable in the context of nD spectroscopy of
bio-polymers.
If the list keeps growing, it will be necessary to split it into sub-
lists.
Being the author of most of the scripts, my personal opinion is that
they are easy to read yet difficult to write! Having so many scripts
ready is very useful, because many frequent statements (and even
blocks) can be recycled in future scripts.
In other words, by copying, pasting and combining existing pieces you
can easily create new scripts, if you need them.
The major difficulty, however, is you can't do everything with a
script. Many corners of iNMR can only be reached with the graphic
interface.
When I discover that there is no way to accomplish a given task with a
script, I simply create a new command and recompile iNMR. You can't do
the same because nobody else has the source code (but me).
Anyway, before being discouraged, if you have any familiarity with
programming, try encouraging yourself: pick a script (possibly one
that you are going to use) and read it: there are chances you will
appreciate the simplicity and brevity of the script.

3) nD SMOOTHING
I first implemented a smoothing algorithm for the sake of baseline
correction in version 2. This year I have recycled it to "clean up"
the DOSY spectra. It has taken only a small step further to make the
"smoothing" a stand-alone algorithm. The effect is a cosmetic
simplification of spectra of any dimensionality. For example, you can
reduce or eliminate shoulders, spikes and the likes and the peaks
(when seen from above in a contour plot) will look more rounded.
For its purpose, smoothing is an alternative to symmetrization and
"t-1 noise removal" but the 3 methods can also be used simultaneously.
For the mechanism of action, smoothing resembles a lot the "visual
weighting" that you have certainly used for 1-D spectra. The fantastic
thing is that this multidimensional processing is (almost) as fast as
the 1-D counterpart (at least on modern computers). The final effect
is similar, yet not strictly equivalent, to weighting with a squared
cosine bell. You can modulate the amount of smoothing in real-time, as
you are used to do with iNMR.
The command can be found under the "Processing" menu.

4) COMPATIBILITY WITH SNOW LEOPARD (ONLY FOR THE CURIOUS READERS)
Apple released a major upgrade of the operating system two weeks ago.
From what I have seen, you can use both iNMR version 3.x.y and version
1.8 with Mac OS 10.6.
The latest version of iNMR is 3.2.2. It's the first version directly
made on Snow Leopard and should be able to run on Tiger (10.4) and
Leopard (10.5) too (I haven't tested everything yet).
It's becoming rather difficult to test iNMR on all possible
combinations of machines and systems. Two platforms are supported
(PowerPC and Intel) and 3 operating systems. Luckily I am not
interested in making a 64-bits version, otherwise the number of
possible combination would be even higher!
When there are so many different platforms to target, there is little
I can do. For example, I see now that the console reports a few
warnings while running iNMR on Snow Leopard. The warnings however come
from the operating system itself. They say something like that: "a
part of the operating system has tried to close a dialog that doesn't
exist anymore, but another part has blocked the attempt and nothing
will happen".
These warnings did not appear with Leopard: I conclude that Apple has
done weird things, probably to be certain that all dialogs are always
closed. I could spend days attempting to make the warnings disappear,
but the chances of success are low while the chances of becoming
incompatible with earlier systems is high. In conclusion, it's safe
and wise to do nothing.
A few advanced users like to run iNMR under "Instruments" and to read
the warnings on the console. This is great for testing, but don't ask
me to explain the meaning of the warnings: they come from the
operating system, not from iNMR.