Greetings again AMMRLers and BUMers, (sorry for the cross-posting)
As always, I received many helpful answers to my previous post, these
mailing lists are tremendous resources. A summary of the answers received
are posted below. A million thanks to all who contributed!
Brian
????????????????????????????????????????????????????????????????????????????
Greetings all,
I have a question concerning optimal processing of 2D spectra. On our
AMX-500, we have been processing the experiments with the following
parameters (because this was how the standard parameter sets were written
when I took over the lab):
gs-HMQC and gs-HMBC: F2: WDW=sine, SSB=2, LB=0.3 F1: WDW=em, SSB=2, LB=30
(Note, this makes the cross peaks in the HMBC appear star shaped)
COSYDFTP: F2: WDW=qsine, SSB=2.5, LB=0 F1: WDW=qsine, SSB=2.5, LB=0
Do these parameters seem optimal for these experiments? Are there any good
sources containing "standard" processing parameters for various experiments?
Brian
????????????????????????????????????????????????????????????????????????????
----------------------------------------------------------------------------
--------------------------------------------------------
Dear Brian,
The sine and qsine functions do not use the parameter LB, only SSB. I
have found as a starting point that using qsine with SSB=2 is a good
compromise in both dimensions for all these experiments. If you take the
time to optimize each one, then there may be better choices on a sample by
sample basis. I would really be interested to hear if anyone else has a
better suggestion because I am self-taught with respect to apodization for
2D.
----------------------------------------------------------------------------
--------------------------------------------------------
The books "100 (in the second edition, 150) And More NMR Experiments"
by Braun, Kalinowsky, Berger have this info.
----------------------------------------------------------------------------
--------------------------------------------------------
I'm frustrated with X-win too. 10 years ago we were using a Unity
for 2D. Varian had interactive phasing from which you could change
the shape of the apodization with a mouse or use mixed weighting
factors to observe the effect on both the fid and 1D slices.
This let you tailor the apodization to the data and was useful.
I'm not sure why this is not included in X-win.
----------------------------------------------------------------------------
--------------------------------------------------------
For gs-HMQC (inv4gs) & gs-HMBC (inv4gslplrnd) I usually process both
dimensions with WDW = sine & leave SSB = 0 on our AMX-600
when I have weak crosspeaks from a complicated multiplet peak then I put in
F2dim. SSB = 2 which elongates the intense peaks in F2 dim. which gives
same result as (F2) EM of 5hz LB also try SSB = 4 in F2 dim.
I always keep the F1 dim. as Sine with SSB=0
with the WDW = sine or qsine etc, it ignores the value of LB
only EM & GM looks at the LB value
----------------------------------------------------------------------------
--------------------------------------------------------
Hane you looked at the Bruker tutorials (under Help)? They are actually
quite good. The recommended parameters seem reasonable.
The valuefor LB in this, "F1: WDW=em, SSB=2, LB=30". depends on the
digital resolution in F1 (SW1 and TD1). Should be near the digital
resolution to reduce/avoid truncation artifacts.
You might also look at linear prediction to increase the quality
(resoluton AND S/N) of 2D experiments. LP is highly under-rated!
----------------------------------------------------------------------------
--------------------------------------------------------
The apodization function should fit the fid. For dqfcosy you should
look at a fid in the middle of the t1 region and make sure that the echo
amplitude fits the apodization maximum.
The apodiaztion for hetcor spectra should begin at unity and end at zero.
----------------------------------------------------------------------------
--------------------------------------------------------
For HMQC and HMBC if the S/N is good, use sine SSB=0 in both dimensions will
give good lineshape.
For HMBC if S/N is not that good, use SSB=2 for better sensitivity.
For COSYDFTP, I think qsine SSB=3 is good.
----------------------------------------------------------------------------
--------------------------------------------------------
wdw = qsine and ssb = 2 in both dimensions usually works
quite well for experiments like HMQC, HSQC, DQF-COSY,
NOESY and the like. However, for HMBC spectra (and, in
general, for all magnitude mode spectra), wdw = sine and
ssb = 0 in both dimensions are by far more advantageous
(can well be reasoned from the mathematical concept of
the 'matched filter'). Using the latter apodization function
may even render cross peaks visible in HMBC spectra which
disappear in the noise when the data are processed with
a qsine function.
As for optimal parameters, I think that these have to be found
separately for each case - in times of fast and powerful computing
equipment most likely by trial and error.
----------------------------------------------------------------------------
--------------------------------------------------------
The WDW parameter tells you which windowfunction is actually used : sine,
qsine => sinm (SSB applies), em => exponential (LB applies)
I generally use sine or qsine functions (no big difference) with ssb=0 or 2
(the two extremes) for phase insensitive and phase sensitive experiments,
respectively.
Check also the ph_mod parameter which determines what kind of phase
correction is applied (no : none, ph : phase correction, mc magnitude, ps:
power).
So : for your COSYDQFTP (phase sensitive) I'd use sine and ssb =2 for both
dimensions. (ssb=2.5 will make the signals less wide, but might create
wiggles (sinc-like line-shape).
For HMQC/HMBC : if it's phase sensitive you might try sine 2 for both
dimensions (else, try sine 0). LB =30 is _enormous_ !!!! : you want to
change that one anyway, even if you stick to em as a window function.
There is a book on these matters, I found helpfull :
NMR Data Processing, Hoch, J.C.; Stern, A.S. Wiley-LISS, Inc: NY, 1996. ISBN
0-4741-03900-4
----------------------------------------------------------------------------
--------------------------------------------------------
For a COSY experiment, a sine function with a shift of 0
ist best.
For most other experiments I use a qsine with ssb at 2.
With these functions LB play no role.
----------------------------------------------------------------------------
--------------------------------------------------------
Don't be afraid to try different window functions, after all you can
always get back to where you came from. With HMQC/HMBC quite
substantial improvements can be had if the processing parameters are
not optimum.
I run a lot of magnitude mode HMQC and HMBC with the same window
functions in F1 and F2 as a starting point ... WDW: QSINE SSB: 4
This usually gives nice round cross peaks.
The LB parameter is only used if the WDW is EM.
Try out the linear prediction function too (ME mod to lpfr etc) This
is great for waek samples when you only have enough time for 64 or
even less increments. You can still get a result without too much
loss of quality.
I tend not to run many phase-sensitive cosy but your numbers look a
good starting point.
Have fun experimenting!
----------------------------------------------------------------------------
--------------------------------------------------------
I have found "150 and More Basic NMR experiments" by S. Braun, H.-O.
Kalinowski, & S. Berger [Wiley-VCH ISBN 3-527-29512-7] to be very
helpful.
----------------------------------------------------------------------------
--------------------------------------------------------
----------------------------------------------------------------------------
--------------------------------------------------------
****************************************************************
K. Brian Killday
Research Specialist
Division of Biomedical Marine Research
Harbor Branch Oceanographic Institution
5600 U. S. 1, North
Ft. Pierce, FL 34946
Phone: 561-465-2400 ext. 456
Fax: 561-461-2221
Email: killday@hboi.edu
****************************************************************
This archive was generated by hypermail 2b29 : Sun Dec 31 2000 - 01:15:02 PST