I am not familiar with version 930601 of UXNMR as we jumped from 920801 to
940501. In the last version you can carry-on linear prediction in
both dimensions (I did not see if it was possible in 3D) simply by fixing
some extra parameters and executing xfb (or xf2 and xf1).
The processing parameters you have to set for each dimension are
ME_mod: it is set to "no" by default for no prediction . you can set it
to "LPfr" or "LPfc"
for forward prediction, and "LPbr" or "LPbc" for backward
prediction. If you use the xfb command, using "LPfr" or "LPfc" has
no importance. The same thing goes for backward prediction.
NCOEFF: It is the number of prediction coefficients. BRUKER recommends
a value of 30 for "regular use".
LPBIN: It seems to be the number of acquired points + the number of points
to be predicted, though I am not quite sure. I use a thumb rule of
LPBIN = 2 * TD. It gives good results.
SI must be set so as 2*SI is higher than LPBIN.
I have also discovered a couple of useful commands:
xtrf2 which allows you to visualize the ser file, and
add2d which can make a linear combination of 2D spectra.
#########################################################################
# Jean-Rene ALATTIA # #
# Centre de Biochimie Structurale # TEL: (33) 67 04 34 33 #
# Faculte de Pharmacie # FAX: (33) 67 52 96 23 #
# 15, Avenue Charles Flahault # #
# 34060 Montpellier Cedex # jrene@cbs.univ-montp1.fr #
#########################################################################
>From wem@bruker.de Wed Jan 11 06:18:09 1995
Received: from bam.bruker.de (bam.bruker.de [149.236.1.1]) by whistler.sfu.ca with SMTP (8.6.8/SFU-2.6H)
id GAA08445 for <dnaugler@sfu.ca> (from wem@bruker.de); Wed, 11 Jan 1995 06:18:05 -0800
Return-Path: <wem@bruker.de>
Received: from bamsoft.software.bruker.de by bam.bruker.de with smtp
(Smail3.1.28.1 #3) id m0rS3V4-000YAUC; Wed, 11 Jan 95 14:54 MET
Received: by bamsoft.software.bruker.de (Smail3.1.28.1 #6)
id m0rS3Up-000lRwC; Wed, 11 Jan 95 14:53 MET
Message-Id: <m0rS3Up-000lRwC@bamsoft.software.bruker.de>
From: wem@bruker.de (Werner Mausshardt)
Subject: Re: Maximum entropy
To: dnaugler@sfu.ca
Date: Wed, 11 Jan 1995 14:53:55 +0100 (MEZ)
X-Mailer: ELM [version 2.4 PL2]
Content-Type: text
Content-Length: 766
Status: RO
Dear David Naugler,
> Is there someone who could advice us on the use of maximum entropy for
> the improvement of 2d spectra. The brucker user manual is unclear to
> us.
Bruker offers the maximum entropy package Maxent; this is an extra
software package with it's own manual; UXNMR only provides an interface
(frontend) to Maxent.
UXNMR itself supplies forward and backward linear prediction
(processing flag ME_mod LPfr, LPfc, LPbr, LPbc) in the 1D, 2D
and 3D processing package. Information about this is found in
the release letter.
For detailed questions send me an email.
Sincerely Yours,
Werner Mausshardt
-- Werner Mausshardt E-mail: wem@bruker.de Software Department Tel: +49 721 5161 448 Bruker Analytische Messtechnik GmbH Fax: +49 721 5161 437>From kfw@lubrizol.com Wed Jan 11 07:00:38 1995 Received: from interlock.lubrizol.com (interlock.lubrizol.com [165.129.2.1]) by whistler.sfu.ca with SMTP (8.6.8/SFU-2.6H) id HAA10621 for <dnaugler@sfu.ca> (from kfw@lubrizol.com); Wed, 11 Jan 1995 07:00:18 -0800 Return-Path: <kfw@lubrizol.com> Received: by interlock.lubrizol.com id AA07402 (InterLock SMTP Gateway 1.1 for dnaugler@sfu.ca); Wed, 11 Jan 1995 09:59:54 -0500 Message-Id: <199501111459.AA07402@interlock.lubrizol.com> Received: by interlock.lubrizol.com (Internal Mail Agent-1); Wed, 11 Jan 1995 09:59:54 -0500 Date: Wed, 11 Jan 95 11:01:25 +0000 From: "Kurt Wollenberg" <kfw@lubrizol.com> Subject: Re: Maximum entropy To: "David Naugler" <dnaugler@sfu.ca> X-Mailer: VersaTerm Link v1.1.1 Status: RO
Kurt Wollenberg Research Chemist The Lubrizol Corp. e-mail kfw@lubrizol.com
>Is there someone who could advice us on the use of maximum entropy for >the improvement of 2d spectra. The Bruker user manual is unclear to >us. > >Thanks, > >David Naugler >Institute of Molecular Biology and Biochemistry >Simon Fraser University >Burnaby, B.C. >Canada V5A 1S6 >tel. (604) 291-5657 >FAX: (604) 291-3765 > > David, Sue Pochapsky of Bruker (Billerica) has a manual titled "Advanced Processing" that contain information on using linear prediction in 2D and 3D processing, as well as additional useful information, in using both forward and reverse linear prediction in 1D applications. This is not maximum entropy but we use the linear prediction capabilities, which gives us increased resolution especially in the f1 domain of 2D inverse experiments.
For linear prediction, typically we use f1 lpbin = 0 and double or quadruple the f1 si. This way you linear predict all the way out from f1 td to f1 si. If you use f1 lpbin different than f1 si then you will zero fill out the difference of f1 lpbin and f1 si, e.g. if f1 td = 1k and f1 lpbin = 2k and f1 si = 4k then you linear predict out from 1k to 2k and zero fill from 2k to 4k. With the old software you also have to define ME_mod as Lpfr for real ft or Lpfc for complex ft, however, with the new software, I believe that this is automatically defined from mc2. The other parameter that you need to define is NCOEF; a suggested setting is TD/8 or TD/6, however this is not possible if you use a large TD since the prediction times then become quite long. Typically we use NCOEF of 64 for larger TD values.
Again, SUe has a manual on using linear prediction which I am sure she would be glad to send to you. Good Luck.
Regard, Kurt