! /buddha/rwb1002/samples/amx/mem/README

R.W.Broadhurst, Department of Biochemistry, University of Cambridge, U.K.
26/Jul/1994

THE DATA:

The main spectrum in this directory is a 15N separated 3D
HNHB experiment.  Data was acquired on a Bruker amx600 in 
simultaneous mode for d1 and using States-TPPI for quadrature 
detection in d2 and d3. Using the FFT with a cosine squared
window function and zero-filling each dimension once yields
a data matrix of size 512 (d1) x 256 (d2) x 64 (d3).  To
improve resolution in d2 and d3, these dimensions were 
reprocessed using maximum entropy, producing a data matrix
of size 512 (d1) x 512 (d2) x 128 (d3).

METHOD:

(1) Preprocessing of the data.  In this case MEM will be used
    only for the indirect dimensions (d2 and d3), so the data
    is first processed in d1 by the FFT, producing a double
    interferogram (eg. by using the script file ljb4_12_3dft.scr
    to produce ljb4_12.spc).  For the present example, the 
    spectrum processed by MEM will be zero-filled twice.  Since
    the data does not have to be apodized for MEM processing,
    the resolution will be better in a MEM spectrum than in a
    FFT spectrum zero-filled equally.

(2) For reference and the selection of appropriate d2 and d3
    planes, the other 2 dimensions are then processed by the
    FFT, producing a 3D spectrum (eg. ljb4_12a.spc).

(3) With reference to the 3D spectrum (ljb4_12a.spc), several
    (d2,d3) planes are chosen.  To make sure that the MEM
    processing is working well it is important to choose a 
    plane with weak signals as well as a plane with strong
    signals.  This is because if the noise level is set too
    high, convergence will be rapid but genuine small signals
    may be lost.  2D (d2,d3) interferograms are then recovered
    from the 3D double interferogram for the chosen points in
    d1 using the extract program (eg. using the script file
    extract_2d.scr to produce ljb4_12_x233tt.spc).

(4) For reference, the 2D (d2,d3) interferograms are processed
    using the FFT with zero-filling twice (eg. using script
    file ljb4_12_3dft.scr to produce ljb4_12_x233ff.spc).  The
    noise level in the FID is then estimated using
    plot2 as follows.  The spectrum is displayed and a box made
    around a region which contains no signal.  If the "stats"
    button in the "Region of Plane" window is tapped, then the 
    standard deviation of data values in the box can be read
    off.  Several measurements should be made and the average 
    taken.  For a 2D box, this figure is related to the noise
    in the FID by the equation:

    (noise in FID) = (av sd noise in spectrum)/(sqrt(n1*n2/4))

    where n1 is the number of points in the first dimension of 
    the 2D interferogram (eg. in this case d2) and n2 is the 
    number of points in the second (eg. d3).

(5) The 2D (d2,d3) interferograms can now be processed by MEM
    (eg. using script ljb4_12_max2.scr to produce
    ljb4_12_x233mm.spc).  The 2D MEM and FFT spectra can now
    be compared and adjustments made to the list of processing
    commands.  (eg. if there are sync wiggles in the MEM spectrum
    or if the search is slow to converge, it may be necessary to 
    apply a decay function to the synthetic FID).

(6) When the 2D MEM spectra are satisfactory, the 3D spectrum
    can be processed (eg. using script ljb4_12_max2.scr to 
    produce ljb4_12m.scr).
    
   

FILE DESCRIPTIONS:

(A) SCRIPT, CONTROL AND LOG FILES:

(1) extract_2d.scr ! control script file for extracting a 2D
    plane from a 3D dataset

(2) ljb4_12_3dft.scr ! control script file for FFT 
    processing of a 3D dataset

(3) ljb4_12_d1.com ! command list for FFT processing in d1

(4) ljb4_12_d2.com ! command list for FFT processing in d2

(5) ljb4_12_d3.com ! command list for FFT processing in d3

(6) ljb4_12_max2.scr ! control script file for 2D maximum 
    entropy processing of a 2D slice or a 3D spectrum

(7) ljb4_12_max2.com ! command and parameter list for 2D 
    maximum entropy processing

(8) max2_ljb4_12_233.log ! log file output from processing
    the 2D slice ljb4_12_233.tt.scr by 2D MEM

(B) DATA AND PARAMETER FILES:

(1) ljb4_12.ser ! serial file from amx600 (in amx format)

(2) ljb4_12.ser.par ! serial file decription

(3) ljb4_12.spc ! spectrum file after processing with FFT in
    dimension d1 only

(4) ljb4_12.spc.par ! default description file

(5) ljb4_12a.spc ! spectrum file after processing with FFT in
    d1, d2 and d3

(6) ljb4_12a.spc.par ! default description file

(7) ljb4_12a.spc.par.old ! referenced description file

(8) ljb4_12m.spc ! spectrum file after processing with FFT in
    d1 and with MEM in d2 and d3

(9) ljb4_12m.spc.par ! default description file

(10) ljb4_12m.spc.par.old ! referenced description file

(11) ljb4_12_x233tt.spc ! double interferogram extacted from
     spectrum file ljb4_12.spc after processing with FFT in 
     d1 only with d1 fixed at point 233

(12) ljb4_12_x233tt.spc.par ! default description file

(13) ljb4_12_x233ff.spc ! 

(14) ljb4_12_x233ff.spc.par ! default description file

(15) ljb4_12_x233ff.spc.par.old ! referenced description file

(16) ljb4_12_x233mm.spc ! 

(17) ljb4_12_x233mm.spc.par ! default description file

(18) ljb4_12_x233mm.spc.par.old ! referenced description file

