ScienceSoft, LLC
www.ScienceSoft.net
TOC PREV NEXT INDEX


CHAPTER 3: Tutorial I: Using NMRanalyst

Five tutorials are provided covering FindIt/VerifyIt and the full NMRanalyst features. This tutorial covers determining likely structures with the findit script and through an NMRanalyst structure elucidation. Input values are provided. CHAPTER 4: "Tutorial II: Setting Analysis Parameters" practises setting analysis parameters for new data sets. CHAPTER 5: "Tutorial III: Combining Analysis Results" introduces the molecular carbon skeleton determination from sensitive indirect detection spectra. CHAPTER 6: "Tutorial IV: Advanced Structure Elucidation" adds placing nitrogen atoms in molecular skeletons and using DQF-COSY and 1,1-ADEQUATE spectra. CHAPTER 7: "Tutorial V: Additional NMRanalyst Features" describes the analysis of 3D spectra and further software features.

The NMR data for the prednisone1 compound used in the first two tutorials were provided by Stephen H. Grode, Ph.D. (MF: C21H26O5, 14.9 mg in DMSO-d6, Dual CryoProbe, Shigemi tube, Bruker 500 MHz DRX-500 spectrometer). The related files are in the directory ANALYST.ARCH/data, where "ARCH" is LINUX on Linux and WIN on MS Windows. On MS Windows, "data" is NMRanalyst's default working directory when the software is started. To start NMRanalyst on Windows (assuming a default installation), click on the menu button at the left bottom of the desktop, click , click , and choose .

On Linux, create a copy of the "data" directory contained in the software installation directory. The location of the installation directory is determined during the software installation. Recommended locations are $HOME, /home/$USER, /usr/local, /opt, /export/home, /usr, and /home. If the installation directory cannot be found, see CHAPTER 9: "Using the NMRanalyst Window". Make a copy of the data directory located inside the installation directory (here assumed to be /usr/local/ANALYST.LINUX) to one's working directory, join this directory, and start the software:

 % cp -r /usr/local/ANALYST.LINUX/data $HOME
 % cd $HOME/data
 % analyst

3.1 findit Script for Molecular Structure Identification

A major application of NMR is the determination of an unknown molecular structure. The NMRanalyst findit script, described here, automates the identification of best matching structures based on molecular formula, 1D proton FID, 1D carbon FID, or HSQC (HMQC) 1D spectral projections. Ideally, the best identified structure among the over 14.5 million considered structures is the correct one. As a nearly infinite number of small molecules could be synthesized, even this huge number of known molecular structures does not guarantee to contain the correct structure. Instead of the correct structure, similar molecular structures may be reported. Only the full structure elucidation identifies an unknown. But the findit identification of similar structures is automated and can be applied routinely. The full structure elucidation requires more NMR data and can be too resource intensive for routine applications.

Start the shown NMRanalyst application window as described in the previous section. This application window shows the state with which the software was last exited. Click on the [UNIX Shell] button at the top right of the application window.

The shown UNIX shell window is displayed. Issue in it the findit command to analyze the carbon, proton 1D FID, and/or HSQC (HMQC) spectral projections in the current directory. The best matching structures from the 15.9 million known structures are listed with the placement number, structure rating, and PubChem Compound ID (CID) in parentheses. Structure names and compound biological activities can be obtained from http://pubchem.ncbi.nlm.nih.gov/search through identified CIDs. If molecular formula or constraints are known, specify them as findit command line options, as described in SECTION 17.2: "FindIt: Identify Database Structures Best Matching NMR Data".

The final findit step is the shown NMRgraph display of best matching structures. The findit FID analysis, identification of best matching structures, and structure display take about 6 minutes (on a 3.06 GHz PC). Close the NMRgraph display by selecting [Exit] from the [File] pull-down menu.

3.2 Selecting Full NMRanalyst Mode

For the structure elucidation of the prednisone datasets, select the full NMRanalyst functionality. From the NMRanalyst [Edit] pull-down menu, select the [Preferences...] item. (This operation is also referred to as selecting [Edit] [Preferences...].) From the started popup, set the Mode: to [Full NMRanalyst] and deselect the [Show All Input Fields] switch as shown. Click the [OK] button or press the keyboard [Enter] key.

3.3 Selecting 2D INADEQUATE Spectrum Type

The window title of NMRanalyst displays the selected spectrum type. The window displayed in SECTION 3.1: "findit Script for Molecular Structure Identification" is set to INADEQUATE. If your window is set to another spectrum type, select the NMRanalyst [Spectrum Type] [INADEQUATE] item as shown.

3.4 The 1D Analysis Workwindow

At the bottom of the NMRanalyst application window, click the [1D Analysis] tab. The 1D analysis has already been completed by the findit script at the beginning of this tutorial and the results are stored in the directory data/FINDIT. To manually run the analysis again, click the [Start] button at the top of the workwindow or simply type the keyboard [Enter] key. This FID transform and spectral analysis takes a few seconds.

3.5 The FFT Workwindow

Click the [2D FFT] tab to switch to the shown FFT workwindow. Bruker AMX/DMX and Varian VNMR format 2D and 3D Free Induction Decays (FIDs) can be transformed, or spectra in Bruker and Varian format can be imported. Click [Start] and the transform of this 2D INADEQUATE FID takes only seconds.

3.6 The nD Analysis Workwindow

Click the [2D Analysis] tab to switch to the shown workwindow. SECTION 1.1: "The Automated Spectral Analysis" explains this 2D INADEQUATE spectrum analysis. Click [Start]. For every examined fitting area, a one-line report is displayed. For identified bonds, the chemical shift frequencies and the determined coupling constant are shown. All numerical values determined by this analysis are saved in the text file named inadequate.corr. For every bond region examined, the initial parameter estimates, the determined parameter values, and the corresponding error (marginal standard deviation) values are saved.

3.7 The Report Workwindow

Generating a report is the final step of the numerical analysis of a multidimensional NMR spectrum. Information from the previous workwindows are summarized and presented as a list of identified correlations (here: carbon-carbon bonds). Display the shown Report workwindow by clicking the [Report] tab. Click [Start] and the generation of this report completes nearly instantaneously.

The following table of correlation patterns is displayed (use the vertical scroll-bar to scroll to about the middle of the program output):

 Index     P(I)      Prob.                 C1   C2    Fa        Fb         J   
 ------------------------------------------------------------------------------
 #   180  10.402  1.0000000  CORRELATION   13   19    48.754    22.829    33.16
 #   191  10.364  1.0000000  CORRELATION   15   17    35.560    33.137    34.82
 #   176   9.134  1.0000000  CORRELATION   13   15    48.756    35.560    36.22
 ...
 #     7   2.747  0.9999912  CORRELATION    1    8   211.416    87.402    46.29
 CUTOFF    2.500  0.9998921
 ------------------------------------------------------------------------------

The identified bonds with determined integrals above the user-specified threshold (CUTOFF) are listed. The table shows from left to right: the correlation index number, the determined integral precision P(I),2 the statistical confidence the bond assignment carries, the location of the bonded carbons in the line list, and the resonance frequencies and the coupling constant determined for this bond pattern. It is more intuitive to inspect correlation information using the displayed correlation table. Since 2D INADEQUATE spectra are homonuclear, each correlation appears twice. This table assigns identified 2D correlations (bond signals) to the previously determined 1D resonance frequencies (SECTION 3.4: "The 1D Analysis Workwindow"):

 CORRELATION to RESONANCE 1    RESONANCE 2       RESONANCE 3        RESONANCE 4... 
===================================================================================
 Shift | Shift  J   Index | Shift  J   Index | Shift  J   Index | Shift  J   Index |
 [ppm] | [ppm] [Hz]   /   | [ppm] [Hz]   /   | [ppm] [Hz]   /   | [ppm] [Hz]   /   |
-----------------------------------------------------------------------------------
 211.41|87.40 46.29      7|66.08 40.71      8|                  |                  |
 210.19|58.81 37.95     28|49.78 37.80     30|                  |                  |
 185.03|127.0 50.81     42|123.7 53.58     43|                  |                  |
 167.18|123.7 63.86     60|41.91 36.15     67|31.52 38.67     71|                  |
 155.05|127.0 62.66     75|41.91 40.53     83|                  |                  |
 126.99|185.0 50.81     42|155.0 62.66     75|                  |                  |
 123.74|185.0 53.58     43|167.2 63.86     60|                  |                  |
 87.396|211.4 46.29      7|50.45 37.17    122|33.56 37.41    127|                  |
 66.077|211.4 40.71      8|                  |                  |                  |
 58.812|210.2 37.95     28|41.91 33.74    148|35.56 28.82    149|                  |
 50.448|87.40 37.17    122|49.78 30.27    156|48.75 31.89    157|15.49 36.34    165|
 49.777|210.2 37.80     30|50.45 30.27    156|                  |                  |
 48.749|50.45 31.89    157|35.56 36.22    176|22.82 33.16    180|                  |
 41.913|167.2 36.15     67|155.0 40.53     83|58.81 33.74    148|18.73 32.83    188|
 35.555|58.81 28.82    149|48.75 36.22    176|33.13 34.82    191|                  |
 33.562|87.40 37.41    127|22.82 32.76    198|                  |                  |
 33.130|35.56 34.82    191|31.52 31.41    201|                  |                  |
 31.518|167.2 38.67     71|33.13 31.41    201|                  |                  |
 22.824|48.75 33.16    180|33.56 32.76    198|                  |                  |
 18.730|41.91 32.83    188|                  |                  |                  |
 15.494|50.45 36.34    165|                  |                  |                  |
----------------------------------------------------------------------------------

For example, the first line:

 211.41|87.40 46.29      7|66.08 40.71      8|                  |                  |

specifies that the carbon atom resonating in the 1D spectrum at 211.41 ppm is bonded to the carbons resonating at:

  1. 87.40 ppm with J = 46.29 Hz (index #7), and
  2. 66.08 ppm with J = 40.71 Hz (index #8).

3.8 Determined Prednisone Structure

The correlation table can be displayed graphically. Select the [Graphic] tab, specify in the File With Plot Data input field the Report workwindow generated structure.plot, and click [Start] to obtain the shown structure display. A carbon atom is labeled by the 1D carbon resonance frequency in ppm. A bond is represented by a line between both bonded atoms and labeled by the determined coupling constant in Hz. Comparison with the prednisone structure at the beginning of this tutorial shows that all the prednisone carbon-carbon bonds are identified.

The prednisone molecular formula indicates that this structure contains oxygen atoms. While NMR does not observe them, observed carbon shifts reflect their presence. Select in the NMRgraph window [Prediction] [Place Heteroatoms...]. In the started popup, make sure that oxygen (O) is selected and click the [OK] button.

The shown structure results.

The structure can be repositioned to a more common orientation: Click on the 50.8 Hz bond to select it. (The bond line changes to a red color upon selection.) Select [Structures] [Rotate Bond To Vertical], [Structures] [Flip Horizontal], and [Structures] [Flip Vertical]. To deselect the bond, click in an empty area. This gives the shown prednisone structure.

3.9 Exiting the Software

To terminate NMRanalyst, click the [Exit] button in the NMRanalyst window. A popup appears, asking for confirmation before exiting NMRanalyst and its started windows.

1Corticosteroid to reduce swelling. It is used for many conditions such as skin diseases, breathing problems, artritis, certain cancers, and for hormone replacement.

2The integral precision is the number of marginal standard deviations that the determined integral value of the bond lies above zero.



ScienceSoft, LLC
www.ScienceSoft.net
TOC PREV NEXT INDEX