ScienceSoft, LLC

User Manual

CHAPTER 1: Overview

NMRanalyst automates the analysis of one to three dimensional NMR spectra. It reduces experimental multidimensional NMR data to a list of detected spin systems, eliminating the tedious manual interpretation of raw NMR data. NMRanalyst's AssembleIt workwindow contains the FindIt, VerifyIt, and AssembleIt components. The FindIt component identifies the best matching structures for analysis results. It contains over 15.9 million common molecular structures and further ones can be added. The VerifyIt component rates a specified structure by its agreement with NMR results. It can compare the rating with all the FindIt structures for further confirmation of the specified structure. The AssembleIt component performs the structure elucidation. It combines the analysis results of several spectra and derives the most likely molecular skeletons from this often incomplete and ambiguous information.

The NMRplot program displays experimental, simulated, and residual spectra as contour, surface, and isosurface plots. The NMRgraph program displays and allows editing molecular structures. It predicts proton and carbon chemical shifts. It completes determined molecular skeletons by adding likely bond multiplicities and NMR unobserved heteroatoms.

The NMRanalyst spectral analysis software is based on a mathematical spin system model. NMRanalyst analyzes all acquired phase components simultaneously for maximum sensitivity. This provides maximum sensitivity for various types of multidimensional spectra. This approach contrasts with other computerized analysis strategies, such as peak picking, which ignore valuable spectral information. By using this novel mathematical approach, NMRanalyst often excels compared to even an experienced spectroscopist in the sensitivity, reliability, accuracy, and speed of the data analysis.

The software supports the analysis of the following and equivalent spectrum types:

1.1 The Automated Spectral Analysis

The first application of NMRanalyst was for 2D INADEQUATE spectra. The schematic explains this application for a >CH-CH2-CH3 molecular fragment. Its 1D proton decoupled 13C spectrum is shown at the back of the schematic. In the 2D spectrum, a pair of bonded carbons (e.g., CH-CH2) gives rise to a pair of anti-phase doublets, centered at the chemical shifts of the two carbons (vA and vB), and split approximately by the carbon-carbon coupling constant (J). Each of these AB spectral patterns is displaced along the F1 axis by the sum of the two relevant chemical shifts, vA+vB.

This spin system results in a bond pattern for each pair of bonded carbons symmetrically disposed about the diagonal as shown. Because the double-quantum frequency (in F1 direction) is the sum of the chemical shifts of the two coupled carbons, an initial analysis of a proton decoupled 1D carbon spectrum can be used to identify the regions of the 2D INADEQUATE spectrum which may contain correlation signals. For each pair of resonances identified in the 1D spectrum, two small regions of the spectrum are defined by the small rectangles shown in the figure above. In case of a bond between the two carbons under consideration, the characteristic pattern is contained within this "fitting area". Here, the AB and BC fitting areas contain the required patterns, but the AC fitting area does not, consistent with the bonding of the three-carbon fragment shown. The objective of NMRanalyst is to examine the fitting area appropriate for each pair of carbon resonances and to determine whether or not it contains a correlation (bond) signal.

This analysis strategy extends to other spectrum types by using the appropriate spin-system model. Shown in this schematic is the DQF-COSY spin system. The spectrum contains an active coupling of two protons. Each correlation (coupling) signal is composed of 16 anti-phase transitions. The spectral diagonal is normally overly crowded and NMRanalyst only examines the off-diagonal transitions in the two shown fitting areas. What can be improved by using this automated spectral analysis?

  1. A major limitation of NMR is its lack of sensitivity. A phase sensitive 2D dataset is acquired with four phase components and each phase sensitive 3D dataset with eight. Half of the phase components are acquired sequentially and hence their noise content is uncorrelated. NMRanalyst evaluates all phase components simultaneously by nonlinear regression analysis, making full use of the acquired data.
  2. The visual spectrum analysis can be described as "pattern recognition". What sticks out of the noise level and looks like an expected signal? Can this signal be assigned to other signals in the same or other spectra or is it perhaps a spectral artifact?NMRanalyst looks for the whole spin system, potentially consisting of several transitions. NMRanalyst starts with 1D resonance information and only searches the areas in the multidimensional spectrum which could contain spin systems. The automated analysis is more selective than the visual pattern recognition. It is also more sensitive as it can reliably detect signals even for signal-to-noise challenged spectra where resonances are not visible.
  3. The molecular structure determination using NMR remains labor intensive. Why not automate the steps which don't require the creative capabilities of a spectroscopist? "Spectroscopist-In-A-Box" is our goal for NMRanalyst. Things remain to be improved, but NMRanalyst already effectively supports using NMR as a structure elucidation tool.

1.2 Identification of Best Matching Molecular Structures

The most powerful NMR structure elucidation is based on 2D INADEQUATE spectra, but they are insensitive. The next best approach are indirect detection methods involving DQF-COSY, HSQC or HMQC, and HMBC spectra. Both approaches are covered in this manual. PubChem ( has published over 15.9 million distinct small molecular structures. Identifying the best matching structures from this collection for available NMR data is faster and can be more reliably automated than a full structure elucidation. FindIt, VerifyIt, and the derived findit script are described in the manual and we have recently published the approach.1

1.3 Using this Manual

This manual consists of three parts. Part one is the "NMRanalyst User Manual". It begins with CHAPTER 1: "Overview". It proceeds to the software installation (CHAPTER 2: "NMRanalyst Installation"). Then five "recipe style" tutorials (CHAPTER 3: "Tutorial I: Using NMRanalyst", CHAPTER 4: "Tutorial II: Setting Analysis Parameters", CHAPTER 5: "Tutorial III: Combining Analysis Results", CHAPTER 6: "Tutorial IV: Advanced Structure Elucidation", and CHAPTER 7: "Tutorial V: Additional NMRanalyst Features") cover using this software effectively. It is highly recommended to work through the tutorials using the software before analyzing and interpreting research datasets.

Part two of this manual is the "NMRanalyst Reference Manual" (starting with CHAPTER 8: "Using the Help System"). It covers all the software features and should be consulted for specific questions.

Part three of this manual is the appendix. It contains the FAQ list (CHAPTER 21: "Frequently Asked Questions"). To find answers, also consult the detailed index.

1.4 Typographic and Keyboard Conventions

Words typeset in:


indicate hypertext links (which might not show in the printed manual). Click on them to locate related information.
Typewriter style
indicates computer terms such as program input and output, shell variables, or command names.

Command line examples are prefixed by a letter indicating the required user privileges:

# execute command with super user (root) privileges,

% execute command with user privileges.

Keyboard keys and user interface buttons or switches are shown in square brackets such as [Enter].

1Reinhard Dunkel, Xinzi Wu, Identification of organic molecules from a structure database using proton and carbon NMR analysis results, Journal of Magnetic Resonance 188 (2007) 97-110.

ScienceSoft, LLC