Metabolomics represents the process of identifying and quantifying metabolites within a cell, tissue, or organism. This metabolite profile, or metabolome, is a reflection of the biochemical processes and pathways that are active under the current conditions. Therefore, a comparison of the metabolomes of biological systems under changing conditions (e.g., disease, toxicity, environmental changes, drug interactions, and gene modification) can provide detailed information on pathways involved in the response to these changes. Metabolomics studies have proven extremely useful in characterizing gene function, analyzing drug toxicity, and identifying biomarkers to diagnose diseases.

Profiling the metabolome of a biological system is inherently complicated by the large number of metabolites and dynamic biochemical reactions. Metabolomics studies need to be conducted carefully to avoid introducing additional variability to the samples being analyzed. Additionally, sample preparation and data collection can be challenging when coupled to the large number of samples that are generated for a metabolomics study. One source of variability can occur during data collection of numerous samples. While NMR spectroscopy is a robust and consistent technology, the necessary high-throughput data collection can be challenging, especially for non-NMR users. Three of NMRFAM’s NMR spectrometers are equipped with Bruker BioSpin SampleJet sample changers, which can hold up to 480 samples at consistent temperatures. Additionally, NMRFAM has utilized the python interpreter added to Bruker’s TopSpin 2.0 software to develop NMRbot, a high-throughput, robust, and reliable tool for automated data collection [1]. NMRbot is routinely used at NMRFAM for metabolomics, high-throughput ligand affinity screening, protein-ligand titration, and spectral data collection for deposition in the Biological Magnetic Resonance Data Bank (BMRB).

In NMR metabolomics, a significant amount of time is devoted to analyzing the overwhelming amount of spectral data collected. There are few reliable tools to pick, compare, and quantitate peak signals in the complex mixture of metabolites represented in the NMR spectra. NMRFAm has been actively involved with developing software to address this problem. NMRFAM has developed the program rNMR [2], an open-source software package for identifying and quantifying metabolites from NMR spectra. rNMR provides a simple graphics-based method for visualizing, identifying, and quantifying metabolites across multiple one- or two-dimensional NMR spectra. rNMR differs from existing software tools for NMR spectroscopy in that analyses are based on regions of interest (ROIs) rather than peak lists. ROI-based analyses support simultaneous views of metabolite signals from up to hundreds of spectra, and ROI boundaries can be adjusted dynamically to ensure that signals corresponding to assigned atoms are analyzed consistently throughout the dataset. rNMR greatly reduces the time required for robust bioanalytical analysis of complex NMR data against NMRFAM’s small molecule database (available from the BMRB) or the user’s own database.

The picking and quantification of NMR peaks has little utility in metabolomics if the metabolites associated with the peaks cannot be identified. Unfortunately, the process of characterizing the metabolites in a mixture by NMR requires multiple, time consuming NMR experiments and lengthy manual data analysis. A more efficient approach is to experimentally determine the chemical shifts of the metabolite standards, store the results in a database, and then compare the chemical shifts in the database with the chemical shifts observed in the metabolomics samples. NMRFAM has collaborated with the BMRB to develop an NMR spectral library with 1D and 2D 1H and 13C data for over 1,300 compounds [3, 4]. The spectral data is available from the BMRB. The data can be queried by software tools at BMRB or by external computer programs [3], and the original time-domain data can be down-loaded by users.

NMRFAM also offers technology for accurate metabolite quantification either alone or in conjunction with compound identification. NMRFAM has developed strategies that can easily quantitate metabolites in 2D 1H-13C HSQC experiments, such as the fast metabolite quantification (FMQ) method [5] and the 2D HSQC0 method [6]. The FMQ method utilizes a large mixture of metabolite standards as a concentration reference for the metabolites expected in the metabolomics sample. The 2D HSQC0 method uses the data from a series of HSQC experiments with incremented repetition times to determine concentrations without the need for standards. NMRFAM has also developed the 2D concurrent HMQC-COSY is an approach for small molecule chemical shift assignment and compound identification [7]. For example, two-dimensional 1H-13C spectra of mixtures can be acquired with acquisition times of 15 min and analyzed by FMLR in the range of 2-5 min per spectrum to identify and quantify constituents present at concentrations of 0.2 mM or greater. Each of these approaches requires the ability to accurately identify and quantify the peaks of interest. To address this issue, NMRFAM has developed a software package called “Newton” that implements an algorithm for fast maximum likelihood reconstruction (FMLR) [8]. Newton performs spectral deconvolution of nD NMR spectra for the purpose of accurate signal quantification, and incorporates the ability to monitor ROIs similarly to rNMR.



  1. Clos LJ, II, Jofre MF, Ellinger JJ, Westler WM, Markley JL. (2013) NMRbot: Python scripts enable high-throughput data collection on current Bruker BioSpin NMR spectrometers. Metabolomics 9(3): 558-63. PMCID: 3651530
  2. Lewis IA, Schommer SC, Markley JL. (2009) rNMR: open source software for identifying and quantifying metabolites in NMR spectra. Magn Reson Chem. 47(Suppl 1): s123-s6. NIHMSID: 165967, PMCID: 2798074
  3. Markley JL, Anderson ME, Cui Q, Eghbalnia HR, Lewis IA, Hegeman AD, Li J, Schulte CF, Sussman MR, Westler WM, Ulrich EL, Zolnai Z. (2007) New bioinformatics resources for metabolomics. Pac. Symp. Biocomput. 157-68.
  4. Ulrich EL, Akutsu H, Doreleijers JF, Harano Y, Ioannidis YE, Lin J, Livny M, Mading S, Maziuk D, Miller Z, Nakatani E, Schulte CF, Tolmie DE, Kent Wenger R, Yao H, Markley JL. (2008) BioMagResBank. Nucleic Acids Res 36(Database issue): D402-8.
  5. Lewis IA, Schommer SC, Hodis B, Robb KA, Tonelli M, Westler WM, Sussman MR, Markley JL. (2007) Method for determining molar concentrations of metabolites in complex solutions from two-dimensional 1H-13C NMR spectra. Anal. Chem. 79(24): 9385-90. NIHMSID: 63552, PMCID: 2533272
  6. Hu K, Ellinger JJ, Chylla RA, Markley JL. (2011) Measurement of Absolute Concentrations of Individual Compounds in Metabolite Mixtures by Gradient-Selective Time-Zero 1H-13C HSQC with Two Concentration References and Fast Maximum Likelihood Reconstruction Analysis. Anal. Chem. 83(24): 9352-60. PMCID: 3253702
  7. Hu K, Westler WM, Markley JL. (2011) Two-dimensional concurrent HMQC-COSY as an approach for small molecule chemical shift assignment and compound identification. J. Biomol. NMR 49:291-296.
  8. Chylla RA, Hu K, Ellinger JJ, Markley JL. (2011) Deconvolution of two-dimensional NMR spectra by fast maximum likelihood reconstruction: application to quantitative metabolomics. Anal. Chem. 83(12): 4871-80. PMCID: 3114465