MagMet: A fully automated web server for targeted nuclear magnetic resonance metabolomics of plasma and serum

Nuclear magnetic resonance (NMR) spectral analysis of biofluids can be a time-consuming process, requiring the expertise of a trained operator. With NMR becoming increasingly popular in the field of metabolomics, there is a growing need to change this paradigm and to automate the process. Here we introduce MagMet, an online web server, that automates the processing and quantification of 1D 1 H NMR spectra from biofluids-specifically, human serum/plasma metabolites, including those associated with inborn errors of metabolism (IEM). MagMet uses a highly efficient data processing procedure that performs automatic Fourier Transformation, phase correction, baseline optimization, chemical shift referencing, water signal removal, and peak picking/peak alignment. MagMet then uses the peak positions, linewidth information, and J-couplings from its own specially prepared standard metabolite reference spectral NMR library of 85 serum/plasma compounds to identify and quantify compounds from experimentally acquired NMR spectra of serum/plasma. MagMet employs linewidth adjustment for more consistent quantification of metabolites from higher field instruments and incorporates a highly efficient data processing procedure for more rapid and accurate detection and quantification of metabolites. This optimized algorithm allows the MagMet webserver to quickly detect and quantify 58 serum/plasma metabolites in 2.6 min per spectrum (when processing a dataset of 50-100 spectra). MagMet's performance was also assessed using spectra collected from defined mixtures (simulating other biofluids), with >100 previously measured plasma spectra, and from spiked serum/plasma samples simulating known IEMs. In all cases, MagMet performed with precision and accuracy matching the performance of human spectral profiling experts. MagMet is available at http://magmet.ca.

Denoising magnetic resonance spectroscopy (MRS) data using stacked autoencoder for improving signal-to-noise ratio and speed of MRS

BACKGROUND

While magnetic resonance imaging (MRI) provides high resolution anatomical images with sharp soft tissue contrast, magnetic resonance spectroscopy (MRS) enables non-invasive detection and measurement of biochemicals and metabolites. However, MRS has low signal-to-noise ratio (SNR) when concentrations of metabolites are in the range of millimolar. Standard approach of using a high number of signal averaging (NSA) to achieve sufficient SNR comes at the cost of a long acquisition time.

PURPOSE

We propose to use deep-learning approaches to denoise MRS data without increasing NSA. This method has potential to reduce the acquisition time as well as improve SNR and quality of spectra, which could enhance the diagnostic value and broaden the clinical applications of MRS.

METHODS

The study was conducted using data collected from the brain spectroscopy phantom and human subjects. We utilized a stack auto-encoder (SAE) network to train deep learning models for denoising low NSA data (NSA = 1, 2, 4, 8, and 16) randomly truncated from high SNR data collected with high NSA (NSA = 192), which were also used to obtain the ground truth. We applied both self-supervised and fully-supervised training approaches and compared their performance of denoising low NSA data based on improvement in SNR. To prevent overfitting, the SAE network was trained in a patch-based manner. We then tested the denoising methods on noise-containing data collected from the phantom and human subjects, including data from brain tumor patients. We evaluated their performance by comparing the SNR levels and mean squared errors (MSEs) calculated for the whole spectra against high SNR "ground truth", as well as the value of chemical shift of N-acetyl-aspartate (NAA) before and after denoising.

RESULTS

With the SAE model, the SNR of low NSA data (NSA = 1) obtained from the phantom increased by 28.5% and the MSE decreased by 42.9%. For low NSA data of the human parietal and temporal lobes, the SNR increased by 32.9% and the MSE decreased by 63.1%. In all cases, the chemical shift of NAA in the denoised spectra closely matched with the high SNR spectra without significant distortion to the spectra after denoising. Furthermore, the denoising performance of the SAE model was more effective in denoising spectra with higher noise levels.

CONCLUSIONS

The reported SAE denoising method is a model-free approach to enhance the SNR of MRS data collected with low NSA. With the denoising capability, it is possible to acquire MRS data with a few NSA, shortening the scan time while maintaining adequate spectroscopic information for detecting and quantifying the metabolites of interest. This approach has the potential to improve the efficiency and effectiveness of clinical MRS data acquisition by reducing the scan time and increasing the quality of spectroscopic data.

Automated identification and quantification of metabolites in human fecal extracts by nuclear magnetic resonance spectroscopy

We report the development of a software program, called MagMet-F, that automates the processing and quantification of 1D 1 H NMR of human fecal extracts. To optimize the program, we identified 82 potential fecal metabolites using 1D 1 H NMR of six human fecal extracts using manual profiling and a literature review of known fecal metabolites. We acquired pure versions of those metabolites and then acquired their 1D 1 H NMR spectra at 700 MHz to generate a fecal metabolite spectral library for MagMet-F. The fitting of these metabolites by MagMet-F was iteratively optimized to replicate manual profiling. We validated MagMet-F's automated profiling using a test set of six fecal extracts. It correctly identified 80% of the compounds and quantified those within <20% of the values determined by manual profiling using Chenomx. We also compared MagMet-F's profiling performance to two other open-access NMR profiling tools, Bayesil and Batman. MagMet-F outperformed both. Bayesil repeatedly overestimated metabolite concentrations by 10% to 40% while Batman was unable to properly quantify any compounds and took 10-20× longer. We have implemented MagMet-F as a freely accessible web server to enable automated, fast and convenient 1D 1 H NMR spectral profiling of fecal samples. MagMet-F is available at https://www.magmet.ca.

Quantitative Analytical and Computational Workflow for Large-Scale Targeted Plasma Metabolomics

Quantifying metabolites from various biological samples is necessary for the clinical and biomedical translation of metabolomics research. One of the ongoing challenges in biomedical metabolomics studies is the large-scale quantification of targeted metabolites, mainly due to the complexity of biological sample matrices. Furthermore, in LC-MS analysis, the response of compounds is influenced by their physicochemical properties, chromatographic conditions, eluent composition, sample preparation, type of MS ionization source, and analyzer used. To facilitate large-scale metabolite quantification, we evaluated the relative response factor (RRF) approach combined with an integrated analytical and computational workflow. This approach considers a compound's individual response in LC-MS analysis relative to that of a non-endogenous reference compound to correct matrix effects. We created a quantitative LC-MS library using the Skyline/Panorama web platform for data processing and public sharing of data. In this study, we developed and validated a metabolomics method for over 280 standard metabolites and quantified over 90 metabolites. The RRF quantification was validated and compared with conventional external calibration approaches as well as literature reports. The Skyline software environment was adapted for processing such metabolomics data, and the results are shared as a "quantitative chromatogram library" with the Panorama web application. This new workflow was found to be suitable for large-scale quantification of metabolites in human plasma samples. In conclusion, we report a novel quantitative chromatogram library with a targeted data analysis workflow for biomedical metabolomic applications.

An efficient LC-MS method for isomer separation and detection of sugars, phosphorylated sugars, and organic acids

Assessing central carbon metabolism in plants can be challenging due to the dynamic range in pool sizes, with low levels of important phosphorylated sugars relative to more abundant sugars and organic acids. Here, we report a sensitive liquid chromatography-mass spectrometry method for analysing central metabolites on a hybrid column, where both anion-exchange and hydrophilic interaction chromatography (HILIC) ligands are embedded in the stationary phase. The liquid chromatography method was developed for enhanced selectivity of 27 central metabolites in a single run with sensitivity at femtomole levels observed for most phosphorylated sugars. The method resolved phosphorylated hexose, pentose, and triose isomers that are otherwise challenging. Compared with a standard HILIC approach, these metabolites had improved peak areas using our approach due to ion enhancement or low ion suppression in the biological sample matrix. The approach was applied to investigate metabolism in high lipid-producing tobacco leaves that exhibited increased levels of acetyl-CoA, a precursor for oil biosynthesis. The application of the method to isotopologue detection and quantification was considered through evaluating 13C-labeled seeds from Camelina sativa. The method provides a means to analyse intermediates more comprehensively in central metabolism of plant tissues.

The influence of cystathionine on neurochemical quantification in brain tumor in vivo MR spectroscopy

PURPOSE

To evaluate the ability of the PRESS sequence (T_E  = 97 ms, optimized for 2-hydroxyglutarate detection) to detect cystathionine in gliomas and the effect of the omission of cystathionine on the quantification of the full neurochemical profile.

METHODS

Twenty-three subjects with a glioma were retrospectively included based on the availability of both MEGA-PRESS and PRESS acquisitions at 3T, and the presence of the cystathionine signal in the edited MR spectrum. In eight subjects, the PRESS acquisition was performed also in normal tissue. Metabolite quantification was performed using LCModel and simulated basis sets. The LCModel analysis for the PRESS data was performed with and without cystathionine.

RESULTS

All subjects with glioma had detectable cystathionine levels >1 mM with Cramér-Rao lower bounds (CRLB) <15%. The mean cystathionine concentrations were 3.49 ± 1.17 mM for MEGA-PRESS and 2.20 ± 0.80 mM for PRESS data. Cystathionine concentrations showed a significant correlation between the two MRS methods (r = 0.58, p = .004), and it was not detectable in normal tissue. Using PRESS, 19 metabolites were quantified with CRLB <50% for more than half of the subjects. The metabolites that were significantly (p < .0028) and mostly affected by the omission of cystathionine were aspartate, betaine, citrate, γ-aminobutyric acid (GABA), and serine.

CONCLUSIONS

Cystathionine was detectable by PRESS in all the selected gliomas, while it was not detectable in normal tissue. The omission from the spectral analysis of cystathionine led to severe biases in the quantification of other neurochemicals that may play key roles in cancer metabolism.

Contribution of macromolecules to brain 1 H MR spectra: Experts' consensus recommendations

Proton MR spectra of the brain, especially those measured at short and intermediate echo times, contain signals from mobile macromolecules (MM). A description of the main MM is provided in this consensus paper. These broad peaks of MM underlie the narrower peaks of metabolites and often complicate their quantification but they also may have potential importance as biomarkers in specific diseases. Thus, separation of broad MM signals from low molecular weight metabolites enables accurate determination of metabolite concentrations and is of primary interest in many studies. Other studies attempt to understand the origin of the MM spectrum, to decompose it into individual spectral regions or peaks and to use the components of the MM spectrum as markers of various physiological or pathological conditions in biomedical research or clinical practice. The aim of this consensus paper is to provide an overview and some recommendations on how to handle the MM signals in different types of studies together with a list of open issues in the field, which are all summarized at the end of the paper.

Automatic 1D 1H NMR Metabolite Quantification for Bioreactor Monitoring

High-throughput metabolomics can be used to optimize cell growth for enhanced production or for monitoring cell health in bioreactors. It has applications in cell and gene therapies, vaccines, biologics, and bioprocessing. NMR metabolomics is a method that allows for fast and reliable experimentation, requires only minimal sample preparation, and can be set up to take online measurements of cell media for bioreactor monitoring. This type of application requires a fully automated metabolite quantification method that can be linked with high-throughput measurements. In this review, we discuss the quantifier requirements in this type of application, the existing methods for NMR metabolomics quantification, and the performance of three existing quantifiers in the context of NMR metabolomics for bioreactor monitoring.

Deep learning-based target metabolite isolation and big data-driven measurement uncertainty estimation in proton magnetic resonance spectroscopy of the brain

PURPOSE

The aim of this study was to develop a method for metabolite quantification with simultaneous measurement uncertainty estimation in deep learning-based proton magnetic resonance spectroscopy (¹ H-MRS).

METHODS

The reliability of metabolite quantification depends on signal-to-noise ratio (SNR), linewidth, and degree of spectral overlap (DSO), and therefore knowledge about these factors may be utilized in measurement uncertainty estimation in deep learning-based ¹ H-MRS. While SNR and linewidth are typically estimated from a representative singlet, DSO needs to be estimated metabolite-specifically. We developed convolutional neural networks (CNNs) capable of isolating target metabolite signal on simulated rat brain spectra at 9.4T, such that, in addition to metabolite content, the signal-to-background ratio (SBR) as a quantitative metric of DSO can be estimated directly from CNN-output for each metabolite. The CNN-predicted SBR was adjusted according to its pre-defined relationship to the ground-truth SBR by exploiting the big spectral data (N = 80 000), and used for measurement uncertainty estimation together with the SNR and linewidth from the CNN-input spectrum. The proposed method was tested first on the simulated spectra in comparison with LCModel and jMRUI and further on in vivo spectra.

RESULTS

The proposed method outperformed LCModel and jMRUI in both quantitative accuracy and measurement uncertainty estimation. Using in vivo data, the metabolite concentrations from the proposed method were close to the reported ranges with the measurement uncertainty of glutamine, glutamate, myo-inositol, N-acetylaspartate, and Tau less than 10%.

CONCLUSION

The proposed method may be used for metabolite quantification with measurement uncertainty estimation in rat brain at 9.4T by exploiting the spectral isolation capability of the CNNs and the availability of big spectral data.

Translational Metabolomics: Current Challenges and Future Opportunities

Metabolomics is one of the latest omics technologies that has been applied successfully in many areas of life sciences. Despite being relatively new, a plethora of publications over the years have exploited the opportunities provided through this data and question driven approach. Most importantly, metabolomics studies have produced great breakthroughs in biomarker discovery, identification of novel metabolites and more detailed characterisation of biological pathways in many organisms. However, translation of the research outcomes into clinical tests and user-friendly interfaces has been hindered due to many factors, some of which have been outlined hereafter. This position paper is the summary of discussion on translational metabolomics undertaken during a peer session of the Australian and New Zealand Metabolomics Conference (ANZMET 2018) held in Auckland, New Zealand. Here, we discuss some of the key areas in translational metabolomics including existing challenges and suggested solutions, as well as how to expand the clinical and industrial application of metabolomics. In addition, we share our perspective on how full translational capability of metabolomics research can be explored.

Intact metabolite spectrum mining by deep learning in proton magnetic resonance spectroscopy of the brain

PURPOSE

To develop a robust method for brain metabolite quantification in proton magnetic resonance spectroscopy (¹ H-MRS) using a convolutional neural network (CNN) that maps in vivo brain spectra that are typically degraded by low SNR, line broadening, and spectral baseline into noise-free, line-narrowed, baseline-removed intact metabolite spectra.

METHODS

A CNN was trained (n = 40 000) and tested (n = 5000) on simulated brain spectra with wide ranges of SNR (6.90-20.74) and linewidth (10-20 Hz). The CNN was further tested on in vivo spectra (n = 40) from five healthy volunteers with substantially different SNR, and the results were compared with those from the LCModel analysis. A Student t test was performed for the comparison.

RESULTS

Using the proposed method the mean-absolute-percent-errors (MAPEs) in the estimated metabolite concentrations were 12.49% ± 4.35% for aspartate, creatine (Cr), γ-aminobutyric acid (GABA), glucose, glutamine, glutamate, glutathione (GSH), myo-Inositol (mI), N-acetylaspartate, phosphocreatine (PCr), phosphorylethanolamine, and taurine over the whole simulated spectra in the test set. The metabolite concentrations estimated from in vivo spectra were close to the reported ranges for the proposed method and the LCModel analysis except mI, GSH, and especially Cr/PCr for the LCModel analysis, and phosphorylcholine to glycerophosphorylcholine ratio (PC/GPC) for both methods. The metabolite concentrations estimated across the in vivo spectra with different SNR were less variable with the proposed method (~10% or less) than with the LCModel analysis.

CONCLUSION

The robust performance of the proposed method against low SNR may allow a subminute ¹ H-MRS of human brain, which is an important technical development for clinical studies.

Comparison of time-gated surface-enhanced raman spectroscopy (TG-SERS) and classical SERS based monitoring of Escherichia coli cultivation samples

The application of Raman spectroscopy as a monitoring technique for bioprocesses is severely limited by a large background signal originating from fluorescing compounds in the culture media. Here, we compare time-gated Raman (TG-Raman)-, continuous wave NIR-process Raman (NIR-Raman), and continuous wave micro-Raman (micro-Raman) approaches in combination with surface enhanced Raman spectroscopy (SERS) for their potential to overcome this limit. For that purpose, we monitored metabolite concentrations of Escherichia coli bioreactor cultivations in cell-free supernatant samples. We investigated concentration transients of glucose, acetate, AMP, and cAMP at alternating substrate availability, from deficiency to excess. Raman and SERS signals were compared to off-line metabolite analysis of carbohydrates, carboxylic acids, and nucleotides. Results demonstrate that SERS, in almost all cases, led to a higher number of identifiable signals and better resolved spectra. Spectra derived from the TG-Raman were comparable to those of micro-Raman resulting in well-discernable Raman peaks, which allowed for the identification of a higher number of compounds. In contrast, NIR-Raman provided a superior performance for the quantitative evaluation of analytes, both with and without SERS nanoparticles when using multivariate data analysis. © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 34:1533-1542, 2018.

Multivoxel proton magnetic resonance spectroscopy in facioscapulohumeral muscular dystrophy

INTRODUCTION

Facioscapulohumeral muscular dystrophy (FSHD) is a hereditary disorder that causes progressive muscle wasting. This study evaluates the use of proton magnetic resonance spectroscopy (¹ H MRS) as a biomarker of muscle strength and function in FSHD.

METHODS

Thirty-six individuals with FSHD and 15 healthy controls underwent multivoxel ¹ H MRS of a cross-section of the mid-thigh. Concentrations of creatine, intramyocellular and extramyocellular lipids, and trimethylamine (TMA)-containing compounds in skeletal muscle were calculated. Metabolite concentrations for individuals with FSHD were compared with those of controls. The relationship between metabolite concentrations and muscle strength was also examined.

RESULTS

The TMA/creatine (Cr) ratio in individuals with FSHD was reduced compared with controls. The TMA/Cr ratio in the hamstrings also showed a moderate linear correlation with muscle strength.

DISCUSSION

¹ H MRS offers a potential method of detecting early muscle pathology in FSHD prior to the development of fat infiltration. Muscle Nerve 57: 958-963, 2018.

1 H-MRS processing parameters affect metabolite quantification: The urgent need for uniform and transparent standardization

Proton magnetic resonance spectroscopy (¹ H-MRS) can be used to quantify in vivo metabolite levels, such as lactate, γ-aminobutyric acid (GABA) and glutamate (Glu). However, there are considerable analysis choices which can alter the accuracy or precision of ¹ H-MRS metabolite quantification. It is currently unknown to what extent variations in the analysis pipeline used to quantify ¹ H-MRS data affect outcomes. The purpose of this study was to evaluate whether the quantification of identical ¹ H-MRS scans across independent and experienced research groups would yield comparable results. We investigated the influence of model parameters and spectral quantification software on fitted metabolite concentration values. Sixty spectra in 30 individuals (repeated measures) were acquired using a 7-T MRI scanner. Data were processed by four independent research groups with the freedom to choose their own individualized and optimal parameter settings using LCModel software. Data were processed a second time in one group using an independent software package (NMRWizard) for an additional comparison with a different post-processing platform. Correlations across research groups of the ratio between the highest and, arguably, the most relevant resonances for neurotransmission [N-acetyl aspartate (NAA), N-acetyl aspartyl glutamate (NAAG) and Glu] over the total creatine [creatine (Cr) + phosphocreatine (PCr)] concentration, using Pearson's product-moment correlation coefficient (r), were calculated. Mean inter-group correlations using LCModel software were 0.87, 0.88 and 0.77 for NAA/Cr + PCr, NAA + NAAG/Cr + PCr and Glu/Cr + PCr, respectively. The mean correlations when comparing NMRWizard results with LCModel fitting results at University Medical Center Utrecht (UMCU) were 0.87, 0.89 and 0.71 for NAA/Cr + PCr, NAA + NAAG/Cr + PCr and Glu/Cr + PCr, respectively. Metabolite quantification using identical ¹ H-MRS data was influenced by processing parameters, basis sets and software choice. Locally preferred processing choices affected metabolite quantification, even when using identical software. Our results reinforce the notion that standard practices should be established to regularize outcomes of ¹ H-MRS studies, and that basis sets used for processing should be made available to the scientific community.

Quantification of γ-aminobutyric acid (GABA) in 1 H MRS volumes composed heterogeneously of grey and white matter

The quantification of γ-aminobutyric acid (GABA) concentration using localised MRS suffers from partial volume effects related to differences in the intrinsic concentration of GABA in grey (GM) and white (WM) matter. These differences can be represented as a ratio between intrinsic GABA in GM and WM: r_M . Individual differences in GM tissue volume can therefore potentially drive apparent concentration differences. Here, a quantification method that corrects for these effects is formulated and empirically validated. Quantification using tissue water as an internal concentration reference has been described previously. Partial volume effects attributed to r_M can be accounted for by incorporating into this established method an additional multiplicative correction factor based on measured or literature values of r_M weighted by the proportion of GM and WM within tissue-segmented MRS volumes. Simulations were performed to test the sensitivity of this correction using different assumptions of r_M taken from previous studies. The tissue correction method was then validated by applying it to an independent dataset of in vivo GABA measurements using an empirically measured value of r_M . It was shown that incorrect assumptions of r_M can lead to overcorrection and inflation of GABA concentration measurements quantified in volumes composed predominantly of WM. For the independent dataset, GABA concentration was linearly related to GM tissue volume when only the water signal was corrected for partial volume effects. Performing a full correction that additionally accounts for partial volume effects ascribed to r_M successfully removed this dependence. With an appropriate assumption of the ratio of intrinsic GABA concentration in GM and WM, GABA measurements can be corrected for partial volume effects, potentially leading to a reduction in between-participant variance, increased power in statistical tests and better discriminability of true effects.

Mass spectrometry as a quantitative tool in plant metabolomics

Metabolomics is a research field used to acquire comprehensive information on the composition of a metabolite pool to provide a functional screen of the cellular state. Studies of the plant metabolome include the analysis of a wide range of chemical species with very diverse physico-chemical properties, and therefore powerful analytical tools are required for the separation, characterization and quantification of this vast compound diversity present in plant matrices. In this review, challenges in the use of mass spectrometry (MS) as a quantitative tool in plant metabolomics experiments are discussed, and important criteria for the development and validation of MS-based analytical methods provided.This article is part of the themed issue 'Quantitative mass spectrometry'.

Spectral fitting using basis set modified by measured B0 field distribution

This study sought to demonstrate and evaluate a novel spectral fitting method to improve quantification accuracy in the presence of large magnetic field distortion, especially with high fields. MRS experiments were performed using a point-resolved spectroscopy (PRESS)-type sequence at 7 T. A double-echo gradient echo (GRE) sequence was used to acquire B0 maps following MRS experiments. The basis set was modified based on the measured B0 distribution within the MRS voxel. Quantification results were obtained after fitting the measured MRS data using the modified basis set. The proposed method was validated using numerical Monte Carlo simulations, phantom measurements, and comparison of occipital lobe MRS measurements under homogeneous and inhomogeneous magnetic field conditions. In vivo results acquired from voxels placed in thalamus and prefrontal cortex regions close to the frontal sinus agreed well with published values. Instead of noise-amplifying complex division, the proposed method treats field variations as part of the signal model, thereby avoiding inherent statistical bias associated with regularization. Simulations and experiments showed that the proposed approach reliably quantified results in the presence of relatively large magnetic field distortion. Published 2015. This article is a U.S. Government work and is in the public domain in the USA.

Minimizing lipid signal bleed in brain (1) H chemical shift imaging by post-acquisition grid shifting

PURPOSE

Low spatial resolution in conventional 1H brain chemical shifting imaging (CSI) studies causes partial volume error (PVE) or signal "bleed" that is especially deleterious to voxels near the scalp. The standard spatial apodization approach adversely affects spatial resolution. Here, a novel automated post-processing strategy of partial volume correction employing grid shifting ("PANGS") is presented, which minimizes residual PVE without compromising spatial resolution.

METHODS

PANGS shifts the locations of the reconstruction coordinates in a designated region of image space-the scalp, to match the tissue "centers-of-mass" instead of the geometric centers of each voxel, by iteratively minimizing the PVE from the scalp into outside voxels. PANGS' performance was evaluated by numerical simulation, and in 3 Tesla 1H CSI human studies employing outer volume suppression and long echo times.

RESULTS

PANGS reduced lipid contamination of cortical spectra by up to 86% (54% on average). Metabolite maps exhibited significantly less lipid artifact than conventional and spatially-filtered CSI. All methods generated quantitatively identical spectral peak areas from central brain locations, but spatial filtering increased spectral linewidths and reduced spatial resolution.

CONCLUSION

PANGS significantly reduces lipid artifacts in 1H brain CSI spectra and metabolite maps, and improves metabolite detection in cortical regions without compromising resolution.

Is the macromolecule signal tissue-specific in healthy human brain? A (1)H MRS study at 7 Tesla in the occipital lobe

PURPOSE

The macromolecule signal plays a key role in the precision and the accuracy of the metabolite quantification in short-TE (1) H MR spectroscopy. Macromolecules have been reported at 1.5 Tesla (T) to depend on the cerebral studied region and to be age specific. As metabolite concentrations vary locally, information about the profile of the macromolecule signal in different tissues may be of crucial importance.

METHODS

The aim of this study was to investigate, at 7T for healthy subjects, the neurochemical profile differences provided by macromolecule signal measured in two different tissues in the occipital lobe, predominantly composed of white matter tissue or of grey matter tissue.

RESULTS

White matter-rich macromolecule signal was relatively lower than the gray matter-rich macromolecule signal from 1.5 to 1.8 ppm and from 2.3 to 2.5 ppm with mean difference over these regions of 7% and 12% (relative to the reference peak at 0.9 ppm), respectively. The neurochemical profiles, when using either of the two macromolecule signals, were similar for 11 reliably quantified metabolites (CRLB < 20%) with relatively small concentration differences (< 0.3 μmol/g), except Glu (± 0.8 μmol/g).

CONCLUSION

Given the small quantification differences, we conclude that a general macromolecule baseline provides a sufficiently accurate neurochemical profile in occipital lobe at 7T in healthy human brain.

Quantification of glutamate and glutamine using constant-time point-resolved spectroscopy at 3 T

Separate quantification of glutamate (Glu) and glutamine (Gln) using conventional MRS on clinical scanners is challenging. In previous work, constant-time point-resolved spectroscopy (CT-PRESS) was optimized at 3 T to detect Glu, but did not resolve Gln. To quantify Glu and Gln, a time-domain basis set was constructed taking into account metabolite T(2) relaxation times and dephasing from B(0) inhomogeneity. Metabolite concentrations were estimated by fitting the basis one-dimensional CT-PRESS diagonal magnitude spectra to the measured spectrum. This method was first validated using seven custom-built phantoms containing variable metabolite concentrations, and then applied to in vivo data acquired in rats exposed to vaporized ethanol and controls. Separate metabolite quantification revealed increased Gln after 16 weeks and increased Glu after 24 weeks of vaporized ethanol exposure in ethanol-treated compared with control rats. Without separate quantification, the signal from the combined resonances of Glu and Gln (Glx) showed an increase at both 16 and 24 weeks in ethanol-exposed rats, precluding the determination of the independent and differential contribution of each metabolite at each time.

Development of quantitative metabolomics for Pichia pastoris

Accurate, reliable and reproducible measurement of intracellular metabolite levels has become important for metabolic studies of microbial cell factories. A first critical step for metabolomic studies is the establishment of an adequate quenching and washing protocol, which ensures effective arrest of all metabolic activity and removal of extracellular metabolites, without causing leakage of metabolites from the cells. Five different procedures based on cold methanol quenching and cell separation by filtration were tested for metabolomics of Pichia pastoris regarding methanol content and temperature of the quenching solution as key parameters. Quantitative evaluation of these protocols was carried out through mass balance analysis, based on metabolite measurements in all sample fractions, those are whole broth, quenched and washed cells, culture filtrate and quenching and washing solution. Finally, the optimal method was used to study the time profiles of free amino acid and central carbon metabolism intermediates in glucose-limited chemostat cultures. Acceptable recoveries (>90%) were obtained for all quenching procedures tested. However, quenching at -27°C in 60% v/v methanol performed slightly better in terms of leakage minimization. We could demonstrate that five residence times under glucose limitation are enough to reach stable intracellular metabolite pools. Moreover, when comparing P. pastoris and S. cerevisiae metabolomes, under the same cultivation conditions, similar metabolite fingerprints were found in both yeasts, except for the lower glycolysis, where the levels of these metabolites in P. pastoris suggested an enzymatic capacity limitation in that part of the metabolism. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s11306-011-0308-1) contains supplementary material, which is available to authorized users.

rNMR: open source software for identifying and quantifying metabolites in NMR spectra

Despite the extensive use of nuclear magnetic resonance (NMR) for metabolomics, no publicly available tools have been designed for identifying and quantifying metabolites across multiple spectra. We introduce here a new open source software tool, rNMR, which 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. ROIs contain all of the underlying NMR data within user-defined chemical shift ranges. ROIs can be inspected visually, and they support robust quantification of NMR signals. 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. We describe how rNMR greatly reduces the time required for robust bioanalytical analysis of complex NMR data. An rNMR analysis yields a compact and transparent way of archiving the results from a metabolomics study so that it can be examined and evaluated by others. The rNMR website at http://rnmr.nmrfam.wisc.edu offers downloadable versions of rNMR for Windows, Macintosh, and Linux platforms along with extensive help documentation, instructional videos, and sample data.