Hierarchical Alignment and Full Resolution Pattern Recognition of 2D NMR Spectra: Application to Nematode Chemical Ecology

Nuclear magnetic resonance (NMR) is the most widely used nondestructive technique in analytical chemistry. In recent years, it has been applied to metabolic profiling due to its high reproducibility, capacity for relative and absolute quantification, atomic resolution, and ability to detect a broad range of compounds in an untargeted manner. While one-dimensional (1D) (1)H NMR experiments are popular in metabolic profiling due to their simplicity and fast acquisition times, two-dimensional (2D) NMR spectra offer increased spectral resolution as well as atomic correlations, which aid in the assignment of known small molecules and the structural elucidation of novel compounds. Given the small number of statistical analysis methods for 2D NMR spectra, we developed a new approach for the analysis, information recovery, and display of 2D NMR spectral data. We present a native 2D peak alignment algorithm we term HATS, for hierarchical alignment of two-dimensional spectra, enabling pattern recognition (PR) using full-resolution spectra. Principle component analysis (PCA) and partial least squares (PLS) regression of full resolution total correlation spectroscopy (TOCSY) spectra greatly aid the assignment and interpretation of statistical pattern recognition results by producing back-scaled loading plots that look like traditional TOCSY spectra but incorporate qualitative and quantitative biological information of the resonances. The HATS-PR methodology is demonstrated here using multiple 2D TOCSY spectra of the exudates from two nematode species: Pristionchus pacificus and Panagrellus redivivus. We show the utility of this integrated approach with the rapid, semiautomated assignment of small molecules differentiating the two species and the identification of spectral regions suggesting the presence of species-specific compounds. These results demonstrate that the combination of 2D NMR spectra with full-resolution statistical analysis provides a platform for chemical and biological studies in cellular biochemistry, metabolomics, and chemical ecology.     

Use of semiselective TOCSY and the pearson correlation for the metabonomic analysis of biofluid mixtures: application to urine

The authors recently proposed an approach to the metabonomic analysis of biofluid mixtures based on the use of the selective TOCSY experiment (Sandusky, P.; Raftery, D. Anal. Chem. 2005, 77, 2455). This method has some significant advantages over standard metabonomic analysis. However, when analyzing overlapped components, the selective TOCSY method can suffer from the relatively high likelihood of simultaneous excitation of several spin systems at once. This multiple excitation can cause problems both with the purity of the individual TOCSY peaks observed and with their assignment into specific spin systems. To address this problem, the possibility of using a more selective excitation is initially explored. Unfortunately, in most cases, greater spin system selectivity can only be gained at the expense of sensitivity. This is obviously an unacceptable tradeoff when dealing with biofluid samples. However, the application of the Pearson product moment correlation to the TOCSY peak integral intensities provides a test for individual TOCSY peak purity and allows for the assignment of the peaks into spin systems. The specific application of this two-stage "semiselective" TOCSY method to rat and human urine is presented. Significantly, it is also demonstrated that the use of semiselective TOCSY spectra as data inputs for PCA calculations provides a more sensitive and reliable method of distinguishing small differences in biofluid composition than the standard metabonomic approach using complete 1D proton NMR spectra of urine samples.     

Plasma phospholipid metabolic profiling and biomarkers of type 2 diabetes mellitus based on high-performance liquid chromatography/electrospray mass spectrometry and multivariate statistical analysis

Liquid chromatography/mass spectrometry (LC/MS) followed by multivariate statistical analysis has been successfully applied to the plasma phospholipids metabolic profiling in type 2 diabetes mellitus (DM-2). Principal components analysis and partial least-squares discriminant analysis (PLS-DA) models were tested and compared in class separation between the DM2 and control. The application of an orthogonal signal correction filtered model highly improved the class distinction and predictive power of PLS-DA models. Additionally, unit variance scaling was also tested. With this methodology, it was possible not only to differentiate the DM2 from the control but also to discover and identify the potential biomarkers with LC/MS/MS. The proposed method shows that LC/MS combining with multivariate statistical analysis is a complement or an alternative to NMR for metabonomics applications.     

Molecular factor analysis applied to collections of NMR spectra

It is often useful to identify and quantify mixture components by analyzing collections of NMR spectra. Such collections arise in metabonomics and many other applications. Many mixtures studied by NMR can contain hundreds of compounds, and it is challenging to analyze the resulting complex spectra. We have approached the problem of separating signals from different molecules in complex mixtures by using self-modeling curve resolution as implemented by the alternating least-squares algorithm. Alternating least squares uses nonnegativity criteria to generate spectra and concentrations from a collection of mixture spectra. Compared to previous applications of alternating least squares, NMR spectra of complex mixtures possess unique features, such as large numbers of components and sample-to-sample variability in peak positions. To deal with these features, we developed a set of data preprocessing methods, and we made modifications to the alternating least-squares algorithm. We use the term "molecular factor analysis" to refer to the preprocessing and modified alternating least-squares methods. Molecular factor analysis was tested using an artificial data set and spectra from a metabonomics study. The results show that the tools can extract valuable information on sample composition from sets of NMR spectra.     

Web server based complex mixture analysis by NMR

Comprehensive metabolite identification and quantification of complex biological mixtures are central aspects of metabolomics. NMR shows excellent promise for these tasks. An automated fingerprinting strategy is presented, termed COLMAR query, which screens NMR chemical shift lists or raw 1D NMR cross sections taken from covariance total correlation spectroscopy (TOCSY) spectra or other multidimensional NMR spectra against an NMR spectral database. Cross peaks are selected using local clustering to avoid ambiguities between chemical shifts and scalar J-coupling effects. With the use of three different algorithms, the corresponding chemical shift list is then screened against chemical shift lists extracted from 1D spectra of a NMR database. The resulting query scores produced by forward assignment, reverse assignment, and bipartite weighted-matching algorithms are combined into a consensus score, which provides a robust means for identifying the correct compound. The approach is demonstrated for a metabolite model mixture that is screened against the metabolomics BioMagResDatabank (BMRB). This NMR-based compound identification approach has been implemented in a public Web server that allows the efficient analysis of a wide range of metabolite mixtures.     

Magic angle spinning NMR and 1H-31P heteronuclear statistical total correlation spectroscopy of intact human gut biopsies

Previously we have demonstrated the use of 1H magic angle spinning (MAS) NMR spectroscopy for the topographical variations in functional metabolic signatures of intact human intestinal biopsy samples. Here we have analyzed a series of MAS 1H NMR spectra (spin-echo, one-dimensional, and diffusion-edited) and 31P-{1H} spectra and focused on analyzing the enhancement of information recovery by use of the statistical total correlation spectroscopy (STOCSY) method. We have applied a heterospectroscopic cross-examination performed on the same samples and between 1H and 31P-{1H} spectra (heteronuclear STOCSY) to recover latent metabolic information. We show that heterospectroscopic correlation can give new information on the molecular compartmentation of metabolites in intact tissues, including the statistical "isolation" of a phospholipid/triglyceride vesicle pool in intact tissue. The application of 31P-1H HET-STOCSY allowed the cross-assignment of major 31P signals to their equivalent 1H NMR spectra, e.g., for phosphorylcholine and phosphorylethanolamine. We also show pathway correlations, e.g., the ascorbate-glutathione pathway, in the STOCSY analysis of intact tissue spectra. These 31P-1H HET-STOCSY spectra also showed different topographical regions, particular for minor signals in different tissue microenvironments. This approach could be extended to allow the detection of altered distributions within metabolic subcompartments as well as conventional metabonomics concentration-based diagnostics.     

Statistical total correlation spectroscopy: an exploratory approach for latent biomarker identification from metabolic 1H NMR data sets

We describe here the implementation of the statistical total correlation spectroscopy (STOCSY) analysis method for aiding the identification of potential biomarker molecules in metabonomic studies based on NMR spectroscopic data. STOCSY takes advantage of the multicollinearity of the intensity variables in a set of spectra (in this case 1H NMR spectra) to generate a pseudo-two-dimensional NMR spectrum that displays the correlation among the intensities of the various peaks across the whole sample. This method is not limited to the usual connectivities that are deducible from more standard two-dimensional NMR spectroscopic methods, such as TOCSY. Moreover, two or more molecules involved in the same pathway can also present high intermolecular correlations because of biological covariance or can even be anticorrelated. This combination of STOCSY with supervised pattern recognition and particularly orthogonal projection on latent structure-discriminant analysis (O-PLS-DA) offers a new powerful framework for analysis of metabonomic data. In a first step O-PLS-DA extracts the part of NMR spectra related to discrimination. This information is then cross-combined with the STOCSY results to help identify the molecules responsible for the metabolic variation. To illustrate the applicability of the method, it has been applied to 1H NMR spectra of urine from a metabonomic study of a model of insulin resistance based on the administration of a carbohydrate diet to three different mice strains (C57BL/6Oxjr, BALB/cOxjr, and 129S6/SvEvOxjr) in which a series of metabolites of biological importance can be conclusively assigned and identified by use of the STOCSY approach.     

Analysis of multiple samples using multiplex sample NMR: selective excitation and chemical shift imaging approaches

Two improved approaches for the rapid analysis of multiple samples using multiplex sample NMR are described. In the first approach, frequency-selective 90 degrees radio frequency pulses and large pulsed field gradients are applied to excite and detect multiple samples in rapid succession. This method is advantageous for samples with relatively long longitudinal (T1) relaxation times. In the second approach, chemical shift imaging is applied to acquire both the spectral and spatial information of multiple samples simultaneously. Chemical shift imaging is more time-consuming than selective excitation; however, it is advantageous for detecting samples with short T1's and for signal averaging. Both approaches demonstrate the potential of multiplex sample NMR for carrying out high-throughput NMR detection.     

Evaluation of the orthogonal projection on latent structure model limitations caused by chemical shift variability and improved visualization of biomarker changes in 1H NMR spectroscopic metabonomic studies

In general, applications of metabonomics using biofluid NMR spectroscopic analysis for probing abnormal biochemical profiles in disease or due to toxicity have all relied on the use of chemometric techniques for sample classification. However, the well-known variability of some chemical shifts in 1H NMR spectra of biofluids due to environmental differences such as pH variation, when coupled with the large number of variables in such spectra, has led to the situation where it is necessary to reduce the size of the spectra or to attempt to align the shifting peaks, to get more robust and interpretable chemometric models. Here, a new approach that avoids this problem is demonstrated and shows that, moreover, inclusion of variable peak position data can be beneficial and can lead to useful biochemical information. The interpretation of chemometric models using combined back-scaled loading plots and variable weights demonstrates that this peak position variation can be handled successfully and also often provides additional information on the physicochemical variations in metabonomic data sets.     

Virtual chromatographic resolution enhancement in cryoflow LC-NMR experiments via statistical total correlation spectroscopy

A new approach to enhancing information recovery from cryogenic probe "on-flow" LC-NMR spectroscopic analyses of complex biological mixtures is demonstrated using a variation on the statistical total correlation spectroscopy (STOCSY) method. Cryoflow probe technology enables sensitive and efficient NMR detection of metabolites on-flow, and the rapid spectral scanning allows multiple spectra to be collected over chromatographic peaks containing several species with similar, but nonidentical, retention times. This enables 1H NMR signal connectivities between close-eluting metabolites to be identified resulting in a "virtual" chromatographic resolution enhancement visualized directly in the NMR spectral projection. We demonstrate the applicability of the approach for structure assignment of drug and endogenous metabolites in urine. This approach is of wide general applicability to any complex mixture analysis problem involving chromatographic peak overlap and with particular application in metabolomics and metabonomics.     

Strategy for automated analysis of dynamic metabolic mixtures by NMR. Application to an insect venom

Elucidation of the composition of chemical-biological samples is a main focus of systems biology and metabolomics. Due to the inherent complexity of these mixtures, reliable, efficient, and potentially automatable methods are needed to identify the underlying metabolites and natural products. Because of its rich chemical information content, nuclear magnetic resonance (NMR) spectroscopy has a unique potential for this task. Here we present a generalization and application of a recently introduced NMR data collection, processing, and analysis strategy that circumvents the need for extensive purification and hyphenation prior to analysis. It uses covariance TOCSY NMR spectra measured on a 1-mm high-temperature cryogenic probe that are analyzed by a spectral trace clustering algorithm yielding 1D NMR spectra of the individual components for their unambiguous identification. The method is demonstrated on a metabolic model mixture and is then applied to the unpurified venom mixture of an individual walking stick insect that contains several slowly interconverting and closely related metabolites.     

Quantification and statistical significance analysis of group separation in NMR-based metabonomics studies

Currently, no standard metrics are used to quantify cluster separation in PCA or PLS-DA scores plots for metabonomics studies or to determine if cluster separation is statistically significant. Lack of such measures makes it virtually impossible to compare independent or inter-laboratory studies and can lead to confusion in the metabonomics literature when authors putatively identify metabolites distinguishing classes of samples based on visual and qualitative inspection of scores plots that exhibit marginal separation. While previous papers have addressed quantification of cluster separation in PCA scores plots, none have advocated routine use of a quantitative measure of separation that is supported by a standard and rigorous assessment of whether or not the cluster separation is statistically significant. Here quantification and statistical significance of separation of group centroids in PCA and PLS-DA scores plots are considered. The Mahalanobis distance is used to quantify the distance between group centroids, and the two-sample Hotelling's T² test is computed for the data, related to an F-statistic, and then an F-test is applied to determine if the cluster separation is statistically significant. We demonstrate the value of this approach using four datasets containing various degrees of separation, ranging from groups that had no apparent visual cluster separation to groups that had no visual cluster overlap. Widespread adoption of such concrete metrics to quantify and evaluate the statistical significance of PCA and PLS-DA cluster separation would help standardize reporting of metabonomics data.     

Adaptive reshaping of objects in (multiparameter) Hilbert space for enhanced detection and classification: an application of receiver operating curve statistics to laser-based mass spectroscopy

We propose a new approach to the classical detection problem of discrimination of a true signal of interest from an interferent signal, which may be applied to the area of chemical sensing. We show that the detection performance, as quantified by the receiver operating curve (ROC), can be substantially improved when the signal is represented by a multicomponent data set that is actively manipulated by means of a shaped laser probe pulse. In this case, the signal sought (agent) and the interfering signal (interferent) are visualized by vectors in a multidimensional detection space. Separation of these vectors can be achieved by adaptive modification of a probing laser pulse to actively manipulate the Hamiltonian of the agent and interferent. We demonstrate one implementation of the concept of adaptive rotation of signal vectors to chemical agent detection by means of strong-field time-of-flight mass spectrometry.     

稳健统计与格拉布斯准则在能力验证结果分析中的应用

分别运用稳健统计（四分位和迭代）法、格拉布斯准则剔除离群值后用经典统计法对热轧带肋钢筋的6组数据进行了统计分析。结果表明:数据组中Z比分值大于5的数据,为离群值。数据组总体服从正态分布或接近正态分布的,3种统计方法结果基本吻合;不存在离群数据的情况下,建议采用稳健统计（四分位或迭代）法进行统计;存在离群数据的情况下,应采用格拉布斯准则剔除离群值后用经典统计法进行统计。对于数据明显偏离正态分布的,不存在离群数据的情况下,建议采用迭代稳健统计法进行统计;存在离群数据的情况下,应采用格拉布斯准则剔除离群值后用经典统计法进行统计。     

Statistical heterospectroscopy, an approach to the integrated analysis of NMR and UPLC-MS data sets: application in metabonomic toxicology studies

Statistical heterospectroscopy (SHY) is a new statistical paradigm for the coanalysis of multispectroscopic data sets acquired on multiple samples. This method operates through the analysis of the intrinsic covariance between signal intensities in the same and related molecules measured by different techniques across cohorts of samples. The potential of SHY is illustrated using both 600-MHz 1H NMR and UPLC-TOFMS data obtained from control rat urine samples (n = 54) and from a corresponding hydrazine-treated group (n = 58). We show that direct cross-correlation of spectral parameters, viz. chemical shifts from NMR and m/z data from MS, is readily achievable for a variety of metabolites, which leads to improved efficiency of molecular biomarker identification. In addition to structure, higher level biological information can be obtained on metabolic pathway activity and connectivities by examination of different levels of the NMR to MS correlation and anticorrelation matrixes. The SHY approach is of general applicability to complex mixture analysis, if two or more independent spectroscopic data sets are available for any sample cohort. Biological applications of SHY as demonstrated here show promise as a new systems biology tool for biomarker recovery.     

Chemometric strategies to assess metabonomic imprinting of food habits in epidemiological studies

Application of metabonomics to nutritional sciences, also termed as nutrimetabonomics, offers the possibility to measure metabolic responses associated with the consumption of specific nutrients and foods. As dietary differences generally only lead to subtle metabolic changes, measuring diet associated metabolic phenotypes is a challenge, and also an opportunity to develop and test new chemometric strategies that can highlight metabolic information in relation to different dietary habits. While multivariate statistical techniques have long been used to analyse dietary data from diet records and questionnaires, to date no attempt has been made to link dietary patterns with metabolic profiles. Using a three-step strategy, it was possible to merge ¹H NMR plasma metabolic profile data with specific dietary patterns as assessed by Principal Component Analysis (PCA) and Partial Least Squares-Discriminant Analysis (PLS-DA). Five dietary patterns (energy intake, plant versus animal based diet, “traditional diet” versus sugar-rich diet, “traditional” versus “modern” diets, and consumption of skim versus whole dairy products) were found by applying PCA to the food frequency questionnaire data which explained 50% of the variation. Metabolic phenotypes associated with these dietary patterns were obtained by PLS-DA and were mainly based on differences in lipids and amino acid profiles in plasma. This new approach to assess relationships between dietary intake and metabolic profiling data will allow greater steps to be made in merging nutritional epidemiology with metabonomics.     

Determination of the composition of acetylglycerol mixtures by (1)H NMR followed by GC investigation

Commercially available partly acetylated glycerols (mono- and diacetins) are a mixture of glycerol, 1- and 2-acetylglycerol, 1,2- and 1,3-diacetylglycerol, and triacetin. No exact analysis method is available. Comparisons of (1)H NMR measurements obtained using deuterated dimethyl sulfoxide (DMSO-d6) and DMSO-d6/15% D2O are sufficient to identify and determine quickly all the components. Advantages compared with the commonly used NMR solvent CDCl3 for fatty acid glycerides include the solubility of all the components and a highly informative OH signal pattern in a region between 4.36 and 5.26 ppm almost free of other signals. 2-Acetylglycerol (2-acetin) is shown to disproportionate even at 50 degrees C into a mixture of glycerol and acetylglycerol thereby making it difficult to quantify by liquid chromatography (LC) and gas chromatography (GC) methods. Complete (1)H chemical shift data for all five components allow for the identification of the components in the mixture and thus the determination of the composition. The NMR method with DMSO-d6 as solvent was used for acetins, propionins, and butyrins. Semipreparative high-performance liquid chromatography (HPLC) on an RP18 column led to moderately pure 1-monoacetin and a mixture of diacylated species. Electron impact mass spectra show for all the components very characteristic fragmentation patterns with loss of AcOCH2 radicals, in contrast to the well-known pattern of longer-chained fatty acid derivatives that show preferred first-step loss of acyl-O radicals.     

Practical approach for the identification and isomer elucidation of biomarkers detected in a metabonomic study for the discovery of individuals at risk for diabetes by integrating the chromatographic and mass spectrometric information

Sensitive and high-resolution chromatographic-driven metabonomomics studies experienced major growth with the aid of new analytical technologies and bioinformatics software packages. Hence, data collections by LC-MS and data analyses by multivariate statistical methods are by far the most straightforward steps, and the detection of biomarker candidates can easily be achieved. However, the unequivocal identification of the detected metabolite candidates, including isomer elucidation, is still a crux of current metabonomics studies. Here we present a comprehensive analytical strategy for the elucidation of the molecular structure of metabolite biomarkers detected in a metabonomics study, exemplified analyzing spot urine of a cohort of healthy, insulin sensitive subjects and clinically well characterized prediabetic, insulin resistant individuals. An integrated approach of LC-MS fingerprinting, multivariate statistic analysis, LC-MSn experiments, micro preparation, FTICR-MS, GC retention index, database search, and generation of an isotope labeled standard was applied. Overall, we could demonstrate the efficiency of our analytical approach by the unambiguous elucidation of the molecular structure of an isomeric biomarker candidate detected in a complex human biofluid. The proposed strategy is a powerful new analytical tool, which will allow the definite identification of physiologically important molecules in metabonomics studies from basic biochemistry to clinical biomarker discovery.