A strategy for modelling dynamic responses in metabolic samples characterized by GC/MS

A multivariate strategy for studying the metabolic response over time in urinary GC/MS data is presented and exemplified by a study of drug-induced liver toxicity in the rat. The strategy includes the generation of representative data through hierarchical multivariate curve resolution (H-MCR), highlighting the importance of obtaining resolved metabolite profiles for quantification and identification of exogenous (drug related) and endogenous compounds (potential biomarkers) and for allowing reliable comparisons of multiple samples through multivariate projections. Batch modelling was used to monitor and characterize the normal (control) metabolic variation over time as well as to map the dynamic response of the drug treated animals in relation to the control. In this way treatment related metabolic responses over time could be detected and classified as being drug related or being potential biomarkers. In summary the proposed strategy uses the relatively high sensitivity and reproducibility of GC/MS in combination with efficient multivariate curve resolution and data analysis to discover individual markers of drug metabolism and drug toxicity. The presented results imply that the strategy can be of great value in drug toxicity studies for classifying metabolic markers in relation to their dynamic responses as well as for biomarker identification.     

Reliable profile detection in comparative metabolomics

A strategy for processing of metabolomic GC/MS data is presented. By considering the relationship between quantity and quality of detected profiles, representative data suitable for multiple sample comparisons and metabolite identification was generated. Design of experiments (DOE) and multivariate analysis was used to relate the changes in settings of the hierarchical multivariate curve resolution (H-MCR) method to quantitative and qualitative characteristics of the output data. These characteristics included number of resolved profiles, chromatographic quality in terms of reproducibility between analytical replicates, and spectral quality defined by purity and number of spectra containing structural information. The strategy was exemplified in two datasets: one containing 119 common metabolites, 18 of which were varied according to a DOE protocol; and one consisting of rat urine samples from control rats and rats exposed to a liver toxin. It was shown that the performance of the data processing could be optimized to produce metabolite data of high quality that allowed reliable sample comparisons and metabolite identification. This is a general approach applicable to any type of data processing where the important processing parameters are known and relevant output data characteristics can be defined. The results imply that this type of data quality optimization should be carried out as an integral step of data processing to ensure high quality data for further modeling and biological evaluation. Within metabolomics, this degree of optimization will be of high importance to generate models and extract biomarkers or biomarker patterns of biological or clinical relevance.     

Statistical multivariate metabolite profiling for aiding biomarker pattern detection and mechanistic interpretations in GC/MS based metabolomics

A strategy for robust and reliable mechanistic statistical modelling of metabolic responses in relation to drug induced toxicity is presented. The suggested approach addresses two cases commonly occurring within metabonomic toxicology studies, namely; 1) A pre-defined hypothesis about the biological mechanism exists and 2) No such hypothesis exists. GC/MS data from a liver toxicity study consisting of rat urine from control rats and rats exposed to a proprietary AstraZeneca compound were resolved by means of hierarchical multivariate curve resolution (H-MCR) generating 287 resolved chromatographic profiles with corresponding mass spectra. Filtering according to significance in relation to drug exposure rendered in 210 compound profiles, which were subjected to further statistical analysis following correction to account for the control variation over time. These dose related metabolite traces were then used as new observations in the subsequent analyses. For case 1, a multivariate approach, named Target Batch Analysis, based on OPLS regression was applied to correlate all metabolite traces to one or more key metabolites involved in the pre-defined hypothesis. For case 2, principal component analysis (PCA) was combined with hierarchical cluster analysis (HCA) to create a robust and interpretable framework for unbiased mechanistic screening. Both the Target Batch Analysis and the unbiased approach were cross-verified using the other method to ensure that the results did match in terms of detected metabolite traces. This was also the case, implying that this is a working concept for clustering of metabolites in relation to their toxicity induced dynamic profiles regardless if there is a pre-existing hypothesis or not. For each of the methods the detected metabolites were subjected to identification by means of data base comparison as well as verification in the raw data. The proposed strategy should be seen as a general approach for facilitating mechanistic modelling and interpretations in metabolomic studies.     

A multivariate screening strategy for investigating metabolic effects of strenuous physical exercise in human serum

A novel hypothesis-free multivariate screening methodology for the study of human exercise metabolism in blood serum is presented. Serum gas chromatography/time-of-flight mass spectrometry (GC/TOFMS) data was processed using hierarchical multivariate curve resolution (H-MCR), and orthogonal partial least-squares discriminant analysis (OPLS-DA) was used to model the systematic variation related to the acute effect of strenuous exercise. Potential metabolic biomarkers were identified using data base comparisons. Extensive validation was carried out including predictive H-MCR, 7-fold full cross-validation, and predictions for the OPLS-DA model, variable permutation for highlighting interesting metabolites, and pairwise t tests for examining the significance of metabolites. The concentration changes of potential biomarkers were verified in the raw GC/TOFMS data. In total, 420 potential metabolites were resolved in the serum samples. On the basis of the relative concentrations of the 420 resolved metabolites, a valid multivariate model for the difference between pre- and post-exercise subjects was obtained. A total of 34 metabolites were highlighted as potential biomarkers, all statistically significant (p < 8.1E-05). As an example, two potential markers were identified as glycerol and asparagine. The concentration changes for these two metabolites were also verified in the raw GC/TOFMS data.The strategy was shown to facilitate interpretation and validation of metabolic interactions in human serum as well as revealing the identity of potential markers for known or novel mechanisms of human exercise physiology. The multivariate way of addressing metabolism studies can help to increase the understanding of the integrative biology behind, as well as unravel new mechanistic explanations in relation to, exercise physiology.     

Magnetic Resonance Metabolic Profiling of Breast Cancer Tissue Obtained with Core Needle Biopsy for Predicting Pathologic Response to Neoadjuvant Chemotherapy

The purpose of this study was to determine whether metabolic profiling of core needle biopsy (CNB) samples using high-resolution magic angle spinning (HR-MAS) magnetic resonance spectroscopy (MRS) could be used for predicting pathologic response to neoadjuvant chemotherapy (NAC) in patients with locally advanced breast cancer. After institutional review board approval and informed consent were obtained, CNB tissue samples were collected from 37 malignant lesions in 37 patients before NAC treatment. The metabolic profiling of CNB samples were performed by HR-MAS MRS. Metabolic profiles were compared according to pathologic response to NAC using the Mann-Whitney test. Multivariate analysis was performed with orthogonal projections to latent structure-discriminant analysis (OPLS-DA). Various metabolites including choline-containing compounds were identified and quantified by HR-MAS MRS in all 37 breast cancer tissue samples obtained by CNB. In univariate analysis, the metabolite concentrations and metabolic ratios of CNB samples obtained with HR-MAS MRS were not significantly different between different pathologic response groups. However, there was a trend of lower levels of phosphocholine/creatine ratio and choline-containing metabolite concentrations in the pathologic complete response group compared to the non-pathologic complete response group. In multivariate analysis, the OPLS-DA models built with HR-MAS MR metabolic profiles showed visible discrimination between the pathologic response groups. This study showed OPLS-DA multivariate analysis using metabolic profiles of pretreatment CNB samples assessed by HR- MAS MRS may be used to predict pathologic response before NAC, although we did not identify the metabolite showing statistical significance in univariate analysis. Therefore, our preliminary results raise the necessity of further study on HR-MAS MR metabolic profiling of CNB samples for a large number of cancers.     

A metabolomics approach to characterise and identify various Mycobacterium species

We investigated the potential use of gas chromatography mass spectrometry (GC-MS), in combination with multivariate statistical data processing, to build a model for the classification of various tuberculosis (TB) causing, and non-TB Mycobacterium species, on the basis of their characteristic metabolite profiles. A modified Bligh-Dyer extraction procedure was used to extract lipid components from Mycobacterium tuberculosis, Mycobacterium avium, Mycobacterium bovis, and Mycobacterium kansasii cultures. Principle component analyses (PCA) of the GC-MS generated data showed a clear differentiation between all the Mycobacterium species tested. Subsequently, the 12 compounds best describing the variation between the sample groups were identified as potential metabolite markers, using PCA and partial least-squares discriminant analysis (PLS-DA). These metabolite markers were then used to build a discriminant classification model based on Bayes' theorem, in conjunction with multivariate kernel density estimation. This model subsequently correctly classified 2 "unknown" samples for each of the Mycobacterium species analysed, with probabilities ranging from 72 to 100%. Furthermore, Mycobacterium species classification could be achieved in less than 16 h, and the detection limit for this approach was 1×10(3)bacteriamL(-1). This study proves the capacity of a GC-MS, metabolomics pattern recognition approach for its possible use in TB diagnostics and disease characterisation.     

Revealing the metabonomic variation of EC using 1H-NMR spectroscopy and its association with the clinicopathological characteristics

In this study, (1)H NMR-based metabonomics has been applied to investigate esophageal cancer metabolic signatures in plasma and urine, purpose of assessing the diagnostic potential of this approach and gaining novel insights into esophageal cancer metabolism and systemic effects. Plasma and urine samples from esophageal cancer patients (n = 108) and a control healthy group (n = 40) were analyzed by Nuclear Magnetic Resonance (NMR) spectroscopy (600 MHz), and their spectral profiles subjected to Orthogonal Projections to Latent Structures (OPLS-DA) for multivariate statistics. Potential metabolic biomarkers were identified using data base comparisons used for examining the significance of metabolites. Compared to healthy controls, esophageal cancer plasma had higher levels of dimethylamine, α-glucose, β-glucose, citric acid, together with lower levels of Leucine, alanine, isoleucine, valine, glycoprotein, lactate, acetone, acetate, choline, isobutyrate, unsaturated lipid, VLDL, LDL, 1-methylhistidine; Compared to healthy controls, esophageal cancer urine had higher levels of Mannitol, glutamate, γ-propalanine, phenylalanine, acetate, allantoin, pyruvate, tyrosine, β-glucose and guinolinate, together with lower levels of N-acetylcysteine, valine, dihydrothymine, hippurate, methylguanidine, 1-methylnicotin- amide and Citric acid; Very good discrimination between cancer and control groups was achieved by multivariate modeling of plasma and urinary profiles. (1)H NMR-based metabolite profiling analysis was shown to be an effective approach to differentiating between patients with EC and healthy subjects. Good sensitivity and selectivity were shown by using the metabolite markers discovered to predict the classification of samples from the healthy control group and the patients with the disease. Plasma and urine metabolic profiling may have potential for early diagnosis of EC and may enhance our understanding of its mechanisms.     

Analytical strategies in metabonomics

To perform metabonomics investigations, it is necessary to generate comprehensive metabolite profiles for complex samples such as biofluids and tissue/tissue extracts. Analytical technologies that can be used to achieve this aim are constantly evolving, and new developments are changing the way in which such profiles' metabolite profiles can be generated. Here, the utility of various analytical techniques for global metabolite profiling, such as, e.g., 1H NMR, MS, HPLC-MS, and GC-MS, are explored and compared.     

Exploratory data analysis of DNA microarrays by multivariate curve resolution

In this work, the application of a multivariate curve resolution procedure based on alternating least squares optimization (MCR-ALS) for the analysis of data from DNA microarrays is proposed. For this purpose, simulated and publicly available experimental data sets have been analyzed. Application of MCR-ALS, a method that operates without the use of any training set, has enabled the resolution of the relevant information about different cancer lines classification using a set of few components; each of these defined by a sample and a pure gene expression profile. From resolved sample profiles, a classification of samples according to their origin is proposed. From the resolved pure gene expression profiles, a set of over- or underexpressed genes that could be related to the development of cancer diseases has been selected. Advantages of the MCR-ALS procedure in relation to other previously proposed procedures such as principal component analysis are discussed.     

Discrimination of Basal Cell Carcinoma from Normal Skin Tissue Using High-Resolution Magic Angle Spinning 1H NMR Spectroscopy

High-resolution magic angle spinning nuclear magnetic resonance (HR-MAS NMR) spectroscopy is a useful tool for investigating the metabolism of various cancers. Basal cell carcinoma (BCC) is the most common skin cancer. However, to our knowledge, data on metabolic profiling of BCC have not been reported in the literature. The objective of the present study was to investigate the metabolic profiling of cutaneous BCC using HR-MAS ¹H NMR spectroscopy. HR-MAS ¹H NMR spectroscopy was used to analyze the metabolite profile and metabolite intensity of histopathologically confirmed BCC tissues and normal skin tissue (NST) samples. The metabolic intensity normalized to the total spectral intensities in BCC and NST was compared, and multivariate analysis was performed with orthogonal partial least-squares discriminant analysis (OPLS-DA). P values < 0.05 were considered statistically significant. Univariate analysis revealed 9 metabolites that showed statistically significant difference between BCC and NST. In multivariate analysis, the OPLS-DA models built with the HR-MAS NMR metabolic profiles revealed a clear separation of BCC from NST. The receiver operating characteristic curve generated from the results revealed an excellent discrimination of BCC from NST with an area under the curve (AUC) value of 0.961. The present study demonstrated that the metabolite profile and metabolite intensity differ between BCC and NST, and that HR-MAS ¹H NMR spectroscopy can be a valuable tool in the diagnosis of BCC.     

Optimization of a sample preparation method for the metabolomic analysis of clinically relevant bacteria

Metabolomics, or metabolite profiling, is an approach that is increasingly used to study the metabolism of diverse organisms, elucidate biological processes and/or find characteristic biomarkers of physiological states. Here, we describe the optimization of a method for global metabolomic analysis of bacterial cultures, with the following steps. Cells are grown to log-phase, starting from an overnight culture and bacterial concentrations are monitored by measuring the optical density of the cultures at 600 nm. At an appropriate density they are harvested by centrifugation, washed three times with NaCl solution and metabolites are extracted using methanol and a bead-mill. Dried extracts are methoxymated and derivatized with methyltrimethylsilyltrifluoroacetamide (MSTFA) then analyzed using gas chromatography coupled to time-of-flight mass spectrometry (GC-MS/TOF). Finally, patterns in the acquired data are examined by multivariate data modeling. This method enabled us to obtain reproducible metabolite profiles of Yersinia pseudotuberculosis, with about 25% compound identification, based on comparison with entries in available GC-MS libraries. To assess the potential utility of the method for comparative analysis of other bacterial species we analyzed cultures of Pseudomonas aeruginosa, Salmonella typhimurium, Escherichia coli and methicillin-sensitive Staphylococcus aureus (MSSA). Multivariate analysis of the acquired data showed that it was possible to differentiate the species according to their metabolic profiles. Our results show that the presented procedure can be used for metabolomic analysis of a wide range of bacterial species of clinical interest.     

Metabolic profiling of human urine by CE-MS using a positively charged capillary coating and comparison with UPLC-MS

The potential of capillary electrophoresis time-of-flight mass spectrometry (CE-TOF-MS) using capillaries coated with a triple layer of polybrene-dextran sulfate-polybrene (PB-DS-PB) was evaluated for metabolic profiling of human urine. The method covers various metabolite classes and stable metabolic profiles of urine samples were obtained with favourable migration time repeatability (RSDs <1%). The PB-DS-PB CE-TOF-MS method was used for the analysis of human urine samples from 30 males and 30 females, which had been previously analyzed by reversed-phase UPLC-TOF-MS. Multivariate data analysis of the obtained data provided clear distinction between urine samples from males and females, emphasizing gender differences in metabolic signatures. Nearly all compounds responsible for male-female classification in CE-TOF-MS were different from the classifying compounds in UPLC-TOF-MS. Almost all compounds causing classification in the CE-TOF-MS study were highly polar and did not exhibit retention in the reversed-phase UPLC system. In addition, the CE-TOF-MS classifiers had an m/z value in the range of 50-150, whereas 95% of the classifying features found with UPLC-TOF-MS had an m/z value above 150. The CE-TOF-MS method therefore appears to be highly complementary to the UPLC-TOF-MS method providing classification based on different classes of metabolites.     

Urine metabolite profiling offers potential early diagnosis of oral cancer

Oral cancer is the sixth most common human cancer, with a high morbidity rate and an overall 5-year survival rate of less than 50%. It is often not diagnosed until it has reached an advanced stage. Therefore, an early diagnostic and stratification strategy is of great importance for oral cancer. In the current study, urine samples of patients with oral squamous cell carcinoma (OSCC, n = 37), oral leukoplakia (OLK, n = 32) and healthy subjects (n = 34) were analyzed by gas chromatography-mass spectrometry (GC–MS). Using multivariate statistical analysis, the urinary metabolite profiles of OSCC, OLK and healthy control samples can be clearly discriminated and a panel of differentially expressed metabolites was obtained. Metabolites, valine and 6-hydroxynicotic acid, in combination yielded an accuracy of 98.9%, sensitivity of 94.4%, specificity of 91.4%, and positive predictive value of 91.9% in distinguishing OSCC from the controls. The combination of three differential metabolites, 6-hydroxynicotic acid, cysteine, and tyrosine, was able to discriminate between OSCC and OLK with an accuracy of 92.7%, sensitivity of 85.0%, specificity of 89.7%, and positive predictive value of 91.9%. This study demonstrated that the metabolite markers derived from this urinary metabolite profiling approach may hold promise as a diagnostic tool for early stage OSCC and its differentiation from other oral conditions.     

Gas chromatography mass spectrometry-based metabolite profiling in plants

The concept of metabolite profiling has been around for decades, but technical innovations are now enabling it to be carried out on a large scale with respect to the number of both metabolites measured and experiments carried out. Here we provide a detailed protocol for gas chromatography mass spectrometry (GC-MS)-based metabolite profiling that offers a good balance of sensitivity and reliability, being considerably more sensitive than NMR and more robust than liquid chromatography-linked mass spectrometry. We summarize all steps from collecting plant material and sample handling to derivatization procedures, instrumentation settings and evaluating the resultant chromatograms. We also define the contribution of GC-MS-based metabolite profiling to the fields of diagnostics, gene annotation and systems biology. Using the protocol described here facilitates routine determination of the relative levels of 300-500 analytes of polar and nonpolar extracts in approximately 400 experimental samples per week per machine.     

The potential of metabolite profiling as a selection tool for genotype discrimination in Populus

Differences between wild-type Populus tremulaxalba and two transgenic lines with modified lignin monomer composition, were interrogated using metabolic profiling. Analysis of metabolite abundance data by GC-MS, coupled with principal components analysis (PCA), successfully differentiated between lines that had distinct phenotypes, whether samples were taken from the cambial zone or non-lignifying suspension tissue cultures. Interestingly, the GC-MS analysis detected relatively few phenolic metabolites in cambial extracts, although a single metabolite associated with the differentiation between lines was directly related to the phenylpropanoid pathway or other down-stream aspects of lignin biosynthesis. In fact, carbohydrates, which have only an indirect relationship with the modified lignin monomer composition, featured strongly in the line-differentiating aspects of the statistical analysis. Traditional HPLC analysis was employed to verify the GC-MS data. These findings demonstrate that metabolic traits can be dissected reliably and accurately by metabolomic analyses, enabling the discrimination of individual genotypes of the same tree species that exhibit marked differences in industrially relevant wood traits. Furthermore, this validates the potential of using metabolite profiling techniques for marker generation in the context of plant/tree breeding for industrial applications.     

A GC-MS metabolic profiling study of plasma samples from mice on low- and high-fat diets

Metabolic profiling of biofluids, based on the quantitative analysis of the concentration profile of their free low molecular mass metabolites, has been playing increasing role employed as a means to gain understanding of the progression of metabolic disorders, including obesity. Chromatographic methods coupled with mass spectrometry have been established as a strategy for metabolic profiling. Among these, GC-MS, targeting mainly the primary metabolism intermediates, offers high sensitivity, good peak resolution and extensive databases. However, the derivatization step required for many involatile metabolites necessitates specific data validation, normalization and analysis protocols to ensure accurate and reproducible performance. In this study, the GC-MS metabolic profiles of plasma samples from mice maintained on 12- or 15-month long low (10 kcal%) or high (60 kcal%) fat diets were obtained. The profiles of the trimethylsilyl(TMS)-methoxime(MeOx) derivatives of the free polar metabolites were acquired through GC-(ion trap)MS, using [U-(13)C]-glucose as the internal standard. After the application of a recently developed data correction and normalization/filtering protocol for GC-MS metabolomic datasets, the profiles of 48 out of the 77 detected metabolites were used in multivariate statistical analysis. Data mining suggested a decrease in the activity of the energy metabolism with age. In addition, the metabolic profiles indicated the presence of subpopulations with different physiology within the high- and low-fat diet mice, which correlated well with the difference in body weight among the animals and current knowledge about hyperglycemic conditions.     

Global urinary metabolic profiling procedures using gas chromatography-mass spectrometry

The role of urinary metabolic profiling in systems biology research is expanding. This is because of the use of this technology for clinical diagnostic and mechanistic studies and for the development of new personalized health care and molecular epidemiology (population) studies. The methodologies commonly used for metabolic profiling are NMR spectroscopy, liquid chromatography mass spectrometry (LC/MS) and gas chromatography-mass spectrometry (GC/MS). In this protocol, we describe urine collection and storage, GC/MS and data preprocessing methods, chemometric data analysis and urinary marker metabolite identification. Results obtained using GC/MS are complementary to NMR and LC/MS. Sample preparation for GC/MS analysis involves the depletion of urea via treatment with urease, protein precipitation with methanol, and trimethylsilyl derivatization. The protocol described here facilitates the metabolic profiling of ～400-600 metabolites in 120 urine samples per week.     

Reference intervals of urinary steroid metabolites using gas chromatography-mass spectrometry in Chinese adults

BACKGROUND: Urinary steroid profiling by GC or GC-MS are established clinical tools to complement other biochemical tests in the diagnosis and investigation of a wide range of adrenocortical disorders, but normative data on adults using the more specific GC-MS are lacking. Our objective was to set up the reference intervals of commonly detected urinary steroid metabolites as well as marker metabolites seen in disease states. METHOD: Apparently healthy adult Chinese males and females were recruited by completing health questionnaires. A 24-h urine specimen was collected from all the participants for urinary steroid profiling by GC-MS in cyclic scan mode. The analyzer was calibrated by using authentic steroid standards. Statistical methods recommended by the National Committee for Clinical Laboratory Standards were followed for setting up the reference intervals of various steroid metabolites. After outliers were excluded, the data were tested for the necessity to partition into sex-, menopausal status- and age-specific reference intervals. RESULTS: 83 males and 89 females were recruited for the study. Necessity to partition into sex-specific reference intervals was demonstrated for almost all steroid metabolites. Menopausal status and age also had a significant impact on steroid metabolite excretion, making separate reference intervals necessary. CONCLUSIONS: We have set up the normative data on the levels of urinary steroid metabolite excretion in Chinese adults for future reference in patient management and research in steroid metabolism.     

A novel approach for nontargeted data analysis for metabolomics. Large-scale profiling of tomato fruit volatiles

To take full advantage of the power of functional genomics technologies and in particular those for metabolomics, both the analytical approach and the strategy chosen for data analysis need to be as unbiased and comprehensive as possible. Existing approaches to analyze metabolomic data still do not allow a fast and unbiased comparative analysis of the metabolic composition of the hundreds of genotypes that are often the target of modern investigations. We have now developed a novel strategy to analyze such metabolomic data. This approach consists of (1) full mass spectral alignment of gas chromatography (GC)-mass spectrometry (MS) metabolic profiles using the MetAlign software package, (2) followed by multivariate comparative analysis of metabolic phenotypes at the level of individual molecular fragments, and (3) multivariate mass spectral reconstruction, a method allowing metabolite discrimination, recognition, and identification. This approach has allowed a fast and unbiased comparative multivariate analysis of the volatile metabolite composition of ripe fruits of 94 tomato (Lycopersicon esculentum Mill.) genotypes, based on intensity patterns of >20,000 individual molecular fragments throughout 198 GC-MS datasets. Variation in metabolite composition, both between- and within-fruit types, was found and the discriminative metabolites were revealed. In the entire genotype set, a total of 322 different compounds could be distinguished using multivariate mass spectral reconstruction. A hierarchical cluster analysis of these metabolites resulted in clustering of structurally related metabolites derived from the same biochemical precursors. The approach chosen will further enhance the comprehensiveness of GC-MS-based metabolomics approaches and will therefore prove a useful addition to nontargeted functional genomics research.     

Metabolite profiling for analysis of yeast stress response during very high gravity ethanol fermentations

A laboratory strain and an industrial strain of Saccharomyces cerevisiae were grown at high substrate concentration, so-called very high gravity (VHG) fermentation. Simultaneous saccharification and fermentation (SSF) was applied in a batch process using 280 g/L maltodextrin as carbon source. It was shown that known ethanol and osmotic stress responses such as decreased growth rate, lower viability, higher energy consumption, and intracellular trehalose accumulation occur in VHG SSF for both strains when compared with standard laboratory medium (20 g/L glucose). The laboratory strain was the most affected. GC-MS metabolite profiling was applied for assessing the yeast stress response influence on cellular metabolism. It was found that metabolite profiles originating from different strains and/or fermentation conditions were unique and could be distinguished with the help of multivariate data analysis. Several differences in the metabolic responses to stressing conditions were revealed, particularly the increased energy consumption of stressed cells was also reflected in increased intracellular concentrations of pyruvate and related metabolites.