Untargeted metabolic profiling identifies altered serum metabolites of type 2 diabetes mellitus in a prospective, nested case control study.

BACKGROUND: Application of metabolite profiling could expand the etiological knowledge of type 2 diabetes mellitus (T2D). However, few prospective studies apply broad untargeted metabolite profiling to reveal the comprehensive metabolic alterations preceding the onset of T2D. METHODS: We applied untargeted metabolite profiling in serum samples obtained from the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam cohort comprising 300 individuals who developed T2D after a median follow-up time of 6 years and 300 matched controls. For that purpose, we used ultraperformance LC-MS with a protocol specifically designed for large-scale metabolomics studies with regard to robustness and repeatability. After multivariate classification to select metabolites with the strongest contribution to disease classification, we applied multivariable-adjusted conditional logistic regression to assess the association of these metabolites with T2D. RESULTS: Among several alterations in lipid metabolism, there was an inverse association with T2D for metabolites chemically annotated as lysophosphatidylcholine(dm16:0) and phosphatidylcholine(O-20:0/O-20:0). Hexose sugars were positively associated with T2D, whereas higher concentrations of a sugar alcohol and a deoxyhexose sugar reduced the odds of diabetes by approximately 60% and 70%, respectively. Furthermore, there was suggestive evidence for a positive association of the circulating purine nucleotide isopentenyladenosine-5'-monophosphate with incident T2D. CONCLUSIONS: This study constitutes one of the largest metabolite profiling approaches of T2D biomarkers in a prospective study population. The findings might help generate new hypotheses about diabetes etiology and develop further targeted studies of a smaller number of potentially important metabolites.

A label-free selected reaction monitoring workflow identifies a subset of pregnancy specific glycoproteins as potential predictive markers of early-onset pre-eclampsia.

Pre-eclampsia (PE) is a serious complication of pregnancy with potentially life threatening consequences for both mother and baby. Presently there is no test with the required performance to predict which healthy first-time mothers will go on to develop PE. The high specificity, sensitivity, and multiplexed nature of selected reaction monitoring holds great potential as a tool for the verification and validation of putative candidate biomarkersfor disease states. Realization of this potential involves establishing a high throughput, cost effective, reproducible sample preparation workflow. We have developed a semi-automated HPLC-based sample preparation workflow before a label-free selected reaction monitoring approach. This workflow has been applied to the search for novel predictive biomarkers for PE. To discover novel candidate biomarkers for PE, we used isobaric tagging to identify several potential biomarker proteins in plasma obtained at 15 weeks gestation from nulliparous women who later developed PE compared with pregnant women who remained healthy. Such a study generates a number of "candidate" biomarkers that require further testing in larger patient cohorts. As proof-of-principle, two of these proteins were taken forward for verification in a 100 women (58 PE, 42 controls) using label-free SRM. We obtained reproducible protein quantitation across the 100 samples and demonstrated significant changes in protein levels, even with as little as 20% change in protein concentration. The SRM data correlated with a commercial ELISA, suggesting that this is a robust workflow suitable for rapid, affordable, label-free verification of which candidate biomarkers should be taken forward for thorough investigation. A subset of pregnancy-specific glycoproteins (PSGs) had value as novel predictive markers for PE.

Effects of platelet-rich plasma on subchondral bone marrow edema and biomarkers in synovial fluid of knee osteoarthritis.

BACKGROUND: The aim of the study was to investigate the effect of platelet-rich plasma (PRP) on subchondral bone marrow edema (BME) and the level of biomarkers in synovial fluid of the knee osteoarthritis. METHODS: Eighty-one patients with symptomatic knee osteoarthritis were randomly divided into two groups according to the number of inpatients. Forty-five cases were treated with intra-articular injection of PRP (PRP group), 36 cases were treated with sodium hyaluronate (SH group), and the clinical effects were evaluated using the Visual Analogue Scale (VAS) and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) scores. The changes of subchondral BME were assessed by magnetic resonance imaging (MRI) before and after treatment. The levels of TNFα, IL-6, MCP-1, MMP-1, MMP-3, and MMP-9 in synovial fluid were also detected. RESULTS: All the patients completed the corresponding treatment and were followed up for 12 months without serious complications. After the treatment, the VAS and WOMAC scores of the two groups were significantly decreased, and the difference was statistically significant at different time points (P < 0.05). The VAS and WOMAC scores of the PRP group were better than those of the SH group (P < 0.05). MRI showed that the subchondral bone edema of the two groups were reduced in varying degrees, and the reduction was more noticeable in the PRP group (P < 0.05). The levels of TNFα, IL-6, MCP-1, MMP-1, MMP-3, and MMP-9 in two groups were decreased, and the difference was statistically significant at different time points (P < 0.05). However, the levels of TNFα, IL-6, MCP-1, MMP-1, MMP-3, and MMP-9 in the PRP group were significantly lower than those in the SH group (P < 0.05). CONCLUSIONS: Intra-articular injection of PRP can significantly reduce the subchondral BME and the level of biomarkers in synovial fluid of the symptomatic knee osteoarthritis.

PRP from Personal Blood Relieves Joint Fluid-Inducing Synovial Injury through NF-κB Pathway and Mitochondrial Apoptosis in Human Synovial Fibroblast Cells.

Background: Platelet-rich plasma (PRP) therapy is a new kind of biological therapy to retune the plasma concentrator into the patient's body for the treatment of osteoarthritis diseases. The present research aimsed to confirm the treatment effects of PRP against osteoarthritis injury and elucidate its potential mechanism via constructing a kind of cellular injury model of human synovial fibroblast cells (HSF cells) induced by synovial fluid from osteoarthritis patients. Materials and Methods: HSF cells wereas firstly treated with the different doses of synovial fluid from osteoarthritis patients, and evaluated for the cellular injury via cell morphology and MTT assay. And then, the protective effect of PRP against cellular injury were examined by cell morphology and MTT assay. Following, flow cytometry and western blot assay were employed to evaluate the effect of PRP on mitochondrial apoptosis. Finally, the effect of PRP on NF-κB pathway-associated inflammation wasere examined by Elisa ELISA assay and western blot. Results: The dilution ratio 1 : 5 of synovial fluid displayed an excellent injury effect against HSF cells and selected as the model condition. The data from cellular image and MTT assay showed that PRP with the doses 1 : 5 and 1 : 10 could alleviate the cellular mounts decrease in the damaged HSF cells. Flow cytometry, western blot, and Elisa ELISA assay displayed that PRP could relieve the cellular mitochondrial apoptosis and NF-κB pathway-associated inflammation in the damaged HSF cells. Conclusion: PRP might relieve HSF cells injury induced by synovial fluid from osteoarthritis patients through alleviating the mitochondrial apoptosis and NF-κB pathway-associated inflammation.

The metabolic transition during disease following infection of Arabidopsis thaliana by Pseudomonas syringae pv. tomato.

The outcome of bacterial infection in plants is determined by the ability of the pathogen to successfully occupy the apoplastic space and deliver a constellation of effectors that collectively suppress basal and effector-triggered immune responses. In this study, we examined the metabolic changes associated with establishment of disease using analytical techniques that interrogated a range of chemistries. We demonstrated clear differences in the metabolome of Arabidopsis thaliana leaves infected with virulent Pseudomonas syringae within 8 h of infection. In addition to confirmation of changes in phenolic and indolic compounds, we identified rapid alterations in the abundance of amino acids and other nitrogenous compounds, specific classes of glucosinolates, disaccharides, and molecules that influence the prevalence of reactive oxygen species. Our data illustrate that, superimposed on defence suppression, pathogens reconfigure host metabolism to provide the sustenance required to support exponentially growing populations of apoplastically localized bacteria. We performed a detailed baseline study reporting the metabolic dynamics associated with bacterial infection. Moreover, we have integrated these data with the results of transcriptome profiling to distinguish metabolomic pathways that are transcriptionally activated from those that are post-transcriptionally regulated.

Enhancement of plant metabolite fingerprinting by machine learning.

Metabolite fingerprinting of Arabidopsis (Arabidopsis thaliana) mutants with known or predicted metabolic lesions was performed by (1)H-nuclear magnetic resonance, Fourier transform infrared, and flow injection electrospray-mass spectrometry. Fingerprinting enabled processing of five times more plants than conventional chromatographic profiling and was competitive for discriminating mutants, other than those affected in only low-abundance metabolites. Despite their rapidity and complexity, fingerprints yielded metabolomic insights (e.g. that effects of single lesions were usually not confined to individual pathways). Among fingerprint techniques, (1)H-nuclear magnetic resonance discriminated the most mutant phenotypes from the wild type and Fourier transform infrared discriminated the fewest. To maximize information from fingerprints, data analysis was crucial. One-third of distinctive phenotypes might have been overlooked had data models been confined to principal component analysis score plots. Among several methods tested, machine learning (ML) algorithms, namely support vector machine or random forest (RF) classifiers, were unsurpassed for phenotype discrimination. Support vector machines were often the best performing classifiers, but RFs yielded some particularly informative measures. First, RFs estimated margins between mutant phenotypes, whose relations could then be visualized by Sammon mapping or hierarchical clustering. Second, RFs provided importance scores for the features within fingerprints that discriminated mutants. These scores correlated with analysis of variance F values (as did Kruskal-Wallis tests, true- and false-positive measures, mutual information, and the Relief feature selection algorithm). ML classifiers, as models trained on one data set to predict another, were ideal for focused metabolomic queries, such as the distinctiveness and consistency of mutant phenotypes. Accessible software for use of ML in plant physiology is highlighted.

Microbiomes attached to fresh perennial ryegrass are temporally resilient and adapt to changing ecological niches.

BACKGROUND: Gut microbiomes, such as the rumen, greatly influence host nutrition due to their feed energy-harvesting capacity. We investigated temporal ecological interactions facilitating energy harvesting at the fresh perennial ryegrass (PRG)-biofilm interface in the rumen using an in sacco approach and prokaryotic metatranscriptomic profiling. RESULTS: Network analysis identified two distinct sub-microbiomes primarily representing primary (≤ 4 h) and secondary (≥ 4 h) colonisation phases and the most transcriptionally active bacterial families (i.e Fibrobacteriaceae, Selemondaceae and Methanobacteriaceae) did not interact with either sub-microbiome, indicating non-cooperative behaviour. Conversely, Prevotellaceae had most transcriptional activity within the primary sub-microbiome (focussed on protein metabolism) and Lachnospiraceae within the secondary sub-microbiome (focussed on carbohydrate degradation). Putative keystone taxa, with low transcriptional activity, were identified within both sub-microbiomes, highlighting the important synergistic role of minor bacterial families; however, we hypothesise that they may be 'cheating' in order to capitalise on the energy-harvesting capacity of other microbes. In terms of chemical cues underlying transition from primary to secondary colonisation phases, we suggest that AI-2-based quorum sensing plays a role, based on LuxS gene expression data, coupled with changes in PRG chemistry. CONCLUSIONS: In summary, we show that fresh PRG-attached prokaryotes are resilient and adapt quickly to changing niches. This study provides the first major insight into the complex temporal ecological interactions occurring at the plant-biofilm interface within the rumen. The study also provides valuable insights into potential plant breeding strategies for development of the utopian plant, allowing optimal sustainable production of ruminants. Video Abstract.

Metabolite signal identification in accurate mass metabolomics data with MZedDB, an interactive m/z annotation tool utilising predicted ionisation behaviour 'rules'.

BACKGROUND: Metabolomics experiments using Mass Spectrometry (MS) technology measure the mass to charge ratio (m/z) and intensity of ionised molecules in crude extracts of complex biological samples to generate high dimensional metabolite 'fingerprint' or metabolite 'profile' data. High resolution MS instruments perform routinely with a mass accuracy of < 5 ppm (parts per million) thus providing potentially a direct method for signal putative annotation using databases containing metabolite mass information. Most database interfaces support only simple queries with the default assumption that molecules either gain or lose a single proton when ionised. In reality the annotation process is confounded by the fact that many ionisation products will be not only molecular isotopes but also salt/solvent adducts and neutral loss fragments of original metabolites. This report describes an annotation strategy that will allow searching based on all potential ionisation products predicted to form during electrospray ionisation (ESI). RESULTS: Metabolite 'structures' harvested from publicly accessible databases were converted into a common format to generate a comprehensive archive in MZedDB. 'Rules' were derived from chemical information that allowed MZedDB to generate a list of adducts and neutral loss fragments putatively able to form for each structure and calculate, on the fly, the exact molecular weight of every potential ionisation product to provide targets for annotation searches based on accurate mass. We demonstrate that data matrices representing populations of ionisation products generated from different biological matrices contain a large proportion (sometimes > 50%) of molecular isotopes, salt adducts and neutral loss fragments. Correlation analysis of ESI-MS data features confirmed the predicted relationships of m/z signals. An integrated isotope enumerator in MZedDB allowed verification of exact isotopic pattern distributions to corroborate experimental data. CONCLUSION: We conclude that although ultra-high accurate mass instruments provide major insight into the chemical diversity of biological extracts, the facile annotation of a large proportion of signals is not possible by simple, automated query of current databases using computed molecular formulae. Parameterising MZedDB to take into account predicted ionisation behaviour and the biological source of any sample improves greatly both the frequency and accuracy of potential annotation 'hits' in ESI-MS data.

A Systems Biology Approach Reveals Differences in the Dynamics of Colonization and Degradation of Grass vs. Hay by Rumen Microbes with Minor Effects of Vitamin E Supplementation.

Increasing the efficiency of utilization of fresh and preserved forage is a key target for ruminant science. Vitamin E is often used as additive to improve product quality but its impact of the rumen function is unknown. This study investigated the successional microbial colonization of ryegrass (GRA) vs. ryegrass hay (HAY) in presence of zero or 50 IU/d supplementary vitamin E, using a rumen simulation technique. A holistic approach was used to link the dynamics of feed degradation with the structure of the liquid-associated (LAB) and solid-associated bacteria (SAB). Results showed that forage colonization by SAB was a tri-phasic process highly affected by the forage conservation method: Early colonization (0-2 h after feeding) by rumen microbes was 2× faster for GRA than HAY diets and dominated by Lactobacillus and Prevotella which promoted increased levels of lactate (+56%) and ammonia (+18%). HAY diets had lower DM degradation (-72%) during this interval being Streptococcus particularly abundant. During secondary colonization (4-8 h) the SAB community increased in size and decreased in diversity as the secondary colonizers took over (Pseudobutyrivibrio) promoting the biggest differences in the metabolomics profile between diets. Secondary colonization was 3× slower for HAY vs. GRA diets, but this delay was compensated by a greater bacterial diversity (+197 OTUs) and network complexity resulting in similar feed degradations. Tertiary colonization (>8 h) consisted of a slowdown in the colonization process and simplification of the bacterial network. This slowdown was less evident for HAY diets which had higher levels of tertiary colonizers (Butyrivibrio and Ruminococcus) and may explain the higher DM degradation (+52%) during this interval. The LAB community was particularly active during the early fermentation of GRA and during the late fermentation for HAY diets indicating that the availability of nutrients in the liquid phase reflects the dynamics of feed degradation. Vitamin E supplementation had minor effects but promoted a simplification of the LAB community and a slight acceleration in the SAB colonization sequence which could explain the higher DM degradation during the secondary colonization. Our findings suggest that when possible, grass should be fed instead of hay, in order to accelerate feed utilization by rumen microbes.

Temporal dynamics of the metabolically active rumen bacteria colonizing fresh perennial ryegrass.

This study investigated successional colonization of fresh perennial ryegrass (PRG) by the rumen microbiota over time. Fresh PRG was incubated in sacco in the rumens of three Holstein × Friesian cows over a period of 8 h, with samples recovered at various times. The diversity of attached bacteria was assessed using 454 pyrosequencing of 16S rRNA (cDNA). Results showed that plant epiphytic communities either decreased to low relative abundances or disappeared following rumen incubation, and that temporal colonization of the PRG by the rumen bacteria was biphasic with primary (1 and 2 h) and secondary (4-8 h) events evident with the transition period being with 2-4 h. A decrease in sequence reads pertaining to Succinivibrio spp. and increases in Pseudobutyrivibrio, Roseburia and Ruminococcus spp. (the latter all order Clostridiales) were evident during secondary colonization. Irrespective of temporal changes, the continually high abundances of Butyrivibrio, Fibrobacter, Olsenella and Prevotella suggest that they play a major role in the degradation of the plant. It is clear that a temporal understanding of the functional roles of these microbiota within the rumen is now required to unravel the role of these bacteria in the ruminal degradation of fresh PRG.

Metagenomic Sequencing of the Chronic Obstructive Pulmonary Disease Upper Bronchial Tract Microbiome Reveals Functional Changes Associated with Disease Severity.

Chronic Obstructive Pulmonary Disease (COPD) is a major source of mortality and morbidity worldwide. The microbiome associated with this disease may be an important component of the disease, though studies to date have been based on sequencing of the 16S rRNA gene, and have revealed unequivocal results. Here, we employed metagenomic sequencing of the upper bronchial tract (UBT) microbiome to allow for greater elucidation of its taxonomic composition, and revealing functional changes associated with the disease. The bacterial metagenomes within sputum samples from eight COPD patients and ten 'healthy' smokers (Controls) were sequenced, and suggested significant changes in the abundance of bacterial species, particularly within the Streptococcus genus. The functional capacity of the COPD UBT microbiome indicated an increased capacity for bacterial growth, which could be an important feature in bacterial-associated acute exacerbations. Regression analyses correlated COPD severity (FEV1% of predicted) with differences in the abundance of Streptococcus pneumoniae and functional classifications related to a reduced capacity for bacterial sialic acid metabolism. This study suggests that the COPD UBT microbiome could be used in patient risk stratification and in identifying novel monitoring and treatment methods, but study of a longitudinal cohort will be required to unequivocally relate these features of the microbiome with COPD severity.

Elevation of brain glucose and polyol-pathway intermediates with accompanying brain-copper deficiency in patients with Alzheimer's disease: metabolic basis for dementia.

Impairment of brain-glucose uptake and brain-copper regulation occurs in Alzheimer's disease (AD). Here we sought to further elucidate the processes that cause neurodegeneration in AD by measuring levels of metabolites and metals in brain regions that undergo different degrees of damage. We employed mass spectrometry (MS) to measure metabolites and metals in seven post-mortem brain regions of nine AD patients and nine controls, and plasma-glucose and plasma-copper levels in an ante-mortem case-control study. Glucose, sorbitol and fructose were markedly elevated in all AD brain regions, whereas copper was correspondingly deficient throughout (all P < 0.0001). In the ante-mortem case-control study, by contrast, plasma-glucose and plasma-copper levels did not differ between patients and controls. There were pervasive defects in regulation of glucose and copper in AD brain but no evidence for corresponding systemic abnormalities in plasma. Elevation of brain glucose and deficient brain copper potentially contribute to the pathogenesis of neurodegeneration in AD.

Molecular phenotyping of a UK population: defining the human serum metabolome.

Phenotyping of 1,200 'healthy' adults from the UK has been performed through the investigation of diverse classes of hydrophilic and lipophilic metabolites present in serum by applying a series of chromatography-mass spectrometry platforms. These data were made robust to instrumental drift by numerical correction; this was prerequisite to allow detection of subtle metabolic differences. The variation in observed metabolite relative concentrations between the 1,200 subjects ranged from less than 5 % to more than 200 %. Variations in metabolites could be related to differences in gender, age, BMI, blood pressure, and smoking. Investigations suggest that a sample size of 600 subjects is both necessary and sufficient for robust analysis of these data. Overall, this is a large scale and non-targeted chromatographic MS-based metabolomics study, using samples from over 1,000 individuals, to provide a comprehensive measurement of their serum metabolomes. This work provides an important baseline or reference dataset for understanding the 'normal' relative concentrations and variation in the human serum metabolome. These may be related to our increasing knowledge of the human metabolic network map. Information on the Husermet study is available at http://www.husermet.org/. Importantly, all of the data are made freely available at MetaboLights (http://www.ebi.ac.uk/metabolights/).

Age-dependent changes in heparan sulfate in human Bruch's membrane: implications for age-related macular degeneration.

PURPOSE: Heparan sulfate (HS) has been implicated in age-related macular degeneration (AMD), since it is the major binding partner for complement factor H (CFH) in human Bruch's membrane (BrM), and CFH has a central role in inhibiting complement activation on extracellular matrices. The aim was to investigate potential aging changes in HS quantity and composition in human BrM. METHODS: Postmortem human ocular tissue was obtained from donors without known retinal disease. The HS was purified from BrM and neurosensory retina, and after digestion to disaccharides, fluorescently labeled and analyzed by reverse-phase HPLC. The HS and heparanase-1 were detected by immunohistochemistry in macular tissue sections from young and old donors, and binding of exogenously applied recombinant CCP6-8 region of CFH (402Y and 402H variants) was compared. RESULTS: Disaccharide analysis demonstrated that the mean quantity of HS in BrM was 50% lower (P = 0.006) in old versus young donors (average 82 vs. 32 years). In addition, there was a small, but significant decrease in HS sulfation in old BrM. Immunohistochemistry revealed approximately 50% (P = 0.02) less HS in macular BrM in old versus young donors, whereas heparanase-1 increased by 24% in old macular BrM (P = 0.56). In young donor tissue the AMD-associated 402H CCP6-8 bound relatively poorly to BrM, compared to the 402Y form. In BrM from old donors, this difference was significantly greater (P = 0.019). CONCLUSIONS: The quantity of HS decreases substantially with age in human BrM, resulting in fewer binding sites for CFH and especially affecting the ability of the 402H variant of CFH to bind BrM.

Use of mass spectrometry fingerprinting to identify urinary metabolites after consumption of specific foods.

BACKGROUND: The lack of robust biological markers of dietary exposure hinders the quantitative understanding of causal relations between diet and health. OBJECTIVE: We aimed to develop an efficient procedure to discover metabolites in urine that may have future potential as biomarkers of acute exposure to foods of high public health importance. DESIGN: Twenty-four participants were provided with a test breakfast in which the cereal component of a standardized breakfast was replaced by 1 of 4 foods of high public health importance; 1.5-, 3-, and 4.5-h postprandial urine samples were collected. Flow infusion electrospray-ionization mass spectrometry followed by supervised multivariate data analysis was used to discover signals resulting from consumption of each test food. RESULTS: Fasted-state urine samples provided a universal comparator for food biomarker lead discovery in postprandial urine. The filtering of data features associated with consumption of the common components of the standardized breakfast improved discrimination models and readily identified metabolites that showed consumption of specific test foods. A combination of trimethylamine-N-oxide and 1-methylhistidine was associated with salmon consumption. Novel ascorbate derivatives were discovered in urine after consumption of either broccoli or raspberries. Sulphonated caffeic acid and sulphonated methyl-epicatechin concentrations increased dramatically after consumption of raspberries. CONCLUSIONS: This biomarker lead discovery strategy can identify urinary metabolites associated with acute exposure to individual foods. Future studies are required to validate the specificity and utility of potential biomarkers in an epidemiologic context.

Development and validation of a standardized protocol to monitor human dietary exposure by metabolite fingerprinting of urine samples.

Conventional tools for measuring dietary exposure have well recognized limitations. Measurement of food-derived metabolites in biofluids provides an alternative approach and our aim was to develop an experimental protocol which ensures that extraneous variability does not obscure metabolic signals from ingested foods. Healthy adults consumed a standardized meal in the evening before each test day and collected pooled overnight urine. On each test day of three different studies, urine was collected in the fasted state and at different time points after consumption of a standardized breakfast. Metabolite fingerprinting of samples using Flow Infusion Electrospray-Ionization Mass Spectrometry followed by multivariate data analysis showed strong discrimination between overnight, fasting and postprandial samples, in each study separately and when data from the three studies were pooled. Such differences were robust and highly reproducible within individuals on separate occasions. Urine volume was an efficient data normalization factor for metabolite fingerprinting data. Postprandial urines had a stable chemical composition over a period of 2-4 h after eating a standardized breakfast, suggesting that there is a flexible time window for urine collection. Fasting urine samples provided a stable baseline for universal comparisons with postprandial samples. A dietary exposure biomarker discovery protocol was validated by demonstrating that top-ranked signals discriminating between fasting and 2-4 h postprandial urine samples could be linked to metabolites abundant in some components of the standardized breakfast. We conclude that the protocol developed will have value in the search for biomarker leads of dietary exposure. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s11306-011-0289-0) contains supplementary material, which is available to authorized users.

Preprocessing, classification modeling and feature selection using flow injection electrospray mass spectrometry metabolite fingerprint data.

Metabolome analysis by flow injection electrospray mass spectrometry (FIE-MS) fingerprinting generates measurements relating to large numbers of m/z signals. Such data sets often exhibit high variance with a paucity of replicates, thus providing a challenge for data mining. We describe data preprocessing and modeling methods that have proved reliable in projects involving samples from a range of organisms. The protocols interact with software resources specifically for metabolomics provided in a Web-accessible data analysis package FIEmspro (http://users.aber.ac.uk/jhd) written in the R environment and requiring a moderate knowledge of R command-line usage. Specific emphasis is placed on describing the outcome of modeling experiments using FIE-MS data that require further preprocessing to improve quality. The salient features of both poor and robust (i.e., highly generalizable) multivariate models are outlined together with advice on validating classifiers and avoiding false discovery when seeking explanatory variables.