Background Quality assurance (QA) and quality control (QC) are two quality management processes that are integral to the success of metabolomics including their application for the acquisition of high quality data in any high-throughput analytical chemistry laboratory. QA defines all the planned and systematic activities implemented before samples are collected, to provide confidence that a subsequent analytical process will fulfil predetermined requirements for quality. QC can be defined as the operational techniques and activities used to measure and report these quality requirements after data acquisition. Aim of review This tutorial review will guide the reader through the use of system suitability and QC samples, why these samples should be applied and how the quality of data can be reported. Key scientific concepts of review System suitability samples are applied to assess the operation and lack of contamination of the analytical platform prior to sample analysis. Isotopically-labelled internal standards are applied to assess system stability for each sample analysed. Pooled QC samples are applied to condition the analytical platform, perform intra-study reproducibility measurements (QC) and to correct mathematically for systematic errors. Standard reference materials and long-term reference QC samples are applied for inter-study and inter-laboratory assessment of data.

Asthma is a heterogeneous disease in which there is a differential response to asthma treatments. This heterogeneity needs to be evaluated so that a personalized management approach can be provided. We stratified patients with moderate-to-severe asthma based on clinicophysiologic parameters and performed an omics analysis of sputum. Partition-around-medoids clustering was applied to a training set of 266 asthmatic participants from the European Unbiased Biomarkers for the Prediction of Respiratory Diseases Outcomes (U-BIOPRED) adult cohort using 8 prespecified clinic-physiologic variables. This was repeated in a separate validation set of 152 asthmatic patients. The clusters were compared based on sputum proteomics and transcriptomics data. Four reproducible and stable clusters of asthmatic patients were identified. The training set cluster T1 consists of patients with well-controlled moderate-to-severe asthma, whereas cluster T2 is a group of patients with late-onset severe asthma with a history of smoking and chronic airflow obstruction. Cluster T3 is similar to cluster T2 in terms of chronic airflow obstruction but is composed of nonsmokers. Cluster T4 is predominantly composed of obese female patients with uncontrolled severe asthma with increased exacerbations but with normal lung function. The validation set exhibited similar clusters, demonstrating reproducibility of the classification. There were significant differences in sputum proteomics and transcriptomics between the clusters. The severe asthma clusters (T2, T3, and T4) had higher sputum eosinophilia than cluster T1, with no differences in sputum neutrophil counts and exhaled nitric oxide and serum IgE levels. Clustering based on clinicophysiologic parameters yielded 4 stable and reproducible clusters that associate with different pathobiological pathways.

Human immunodeficiency virus type 1(HIV-1) infects and activates innate immune cells in the brain resulting in inflammation and neuronal death with accompanying neurological deficits. Induction of inflammasomes causes cleavage and release of IL-1β and IL-18, representing pathogenic processes that underlie inflammatory diseases although their contribution HIV-associated brain disease is unknown.

Introduction: Metabolomics is increasingly being used in the clinical setting for disease diagnosis, prognosis and risk prediction. Machine learning algorithms are particularly important in the construction of multivariate metabolite prediction. Historically, partial least squares (PLS) regression has been the gold standard for binary classification. Nonlinear machine learning methods such as random forests (RF), kernel support vector machines (SVM) and artificial neural networks (ANN) may be more suited to modelling possible nonlinear metabolite covariance, and thus provide better predictive models.Objectives: We hypothesise that for binary classification using metabolomics data, non-linear machine learning methods will provide superior generalised predictive ability when compared to linear alternatives, in particular when compared with the current gold standard PLS discriminant analysis.Methods: We compared the general predictive performance of eight archetypal machine learning algorithms across ten publicly available clinical metabolomics data sets. The algorithms were implemented in the Python programming language. All code and results have been made publicly available as Jupyter notebooks.Results: There was only marginal improvement in predictive ability for SVM and ANN over PLS across all data sets. RF performance was comparatively poor. The use of out-of-bag bootstrap confidence intervals provided a measure of uncertainty of model prediction such that the quality of metabolomics data was observed to be a bigger influence on generalised performance than model choice.Conclusion: The size of the data set, and choice of performance metric, had a greater influence on generalised predictive performance than the choice of machine learning algorithm.

In this study, we sought to determine whether asthma has a metabolic profile and whether this profile is related to disease severity.We characterised the serum from 22 healthy individuals and 54 asthmatics (12 mild, 20 moderate, 22 severe) using liquid chromatography–high-resolution mass spectrometry-based metabolomics. Selected metabolites were confirmed by targeted mass spectrometry assays of eicosanoids, sphingolipids and free fatty acids.We conclusively identified 66 metabolites; 15 were significantly altered with asthma (p≤0.05). Levels of dehydroepiandrosterone sulfate, cortisone, cortisol, prolylhydroxyproline, pipecolate and N-palmitoyltaurine correlated significantly (p<0.05) with inhaled corticosteroid dose, and were further shifted in individuals treated with oral corticosteroids. Oleoylethanolamide increased with asthma severity independently of steroid treatment (p<0.001). Multivariate analysis revealed two patterns: 1) a mean difference between controls and patients with mild asthma (p=0.025), and 2) a mean difference between patients with severe asthma and all other groups (p=1.7×10⁻⁴). Metabolic shifts in mild asthma, relative to controls, were associated with exogenous metabolites (e.g.dietary lipids), while those in moderate and severe asthma (e.g. oleoylethanolamide, sphingosine-1-phosphate, N-palmitoyltaurine) were postulated to be involved in activating the transient receptor potential vanilloid type 1 (TRPV1) receptor, driving TRPV1-dependent pathogenesis in asthma.Our findings suggest that asthma is characterised by a modest systemic metabolic shift in a disease severity-dependent manner, and that steroid treatment significantly affects metabolism.

High-resolution mass spectrometry (HRMS)-based metabolomics approaches have made significant advances. However, metabolite identification is still a major challenge with significant bottleneck in translating metabolomics data into biological context. In the current study, a liquid chromatography (LC)–HRMS metabolomics method was developed using an all ion fragmentation (AIF) acquisition approach. To increase the specificity in metabolite annotation, four criteria were considered: (i) accurate mass (AM), (ii) retention time (RT), (iii) MS/MS spectrum, and (iv) product/precursor ion intensity ratios. We constructed an in-house mass spectral library of 408 metabolites containing AMRT and MS/MS spectra information at four collision energies. The percent relative standard deviations between ion ratios of a metabolite in an analytical standard vs sample matrix were used as an additional metric for establishing metabolite identity. A data processing method for targeted metabolite screening was then created, merging m/z, RT, MS/MS, and ion ratio information for each of the 413 metabolites. In the data processing method, the precursor ion and product ion were considered as the quantifier and qualifier ion, respectively. We also included a scheme to distinguish coeluting isobaric compounds by selecting a specific product ion as the quantifier ion instead of the precursor ion. An advantage of the current AIF approach is the concurrent collection of full scan data, enabling identification of metabolites not included in the database. Our data acquisition strategy enables a simultaneous mixture of database-dependent targeted and nontargeted metabolomics in combination with improved accuracy in metabolite identification, increasing the quality of the biological information acquired in a metabolomics experiment.