Statistical analysis in metabolic phenotyping.

Metabolic phenotyping is an important tool in translational biomedical research. The advanced analytical technologies commonly used for phenotyping, including mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy, generate complex data requiring tailored statistical analysis methods. Detailed protocols have been published for data acquisition by liquid NMR, solid-state NMR, ultra-performance liquid chromatography (LC-)MS and gas chromatography (GC-)MS on biofluids or tissues and their preprocessing. Here we propose an efficient protocol (guidelines and software) for statistical analysis of metabolic data generated by these methods. Code for all steps is provided, and no prior coding skill is necessary. We offer efficient solutions for the different steps required within the complete phenotyping data analytics workflow: scaling, normalization, outlier detection, multivariate analysis to explore and model study-related effects, selection of candidate biomarkers, validation, multiple testing correction and performance evaluation of statistical models. We also provide a statistical power calculation algorithm and safeguards to ensure robust and meaningful experimental designs that deliver reliable results. We exemplify the protocol with a two-group classification study and data from an epidemiological cohort; however, the protocol can be easily modified to cover a wider range of experimental designs or incorporate different modeling approaches. This protocol describes a minimal set of analyses needed to rigorously investigate typical datasets encountered in metabolic phenotyping.

peakPantheR, an R package for large-scale targeted extraction and integration of annotated metabolic features in LC-MS profiling datasets.

SUMMARY: Untargeted liquid chromatography-mass spectrometry (LC-MS) profiling assays are capable of measuring thousands of chemical compounds in a single sample, but unreliable feature extraction and metabolite identification remain considerable barriers to their interpretation and usefulness. peakPantheR (Peak Picking and ANnoTation of High-resolution Experiments in R) is an R package for the targeted extraction and integration of annotated features from LC-MS profiling experiments. It takes advantage of chromatographic and spectral databases and prior information of sample matrix composition to generate annotated and interpretable metabolic phenotypic datasets and power workflows for real-time data quality assessment. AVAILABILITY AND IMPLEMENTATION: peakPantheR is available via Bioconductor (https://bioconductor.org/packages/peakPantheR/). Documentation and worked examples are available at https://phenomecentre.github.io/peakPantheR.github.io/ and https://github.com/phenomecentre/metabotyping-dementia-urine. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The nPYc-Toolbox, a Python module for the pre-processing, quality-control and analysis of metabolic profiling datasets.

SUMMARY: As large-scale metabolic phenotyping studies become increasingly common, the need for systemic methods for pre-processing and quality control (QC) of analytical data prior to statistical analysis has become increasingly important, both within a study, and to allow meaningful inter-study comparisons. The nPYc-Toolbox provides software for the import, pre-processing, QC and visualization of metabolic phenotyping datasets, either interactively, or in automated pipelines. AVAILABILITY AND IMPLEMENTATION: The nPYc-Toolbox is implemented in Python, and is freely available from the Python package index https://pypi.org/project/nPYc/, source is available at https://github.com/phenomecentre/nPYc-Toolbox. Full documentation can be found at http://npyc-toolbox.readthedocs.io/ and exemplar datasets and tutorials at https://github.com/phenomecentre/nPYc-toolbox-tutorials.

mzTab-M: A Data Standard for Sharing Quantitative Results in Mass Spectrometry Metabolomics.

Mass spectrometry (MS) is one of the primary techniques used for large-scale analysis of small molecules in metabolomics studies. To date, there has been little data format standardization in this field, as different software packages export results in different formats represented in XML or plain text, making data sharing, database deposition, and reanalysis highly challenging. Working within the consortia of the Metabolomics Standards Initiative, Proteomics Standards Initiative, and the Metabolomics Society, we have created mzTab-M to act as a common output format from analytical approaches using MS on small molecules. The format has been developed over several years, with input from a wide range of stakeholders. mzTab-M is a simple tab-separated text format, but importantly, the structure is highly standardized through the design of a detailed specification document, tightly coupled to validation software, and a mandatory controlled vocabulary of terms to populate it. The format is able to represent final quantification values from analyses, as well as the evidence trail in terms of features measured directly from MS (e.g., LC-MS, GC-MS, DIMS, etc.) and different types of approaches used to identify molecules. mzTab-M allows for ambiguity in the identification of molecules to be communicated clearly to readers of the files (both people and software). There are several implementations of the format available, and we anticipate widespread adoption in the field.

PhenoMeNal: processing and analysis of metabolomics data in the cloud.

BACKGROUND: Metabolomics is the comprehensive study of a multitude of small molecules to gain insight into an organism's metabolism. The research field is dynamic and expanding with applications across biomedical, biotechnological, and many other applied biological domains. Its computationally intensive nature has driven requirements for open data formats, data repositories, and data analysis tools. However, the rapid progress has resulted in a mosaic of independent, and sometimes incompatible, analysis methods that are difficult to connect into a useful and complete data analysis solution. FINDINGS: PhenoMeNal (Phenome and Metabolome aNalysis) is an advanced and complete solution to set up Infrastructure-as-a-Service (IaaS) that brings workflow-oriented, interoperable metabolomics data analysis platforms into the cloud. PhenoMeNal seamlessly integrates a wide array of existing open-source tools that are tested and packaged as Docker containers through the project's continuous integration process and deployed based on a kubernetes orchestration framework. It also provides a number of standardized, automated, and published analysis workflows in the user interfaces Galaxy, Jupyter, Luigi, and Pachyderm. CONCLUSIONS: PhenoMeNal constitutes a keystone solution in cloud e-infrastructures available for metabolomics. PhenoMeNal is a unique and complete solution for setting up cloud e-infrastructures through easy-to-use web interfaces that can be scaled to any custom public and private cloud environment. By harmonizing and automating software installation and configuration and through ready-to-use scientific workflow user interfaces, PhenoMeNal has succeeded in providing scientists with workflow-driven, reproducible, and shareable metabolomics data analysis platforms that are interfaced through standard data formats, representative datasets, versioned, and have been tested for reproducibility and interoperability. The elastic implementation of PhenoMeNal further allows easy adaptation of the infrastructure to other application areas and 'omics research domains.

Quantitative Lipoprotein Subclass and Low Molecular Weight Metabolite Analysis in Human Serum and Plasma by 1H NMR Spectroscopy in a Multilaboratory Trial.

We report an extensive 600 MHz NMR trial of quantitative lipoprotein and small-molecule measurements in human blood serum and plasma. Five centers with eleven 600 MHz NMR spectrometers were used to analyze 98 samples including 20 quality controls (QCs), 37 commercially sourced, paired serum and plasma samples, and two National Institute of Science and Technology (NIST) reference material 1951c replicates. Samples were analyzed using rigorous protocols for sample preparation and experimental acquisition. A commercial lipoprotein subclass analysis was used to quantify 105 lipoprotein subclasses and 24 low molecular weight metabolites from the NMR spectra. For all spectrometers, the instrument specific variance in measuring internal QCs was lower than the percentage described by the National Cholesterol Education Program (NCEP) criteria for lipid testing [triglycerides <2.7%; cholesterol <2.8%; low-density lipoprotein (LDL) cholesterol <2.8%; high-density lipoprotein (HDL) cholesterol <2.3%], showing exceptional reproducibility for direct quantitation of lipoproteins in both matrixes. The average relative standard deviations (RSDs) for the 105 lipoprotein parameters in the 11 instruments were 4.6% and 3.9% for the two NIST samples, whereas they were 38% and 40% for the 37 commercially sourced plasmas and sera, respectively, showing negligible analytical compared to biological variation. The coefficient of variance (CV) obtained for the quantification of the small molecules across the 11 spectrometers was below 15% for 20 out of the 24 metabolites analyzed. This study provides further evidence of the suitability of NMR for high-throughput lipoprotein subcomponent analysis and small-molecule quantitation with the exceptional required reproducibility for clinical and other regulatory settings.

The effects of kisspeptin on ß-cell function, serum metabolites and appetite in humans.

AIMS: To investigate the effect of kisspeptin on glucose-stimulated insulin secretion and appetite in humans. MATERIALS AND METHODS: In 15 healthy men (age: 25.2 ± 1.1 years; BMI: 22.3 ± 0.5 kg m-2 ), we compared the effects of 1 nmol kg-1 h-1 kisspeptin versus vehicle administration on glucose-stimulated insulin secretion, metabolites, gut hormones, appetite and food intake. In addition, we assessed the effect of kisspeptin on glucose-stimulated insulin secretion in vitro in human pancreatic islets and a human ß-cell line (EndoC-ßH1 cells). RESULTS: Kisspeptin administration to healthy men enhanced insulin secretion following an intravenous glucose load, and modulated serum metabolites. In keeping with this, kisspeptin increased glucose-stimulated insulin secretion from human islets and a human pancreatic cell line in vitro. In addition, kisspeptin administration did not alter gut hormones, appetite or food intake in healthy men. CONCLUSIONS: Collectively, these data demonstrate for the first time a beneficial role for kisspeptin in insulin secretion in humans in vivo. This has important implications for our understanding of the links between reproduction and metabolism in humans, as well as for the ongoing translational development of kisspeptin-based therapies for reproductive and potentially metabolic conditions.

Development and Application of Ultra-Performance Liquid Chromatography-TOF MS for Precision Large Scale Urinary Metabolic Phenotyping.

To better understand the molecular mechanisms underpinning physiological variation in human populations, metabolic phenotyping approaches are increasingly being applied to studies involving hundreds and thousands of biofluid samples. Hyphenated ultra-performance liquid chromatography-mass spectrometry (UPLC-MS) has become a fundamental tool for this purpose. However, the seemingly inevitable need to analyze large studies in multiple analytical batches for UPLC-MS analysis poses a challenge to data quality which has been recognized in the field. Herein, we describe in detail a fit-for-purpose UPLC-MS platform, method set, and sample analysis workflow, capable of sustained analysis on an industrial scale and allowing batch-free operation for large studies. Using complementary reversed-phase chromatography (RPC) and hydrophilic interaction liquid chromatography (HILIC) together with high resolution orthogonal acceleration time-of-flight mass spectrometry (oaTOF-MS), exceptional measurement precision is exemplified with independent epidemiological sample sets of approximately 650 and 1000 participant samples. Evaluation of molecular reference targets in repeated injections of pooled quality control (QC) samples distributed throughout each experiment demonstrates a mean retention time relative standard deviation (RSD) of <0.3% across all assays in both studies and a mean peak area RSD of <15% in the raw data. To more globally assess the quality of the profiling data, untargeted feature extraction was performed followed by data filtration according to feature intensity response to QC sample dilution. Analysis of the remaining features within the repeated QC sample measurements demonstrated median peak area RSD values of <20% for the RPC assays and <25% for the HILIC assays. These values represent the quality of the raw data, as no normalization or feature-specific intensity correction was applied. While the data in each experiment was acquired in a single continuous batch, instances of minor time-dependent intensity drift were observed, highlighting the utility of data correction techniques despite reducing the dependency on them for generating high quality data. These results demonstrate that the platform and methodology presented herein is fit-for-use in large scale metabolic phenotyping studies, challenging the assertion that such screening is inherently limited by batch effects. Details of the pipeline used to generate high quality raw data and mitigate the need for batch correction are provided.

Power Analysis and Sample Size Determination in Metabolic Phenotyping.

Estimation of statistical power and sample size is a key aspect of experimental design. However, in metabolic phenotyping, there is currently no accepted approach for these tasks, in large part due to the unknown nature of the expected effect. In such hypothesis free science, neither the number or class of important analytes nor the effect size are known a priori. We introduce a new approach, based on multivariate simulation, which deals effectively with the highly correlated structure and high-dimensionality of metabolic phenotyping data. First, a large data set is simulated based on the characteristics of a pilot study investigating a given biomedical issue. An effect of a given size, corresponding either to a discrete (classification) or continuous (regression) outcome is then added. Different sample sizes are modeled by randomly selecting data sets of various sizes from the simulated data. We investigate different methods for effect detection, including univariate and multivariate techniques. Our framework allows us to investigate the complex relationship between sample size, power, and effect size for real multivariate data sets. For instance, we demonstrate for an example pilot data set that certain features achieve a power of 0.8 for a sample size of 20 samples or that a cross-validated predictivity QY(2) of 0.8 is reached with an effect size of 0.2 and 200 samples. We exemplify the approach for both nuclear magnetic resonance and liquid chromatography-mass spectrometry data from humans and the model organism C. elegans.

Precision high-throughput proton NMR spectroscopy of human urine, serum, and plasma for large-scale metabolic phenotyping.

Proton nuclear magnetic resonance (NMR)-based metabolic phenotyping of urine and blood plasma/serum samples provides important prognostic and diagnostic information and permits monitoring of disease progression in an objective manner. Much effort has been made in recent years to develop NMR instrumentation and technology to allow the acquisition of data in an effective, reproducible, and high-throughput approach that allows the study of general population samples from epidemiological collections for biomarkers of disease risk. The challenge remains to develop highly reproducible methods and standardized protocols that minimize technical or experimental bias, allowing realistic interlaboratory comparisons of subtle biomarker information. Here we present a detailed set of updated protocols that carefully consider major experimental conditions, including sample preparation, spectrometer parameters, NMR pulse sequences, throughput, reproducibility, quality control, and resolution. These results provide an experimental platform that facilitates NMR spectroscopy usage across different large cohorts of biofluid samples, enabling integration of global metabolic profiling that is a prerequisite for personalized healthcare.

Ultra performance liquid chromatography-mass spectrometry profiling of bile acid metabolites in biofluids: application to experimental toxicology studies.

We have developed an ultra performance liquid chromatography-mass spectrometry (UPLC-MS(E)) method to measure bile acids (BAs) reproducibly and reliably in biological fluids and have applied this approach for indications of hepatic damage in experimental toxicity studies. BAs were extracted from serum using methanol, and an Acquity HSS column coupled to a Q-ToF mass spectrometer was used to separate and identify 25 individual BAs within 5 min. Employing a gradient elution of water and acetonitrile over 21 min enabled the detection of a wide range of endogenous metabolites, including the BAs. The utilization of MS(E) allowed for characteristic fragmentation information to be obtained in a single analytical run, easily distinguishing glycine and taurine BA conjugates. The proportions of these conjugates were altered markedly in an experimental toxic state induced by galactosamine exposure in rats. Principally, taurine-conjugated BAs were greatly elevated ( approximately 50-fold from control levels), and were highly correlated to liver damage severity as assessed by histopathological scoring (r = 0.83), indicating their potential as a sensitive measure of hepatic damage. The UPLC-MS approach to BA analysis offers a sensitive and reproducible tool that will be of great value in exploring both markers and mechanisms of hepatotoxicity and can readily be extended to clinical studies of liver damage.

Robust algorithms for automated chemical shift calibration of 1D 1H NMR spectra of blood serum.

In biofluid NMR spectroscopy, the frequency of each resonance is typically calibrated by addition of a reference compound such as 3-(trimethylsilyl)-propionic acid- d 4 (TSP) to the sample. However biofluids such as serum cannot be referenced to TSP, due to shifts resonance caused by binding to macromolecules in solution. In order to overcome this limitation we have developed algorithms, based on analysis of derivative spectra, to locate and calibrate (1)H NMR spectra to the alpha-glucose anomeric doublet. We successfully used these algorithms to calibrate 77 serum (1)H NMR spectra and demonstrate the greater reproducibility of the calculated chemical-shift corrections ( r = 0.97) than those generated by manual alignment ( r = 0.8-0.88). Hence we show that these algorithms provide robust and reproducible methods of calibrating (1)H NMR of serum, plasma, or any biofluid in which glucose is abundant. Precise automated calibration of complex biofluid NMR spectra is an important tool in large-scale metabonomic or metabolomic studies, where hundreds or even thousands of spectra may be analyzed in high-resolution by pattern recognition analysis.