Bashar Amer - Metabolomics Application Development, Thermo Fisher Scientific
Introduction
Metabolomics, using Liquid Chromatography-Mass Spectrometry (LC-MS), is one of the most powerful ways to deliver molecular profiles from patient samples. The clear annotation of metabolites enables scientists to accurately characterize biomarkers for disease diagnosis, prognosis and treatment, while quantification of metabolites enables validation of biomarker discoveries and provides benchmarks for normal metabolite levels (e.g., normal and abnormal sugar concentrations in the blood).
The study of metabolites usually falls under one of two main analytical chemistry strategies: targeted or untargeted analysis. Targeted analysis provides accurate quantitation and annotation of a predefined group of metabolites, while untargeted analysis gives an overview of changes in metabolite relative concentrations across known and unknown metabolites in a sample. Both approaches have their own strengths and drawbacks when using LC-MS, so researchers are often faced with a choice: a low accuracy overview of total molecular changes, or a detailed yet blinkered snapshot of a select group of metabolites.
However, a new semi-targeted workflow combining both untargeted and targeted workflows has recently emerged as a middle ground, addressing the limitations of these approaches. The primary focus of this semi-targeted approach is the confident annotation, and optionally the accurate quantitation, of a targeted set of metabolites. In conjunction, the secondary focus is to find new molecular connections in a system by performing untargeted analysis on a single injection. Here, we discuss how semi-targeted metabolomics could offer the best of both worlds, and how this technique can be implemented by laboratories to overcome barriers across multiple fields for their analytical needs.
Targeted and Untargeted Metabolomics
To make meaningful biological interpretations of metabolomic data, researchers need two main insights: specific identification and quantification of metabolite changes (targeted analysis), and an overview of general changes to the metabolome (untargeted analysis).
Targeted metabolomics is used in hypothesis-driven research to annotate and quantify a biologically relevant subset of known metabolites, defined at the start of the study before data acquisition. The workflow requires ultimate sensitivity and specificity, and typically uses triple quadrupole mass spectrometers where only the metabolites of interest are detected. For example, researchers can identify a select group of metabolites associated with type II diabetes to be analyzed in a sample from a suspected diabetes patient. These metabolites can then be chemically annotated and quantified, making the approach highly accurate at detecting these specific metabolite changes. However, targeted analysis such as this has a major drawback: a limited coverage of the metabolome. As such, researchers might miss metabolite changes outside of their pre-defined subset — information that can be of biological significance to the question at hand.
In comparison, untargeted discovery metabolomics offers a wider overview of metabolites and their relative levels. Here, researchers will acquire full scan MS1 data to capture all ions in the sample. The advantage of this is that analysts can maximize the number of metabolites detected, providing the opportunity to find unexpected changes not part of the original study goal. But, of course, the general, untargeted overview comes with several limitations. For one, the varying physiochemical properties of the metabolome require specific detection criteria for different metabolite groups. This means researchers must compromise on a combination of analytical parameters (e.g., stationary phase for separation and ionization mode for MS-based detection), which may improve the detection of some substances, but reduce the detection of others. Moreover, untargeted analysis can suffer from signal bias and mass drift introduced by complex sample matrix effects, which complicates metabolite identification and reduces overall sensitivity. Therefore, it is critical to use High-Resolution Accurate-Mass (HRAM) detection, such as that provided by Orbitrap mass analyzers — when feasible — for untargeted metabolomics studies.
What is Semi-Targeted Metabolomics?
The recently developed semi-targeted metabolomics technique is a promising alternative, offering researchers a way to strike the balance between untargeted and targeted approaches in one single experiment. Semi-targeted workflows begin in much the same way as targeted approaches: researchers annotate and quantify a pre-selected group of metabolites in a sample. However, the data can then be reanalyzed (or retro-mined) to look for global metabolic changes that were not part of the original focus. Semi-targeted analysis can, therefore, identify other biologically meaningful metabolite changes that the scientist may not have been aware of in their signaling pathway of interest.
One of the biggest strengths of semi-targeted metabolomics is the ability to perform targeted and untargeted analysis in a single sample injection. Traditionally, in metabolomics experiments a sample is injected (analyzed) twice; once for untargeted metabolomics analysis and a second time for targeted analysis. In comparison, the single injection workflow is particularly advantageous for laboratories that have limited access to samples, time and resources, and offers a powerful and efficient way to gain more knowledge from valuable biological samples.
Exemplifying the potential of semi-targeted workflows, a recent study demonstrated how the technique can benefit both hypothesis-based targeted verification and discovery untargeted data acquisition. Here, semi-targeted workflows were utilized for cancer metabolomics using Ultra-High-Performance Liquid Chromatography coupled with High-Resolution Tandem MS (UHPLC-HRMS/MS). In the workflow, a targeted quantification of 110 cancer-related metabolites was analyzed simultaneously with untargeted profiling of 4651 features.
In the targeted workflow, a total of 78 metabolites were confirmed in Parallel Reaction Monitoring (PRM) and data-independent All-Ion Fragmentation (AIF) methods, of which 88 were confidently identified and precisely quantified. Meanwhile, in the untargeted workflow, all 4651 metabolites analyzed had high reliability and validity. By merging targeted and untargeted approaches into one injection and workflow, the semi-targeted workflows used in this study provided the authors with a flexible, accurate and confident method for large-scale metabolomics analysis.
The Semi-Targeted Analysis Workflow
Figure 1 illustrates key aspects of the semi-targeted metabolomics workflow: utilizing high-resolution accurate mass spectrometry on Orbitrap technology; sophisticated data processing and analysis application solutions for targeted confident identification and quantification of metabolites; and untargeted differential analysis and unknown annotation for biomarker discovery.
Overcoming Barriers to Adopting Semi-Targeted Analysis
Despite the clear advantages that semi-targeted analysis offers for metabolomics, its adoption has been relatively slow. Why is this, and how are vendors overcoming these barriers to enable its wider use in laboratories?
First, to perform accurate and confident annotation with semi-targeted analysis, scientists must have access to pure and diverse chemical standards for use in the MS workflow development. This is to confirm the identity of metabolites by matching multiple properties such as separation retention time, accurate molecular mass and MS fragmentation patterns. While many of these authentic metabolite standards are commercially available, there are plenty that are not. The main barrier, therefore, is the need for custom chemical synthesis, when true unknown compounds are identified, which significantly increases the cost. The untargeted part of semi-targeted workflows can help the community by providing insights toward important and relevant metabolites to be synthesized, which could then feed the targeted part of future semi-targeted studies.
Second, high-quality data must be obtained to ensure a confident interpretation of the results. In short, analyte data measured with LC-MS must have mass accuracy, high resolution, and biological and statistical robustness to ensure confident annotation of unknown metabolites and accurate differential analysis. While older LC-MS systems have often struggled to meet these demands for semi-targeted metabolomics, the latest commercially available Orbitrap-based MS technologies and approaches are designed to do exactly that — providing researchers with the tools to gain high-resolution MS data. Moreover, they are also ideal for performing large-scale experiments with high throughput and robust analysis, helping laboratories to analyze large numbers of samples with high consistency and replicability.
As discussed earlier, one of the benefits of semi-targeted analysis is enabling scientists to dive deeper into the sample, while providing an overview of known metabolites. One way in which vendors are helping scientists to achieve a deeper coverage of the metabolome is by creating intelligence-driven data acquisition strategies, which maximize the number of relevant compounds interrogated by MS/MS. These strategies offer several benefits to users, including integrating independent experiments into automated workflows. This thereby increases the efficiency and ease-of-use of LC-MS-based semi-targeted analysis, resulting in deeper coverage of the metabolome with higher confidence in the annotation.
Importantly, data processing is another key component of semi[1]targeted analysis, and retro-mining data requires powerful data analysis capabilities. As such, laboratories need sophisticated and comprehensive software solutions that any scientist can learn to use without a major training burden. These solutions should be able to enable fast data processing and analysis with accurate quantification of metabolites. Moreover, data solutions should facilitate differential analysis and confident metabolite annotation for accurate discovery analysis. Luckily, powerful software solutions are becoming more readily available to overcome these challenges. Thermo Scientific™ TraceFinder™ and Compound Discoverer™ applications, for example, have demonstrated an extended capability for metabolite quantification and biomarker discovery, respectively (Figure 1). Such systems unify data analytics and retro-mining processes along with spectral libraries and databases for compound identification/ annotation, statistical analysis and biological interpretation.
Can Semi-Targeted Workflows Increase Access to Metabolomics?
Semi-targeted workflows offer laboratories a new way to approach metabolomics. By enabling simultaneous acquisition of hypothesis-led and discovery-led datasets, semi-targeted workflows allow scientists to gain more knowledge about biological systems in a single experiment. Weighing up the pros and cons of untargeted and targeted metabolomics has often held back the tremendous potential of metabolomics in life sciences research. Now, semi-targeted analysis provides the best of both worlds to unlock this potential for scientists worldwide.
As a member of the vertical marketing metabolomics team at Thermo Fisher Scientific, Bashar Amer aims to develop robust and sophisticated metabolomics applications by employing state-of-the-art technologies and best practices to meet market needs. Before joining the Omics team at Thermo Fisher Scientific in San Jose, California, Bashar developed and implemented targeted and untargeted LC-MS metabolomics workflows to support microbial- and plant-based metabolomics studies at the Lawrence Berkeley National Lab in Emeryville, California. Bashar received a Ph.D. degree in GC-MS and LC-MS nutritional metabolomics from Aarhus University in Denmark.
Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special
offers from American Pharmaceutical Review – all delivered right to your inbox!
Sign up now!