The Five Key Elements of a Successful Metabolomics Study

Metabolomics: Completing the Biological Picture

Metabolomics offers new insights into systems biology, empowering biomarker discovery, and advancing research across the life sciences. Metabolomics is the systematic measurement of the low molecular weight biochemicals (~50 to 1,500 Da) known as metabolites. Nearly every variable, from genes and microbiota to disease and diet, exerts influence by subtly altering metabolite levels. In this way, metabolites are sensitive indicators of homeostatic imbalances. Measuring metabolites using metabolomics improves our understanding of how disease, drug exposure, genetics, the microbiome, diet, and lifestyle all influence the phenotype.

Leveraging Metabolomics to Advance Research

Researchers across multiple areas of the life sciences are using metabolomics to answer key research questions and advance programs in translational science, biomarker identification, and the study of disease and treatment mechanisms. While metabolomics has proven itself to be an incredibly valuable research tool, metabolomics studies also have unique characteristics to consider.
This blog will provide you with the five key elements of a successful metabolomics study:

1. Define a Clear Objective
2. Use Strong Study Design Elements
3. Power Your Study for Success
4. Choose the Right Profiling Technology for Your Goals
5. Have a Data Interpretation Plan in Place

1. Define a Clear Objective

The first and most critical step in any successful scientific study is to clearly define the study’s objective. Are you seeking general information to help you form a hypothesis? Do you have a hypothesis in mind that you wish to validate? Are you hoping to discover biomarkers for a disease? Do you want to understand the mechanism of action (MOA) of a potential drug candidate? Metabolon offers every investigator study design assistance from an experienced study director with a Ph.D. background. A clear objective is critical to identifying the ideal metabolomics approach for your study. For example, if you want a comprehensive overview of the changes occurring in the metabolome between a disease and healthy cohort or want to measure an intervention on different groups, then an untarageted metabolomics approach is suitable. On the other hand, if you already have some understanding of the metabolites implicated in your study, you may want to narrow your focus and hone in on specific metabolite classes (such as lipids), metabolic pathways (such as the glycolysis pathway), or a biomarker.

2. Use Strong Study Design Elements

Strong study design elements are central to uncovering biologically significant results.

• Select the appropriate sample matrix (or combination of matrices).
• Collect adequate exposures (dose and time of collection).
• Employ controls for each tested variable—don’t skimp!
• Take steps to minimize excess variation—maintain consistency!

Depending on what you are hoping to uncover, select an appropriate sample matrix or matrices. Try to use the matrix or combination of matrices that are “closest” to the system you are attempting to study. For example, when studying the effect of drug exposure on the brain, a CSF sample will provide significantly more information than a urine sample. On the other hand, urine would be an appropriate matrix if you are attempting to study a treatment effect on kidney function. When using animal models, it may be advisable to include organ samples, in addition to plasma, which could improve understanding of the physiology’s precise site of action.

Sampling at only a single exposure point is a gamble and may result in too few metabolic changes to derive any biomarkers or enhanced understanding. At the other extreme, a single point sample may reveal so many changes that it can be difficult to assign cause and effect. Multi-exposure point studies, such as a time-course study or a multi-dose study, increase your chance of seeing meaningful variations. A good rule of thumb is to collect samples at “mild,” “moderate,” and “severe” exposures to ensure that all salient cause and effect changes are captured.

Sometimes our desire for robust results around a research project makes us overly ambitious. It’s important to avoid the temptation of testing multiple variables unless you are prepared to have each variable accompanied by rigorous controls. Embedding too many questions in a study without sufficient controls built around each question can leave you with a lot of data and few meaningful answers. Only incorporate as many questions and variables as you can realistically test with adequate controls.

For every study, you need to consider how much variance is inherent in the cohort or study samples. For example, inbred rodent models have significantly less variability than human study subjects with different ages, genders, genotypes, BMIs, and lifestyles. To accommodate the off-target variables, collect adequate samples, employ strong controls, and attempt to balance the groups where appropriate. Also, consistency in sample collection and handling is essential. Having rigorous collection and handling protocols minimizes variance so that you can obtain meaningful results with fewer samples.

For more information on Study Design Success, download our Guide.

3. Power Your Study for Success

A strong study design can deliver biologically significant results, but a well-powered study can provide statistically significant results. Several factors can influence the number of samples required for a given study. Studies with a strong phenotype or treatment effect, such as a toxicology study, can require fewer samples. Studies with repeated sampling from the same subject, multiple time points, or multiple doses of a drug/inhibitor can also require fewer samples. Conversely, studies with a subtle phenotype or treatment effect, such as studies looking at dietary supplements or exercise-induced changes would require more samples. Studies with mixed populations of subjects (mixed gender, wide-ranging age, or BMI) or multiple-site collections would also require more samples to be sufficiently powered. Your study director will work with you to determine the optimal number of samples and quantities for your study. 

4. Choose the Right Technology for Your Goals

Metabolomics requires a different and more sophisticated profiling platform than many other types of “omic” research (such as genomic or proteomic research). Metabolites, which are small and highly abundant, are also extremely chemically diverse, and they vary substantially in their molecular weight and solubility. This makes it difficult for a single analytical method to accurately measure all of the metabolites present in a sample. In order to have meaningful study results, you should select a metabolomics “platform” (screening technology) with both advanced instrumentation and sophisticated informatics. Sensitive and precise instrumentation is necessary to detect all of the low-level metabolites present in a sample, while sophisticated informatics help rapidly and accurately identify and quantify these metabolites.

There are many approaches to generating metabolomics data. Some of the most popular approaches include liquid or gas chromatography coupled to mass spectrometry (LC/MS, GC/MS), nuclear magnetic resonance (NMR) spectroscopy, and capillary electrophoresis (CE) chromatography mass spectrometry methods. Each approach has pros and cons, and some may be more suited to your study than others. NMR, for example, does not destroy the sample, but it has a very limited breadth of metabolite coverage. CE allows you to profile with smaller sample quantities and great resolving power, but it has limitations for detecting “neutral” metabolites. UPLC/MS technology is among the most sensitive and highly resolving metabolomics technologies. LC-MS offers a broad and sensitive screening of most metabolite classes, while GC-MS and LC-MS+ HILIC are able to detect certain polar metabolites. For a metabolomics discovery study, you generally have the best chance of capturing the full range of metabolites present in a sample if you select a multi-system approach that runs multiple detection instruments in tandem.

Regardless of the instrumentation used, a single metabolomics study is likely to identify hundreds of metabolites and produce thousands to tens of thousands of data points. Attempting to identify metabolites and derive biological significance from the mountain of data collected can be time-consuming and resource-intensive. Many metabolomics technologies use an “ion-centric” approach. This approach requires manual identification of all of the metabolites using ion features. This method has a high potential for false discovery and ambiguous identification, leading to substantially fewer accurately identified metabolites. On the other hand, an automated, “chemocentric” approach incorporates sophisticated software tools that distill the thousands of ion features, eliminate noise, and automatically classify metabolites by key ion features (mass, retention time, and peak area). This approach requires an exhaustive chemical reference library of metabolite standards and sophisticated informatics software that can identify the metabolites. When this approach is used for broad, “whole metabolome” profiling, it ensures a more accurate measurement of the metabolites and their perturbation. It also allows for faster, easier, and more accurate statistical analysis.

5. Have a Data Interpretation Plan in Place

Nearly any metabolomics profiling technology will produce large amounts of complex data. You must have a plan in place for translating that data into meaningful biological insight. Your data interpretation plan should be driven by the “deliverables” you expect to receive from your metabolomics study. If you will receive only the raw study data, be prepared to dedicate substantial time and resources to verifying the metabolite assignments, performing statistical analysis, and researching the biological implications of any metabolic perturbations. Be aware that this process can be time-consuming and challenging, particularly if your analyst does not provide you with specialized interpretation tools. Make sure to budget your resources accordingly. On the other hand, if your metabolomics study provider offers specialized interpretation tools or verified statistical analysis and pathway interpretation, your resources can be devoted to conducting any recommended follow-up studies and preparing your work for publication.

When all of these components are present in your study, you dramatically improve your chances of deriving meaningful findings that translate into actionable results.