What is metabolomics?

The metabolome is defined as the collection of all metabolites found in a cell, an organ, or an organism, whereas metabolites are the small molecules that are synthesised, transformed or excreted, and used as building blocks, signals or fuel by living beings.Metabolomics aims at identifying and quantifying all these metabolites. When subclasses of metabolites are targeted, the prefix is changed, as for example lipidomics (hydrophobic compounds) or ionomics (ions). The term “metabolome” was first proposed by Oliver et al. in 1998 who were investigating yeast using a functional genomics approach.

While identifying or quantifying metabolites was initially associated to diagnostics, metabolomics has much more to offer. In metabolomics, one studies the small-molecule metabolites that both result from and influence cellular processes, as a way of learning about cellular biology, systems biology, and disease. Metabolite levels and fluxes are often seen as being at the receiving end of the information flow from the genotype to the phenotype, but such a view is too simplistic. Many metabolites provide feedback inhibition to protein expression, and thus regulate the flux of metabolites through the system, and others signals or hormones affect gene expression and activity. Thus, there is not a linear flow of information from the genotype to the phenotype; rather, a systems view will provide a more accurate understanding of the mechanisms of biology, physiology, and disease. Given the complexity of biological systems, studying the metabolome is a challenging and undertaking.

Metabolomics for what purpose?

Metabolomics can be used to diagnose the quality and nutritional value of food, but also identify potentially bioactive molecules. For example, molecules of defence against pathogens can be discovered, as well as metabolites associated with resistance to drought, cold, or nutrient deficiency. Then, the use of metabolomics in quantitative genetics can combine such metabolic traits with molecular markers, eventually leading to a better understanding of how metabolism and plant performance are integrated, paving the way for new selection strategies.

Recent advances even allow the uncovering of the molecular bases of quantitative traits in one single experiment. Although they require the handling of very large numbers of samples, strategies of high-throughput metabolic phenotyping are already available. Metabolomic data can also be integrated with other data types (genomic, transcriptomic, proteomic, ecophysiological…) using multivariate analyses. Integrating various dimensions of life can be viewed as integrative biology. For example, the discovery of correlations between metabolite levels and gene expression enables the generation of new hypotheses about how metabolites are perceived and regulated; hypotheses that can then be tested using reverse genetics approaches. In the emerging field of systems biology, metabolomics also appears as a key step, as it enables the parameterization and validation of models that predict metabolic fluxes or concentrations of molecules of interest.

Finally, the metabolomics community has begun to organise and store metabolite data, but also information about the samples under study, i.e. “metadata”; about taxonomy, organ or cell type, date of collection, environmental conditions and technical details of the analysis. This apparently bureaucratic habit will actually enable future data mining approaches, which will soon revolutionise the exploration of the metabolic diversity of plants by adding an explanatory dimension.

What methodologies for metabolomics?

Metabolomics are often defined as non-targeted, which implies little or no a priori on the molecules to be detected. Ideally, every molecule of a given metabolome should be detected. However, due to the huge physical and chemical diversity of metabolites, it is not yet possible to detect all of them at the same time. As a consequence, complementary methods of extraction and analysis have to be used.

Nuclear magnetic resonance (NMR) and mass spectrometry (MS) are nowadays the most widespread technologies used to perform untargeted metabolomics, both offering a wide range of applications. NMR can detect and quantify metabolites in complex mixtures with high reliability. It is also used to track the fate of isotopes in metabolic pathways, cells or living tissues.

Finally, NMR is used to determine the structure of molecules. MS, coupled with gas chromatography (GC-MS) or liquid (LC-MS), has the advantage of being more sensitive than NMR and despite a less efficient structure resolution and a lower reproducibility of results it can detect a much greater number of metabolite signatures.

The development of mass spectrometers with increasing mass resolution already enables the detection of several thousands of analytes in a sample. Untargeted methods are demanding when it comes to transforming raw data into concentrations of metabolites, an operation that requires sophisticated algorithms, databases and much expertise. An important consequence is slowness and indeed, typical experiments do not exceed a few hundreds of samples. In some cases, however, it is possible to work directly on the shape of the spectra (proton NMR) and mass chromatograms (LC-MS and GC-Q-TOF) using multivariate analysis. Thus, groups of samples and discriminating markers can be distinguished quite easily. This approach called”fingerprinting”; enables the processing of large numbers of samples. Another way to increase the throughput of samples without breaking the cost of analysis is to use targeted approaches. In that case, only a few information-rich metabolites are measured, which dramatically speeds up the analysis of the raw data. LC-MS/MS (liquid chromatography coupled to a triple quadrupole MS), which allows to target up to tens of metabolites with great sensitivity (atomoles) is a promising technology for that purpose.

Finally, conventional methods involving chemical or biochemical reactions whose products are measured by spectrophotometry, fluorimetry or luminometry can advantageously be used in high-throughput. They can be performed with microplates, which enable the parallel processing of tens to hundreds of samples (actually up to 9600!) and are well suited to automation. Last but not least, metabolite data and their associated metadata are best stored in public databases using dedicated concepts and software, which are under intense development, eventually leading to their re-analysis by the scientific community.

Are structural analyses necessary for metabolomics?

Structural analyses are used for metabolite annotation and provide information for 3D structures. As mentioned above, two-dimensional NMR enables the determination of the structure of unknown compounds. It is for example essential to the study of the so-called secondary metabolites, which are highly diverse in plant and involved in many mechanisms.

Such identification requires at first isolation of the compounds under study, a step which is slow and expensive. The coupling between liquid chromatography and NMR spectroscopy can nevertheless overcome this limitation. The use of LC-MS also proves useful, as it allows a preliminary identification of compounds in complex extracts. The coupling of LC-MS and LC-NMR actually allows a significant time saving by restricting the slow and expensive LC-NMR analysis to unknown compounds.

Two-dimensional NMR coupled with molecular modelling allows the analysis of the 3D structures of the compounds under study. Indeed, the spatial conformation of these compounds is often correlated with observed biological activity. These techniques are typically used also to study the interactions between secondary metabolites such as polyphenolics and macromolecules such as proteins.

What is the application of metabolomics?

There are four conceptual approaches in metabolomics: target analysis, metabolite profiling, metabolomics, and metabolic fingerprinting. Target analysis has been applied for many decades and includes the determination and quantification of a small set of known metabolites (targets) using one particular analytical technique of best performance for the compounds of interest. Metabolite profiling, on the other hand, aims at the analysis of a larger set of compounds, both identified and unknown with respect to their chemical nature. This approach has been applied for many different biological systems using GC-MS, including plants, microbes, urine, and plasma samples. Metabolomics employs complementary analytical methodologies, for example, LC-MS/MS, GC-MS, and/or NMR, in order to determine and quantify as many metabolites as possible, either identified or unknown compounds. The fourth conceptual approach is metabolic finger-printing (or footprinting for external and/or secreted metabolites). Here a metabolic “signature” or mass profile of the sample of interest is generated and then compared in a large sample population to screen for differences between the samples. When signals that can significantly discriminate between samples are detected, the metabolites are identified and the biological relevance of that compound can be elucidated, greatly reducing the analysis time.

Since metabolites are so closely linked to the phenotype of an organism, metabolomics can be used for a large range of applications, including phenotyping of genetically modified plants and substantial equivalence testing, determination of gene function, and monitoring responses to biotic and abiotic stress. Metabolomics can therefore be seen as bridging the gap between genotype and phenotype, providing a more comprehensive view of how cells function, as well as identifying novel or striking changes in specific metabolites. Analysis and data mining of metabolomic data sets and their metadata can also lead to new hypotheses and new targets for biotechnology.

What is the difference between proteomics & metabolomics?

Metabolomics can be used to determine differences between the levels of thousands of molecules between a healthy and diseased plant. Proteomics studies dynamic protein products and their interactions, while metabolomics is also an intermediate step in understanding organism’s entire metabolism.Why hasn’t metabolomics taken off in industry?
Metabolomics is being used heavily in industry, for example, in the agro-biotechnology business. Monsanto (U.S.) and Metanomics (Germany) have a large-scale agreement on use of intellectual property that was gained by industry-sized metabolomics screenings of hundreds of thousands of plants. Metabolomics is also used
heavily in the biomedical industry, including pharma, again using a service provider (Metabolon, Durham, North Carolina). Metabolon has gained several rounds of investment and has a strong standing in the market. There are ranges of other metabolomics service providers, from Stemina (Madison, Wisconsin), to Biocrates (Austria) and multiple others. Within the pharma or agbiotech industry, it is simply cheaper to outsource services for large-scale operations than to implement all equipment and expertise in-house, although most companies have smaller in-house teams as well to validate metabolomics findings.

What are the main bottlenecks occur in metabolomics?

The biggest bottleneck is correct compound annotations. There are about 40,000,000 small molecules represented in PubChem, but all publicly available tandem mass spectrometry (MS-MS) libraries (Metlin, NIST14, and Mass Bank) combined only comprise 40,000 unique compounds. So currently we have at best 0.1% coverage. That is just a sad state of affairs, mostly a result of poor funding decisions by the scientific agencies. Some scientists may argue that there are only 15,000 compounds in biochemical databases like KEGG, but others correctly reply that we are entering the era of the “exposome,” integrating data and knowledge of both endogenous metabolites and exogenous compounds such as food, cosmetics, drugs, and pesticides, all of which have enormous implications on health and disease.