Metabolomics identifies and quantifies both known and unknown metabolites resulting from biological pathways, providing an understanding of health vs disease in both static and dynamic states. Resulting data offers a level of analysis above and beyond that obtained from genomics and proteomics by measuring phenotype. The ability to get a comprehensive view of pathway metabolism delivered in a timely manner makes metabolomics attractive to researchers answering various biological questions.
Technological advancements have continued to bring metabolomics into the forefront of biomarker discovery, drug discovery, and personalized medicine. From improvements in instrumentation that both enhance analytical sensitivity and capture more data from a single experiment to innovations in software analysis, metabolites are fast becoming integral to biological investigations as well as in the clinic.
The balance between confidence and coverage
Recently scientists have been gaining a better understanding of phenotype, the ability to distinctly measure disease characteristics within the body instead of inferring or predicting risk for a disease that may or may not develop. Phenotype analysis is expanding as advancements in techniques offer quicker, easier, and more comprehensive metabolite characterization.
Precision Metabolomics, a technology platform developed by Metabolon, comprehensively measures metabolism, offering the capability to both discover and then monitor molecular phenotypic change longitudinally. It provides the ability to observe individuals over time, learn how their metabolism is changing as disease develops, and identify markers of early disease as well as markers that indicate staging of disease, ultimately permitting assessment of whether treatment or lifestyle changes are effective.
“It is simplistic to try to single out an individual technology as being key to our metabolomics abilities. The subtler truth is that successful metabolomics experiments require a significant infrastructure behind them,” says Mike Milburn, Metabolon’s CSO. High-resolution mass spectrometry (MS) coupled with fast liquid chromatography (LC) is a core component of providing the sensitivity and power to support identifying hundreds of metabolites from a sample. In combination with an extensive software infrastructure that sifts through the massive amounts of data these instruments produce and an extensive library of curated reference standards, metabolites can be rapidly and accurately analyzed.
When discussing metabolite characterization using the latest technologies, Drew Jones, assistant professor at New York University and director of the Metabolomics Core Resource Laboratory, highlights how confidence in results is influenced by the amount of coverage an experiment aims to focus on. “Balancing confidence and coverage in metabolomics applies to how one can approach an experiment,” he explains. At the top of a pyramid, researchers observe the least amount of coverage but can obtain the highest confidence in measurements using verified standards. At the bottom of the pyramid is untargeted analysis with a wide coverage range but confidence levels drop. “Applying the latest technologies is allowing a switch to untargeted studies, where more and more detailed information can be taken from one study with higher confidence.”
Jones mentions that MS advancements in HILIC (hydrophilic interaction liquid chromatography) columns, in particular, offer retention of compounds that had previously only been available for analysis by GC-MS or ion pairing reversed phase methods. Specific improvements in HILIC columns allow a flexibility for solvent switching and greater access to the polar metabolome, delivering increased sensitivity by LC-MS and a growing data source. One such column is Millipore’s SeQuant HILIC HPLC columns designed to separate polar hydrophilic compounds by hydrophilic partitioning combined with weak ionic interactions for greater selectivity and easier method optimization.
Biocompare’s Mass Spec Search Tool
Find, compare and review mass spec
tools from different suppliers Search
Advances in UHPLC and MS combinations creates an extremely powerful tool, delivering increased capacity and faster analysis and bringing a greater focus to larger epidemiological studies, biobank screening, and improved decision-making for metabolite functions and applications. Ion mobility mass spectrometry as in Waters SYNAPT and Vion systems, combine ion-mobility separations with high-resolution MS, allowing measurement of collision cross section values specific to each metabolite.
Jose Castro-Perez, Ph.D., director of biomedical research, global markets at Waters, explains, “Obtaining molecular weight and fragmentation patterns of a metabolite allows for a greater degree of confidence with first pass analysis without the need for synthesizing large numbers of metabolites, which can be extremely cost-prohibitive. In silico tools using the power of ion mobility can really help identify ‘known’-unknowns.”
The big data challenge
The ability to use and understand all of this data collected from better identification and quantification techniques is a current bottleneck in metabolomics. Even with comprehensive databases available containing larger numbers of entries, classification of metabolites is time-consuming and can be hit or miss. Jones says software is key to making advances in the field, especially moving from narrow targeted assays to verifying metabolites in large batches from increasingly untargeted and global analyses, mirroring advancing proteomics approaches.
The ability to use and understand all of this data collected from better identification and quantification techniques is a current bottleneck in metabolomics.
Many labs like Jones’ and Metabolon’s develop their own software that can effectively mine data and define biological significance such as associating a novel metabolite with the microbiome or a specific cancer cell type. Commercial bioinformatics tools like Waters’ Progenesis QI provide differential analysis to compare controls with disease data, using an extensive database to search for hits. Additional options include software suites, including SimMet by PREMIER Biosoft, specific to metabolite analysis that facilitates LC-MS data processing, metabolite identification, quantification, and statistical analysis.
“The big data challenge of multi-omics must be addressed,” asserts Milburn. Fortunately, the continued development of software tools and AI solutions to support metabolomics data interpretation, particularly in conjunction with other -omics data, is beginning to receive more attention.
Castro-Perez refers to machine learning (ML) and artificial intelligence algorithms as the future of metabolite analysis. Applying ML helps identify metabolites that are not a hit on any database, combining all data including cross-sectional molecular measures to provide more value. ML analysis also offers a higher degree of confidence when doing untargeted studies, where ML algorithms integrated into public databases could progressively improve the accuracy of results.
Milburn agrees, “While data acquisition and the data itself have focused much attention on the why and how of global metabolomics, another key advancement has been in developing data interpretation tools using artificial intelligence and software engineering. We’re investing heavily in this area to continuously refine our knowledge infrastructure and tools that support a scientist’s ability to obtain actionable information from the data, rather than just providing a large spreadsheet with a list of assorted metabolites.”
Clinical translation and personalized medicine
Metabolomics applications are rapidly expanding across every area of the life sciences, as well as moving into the clinic. Metabolon has identified clear trends in recent years where metabolomics is increasingly being applied in population studies to identify phenotypic biomarkers for disease and disease mechanisms in cohorts with hundreds to thousands of individuals, precision medicine research and diagnostic use, and areas that have up to this point been predominantly focused on genomics technologies, including microbiome research.
Castro-Perez adds that as technology is becoming more accessible and easier to implement, more discoveries can be made with a greater potential to translate biomarkers to the clinic. Converting biomarkers to actionable health benefits could be successful with the development of metabolomic fingerprints. Examining a patient’s metabolome and basing treatment on personalized patterns could bridge the gap between the pharmacodynamic effect of a drug and efficacy.
We already see this with cancer therapies, where human cells are extracted and tested ex vivo to assess drug effects before treating a patient. Integrating metabolomics data and supplementing with metabolite profiling would provide further information for a more informed assessment. Given cancer’s extremely heterogeneous nature, personalization to this extent would add incredible value. The ability to derive genomic, proteomic, and metabolomic information in an efficient manner could shape pharmacodynamics and predictive biomarker assessments using liquid biopsies to provide comprehensive real-time molecular examination of disease.
MS imaging using desorption electrospray ionization (DESI) is also gaining traction by complementing metabolomics analysis in clinical research, providing spatial distribution assessed in conjunction with peak analysis. All of these tools can help metabolomics evolve with the ability to analyze data and translate to clinical advancements, ultimately intersecting the value proposition of metabolomics with improving human health.