Metabolomics and Biomarkers: Unveiling the Secrets of Biological Signatures

Unraveling Metabolomics: Exploring Concepts and Characteristics

Metabolomics is a science that studies biological systems (cells, tissues, or organisms) by examining the changes in their metabolites after stimulation or perturbation (e.g., after mutation of a specific gene or environmental changes) or changes over time. The metabolome, which represents the downstream and final products of the genome, is a collection of small molecular compounds. They are mainly endogenous small molecules with a relative molecular mass of <1000 that are involved in the metabolism of an organism and in the maintenance of its normal functioning in growth and development.

The nomenclature of metabolomics used to be controversial, with two terms, metabolomics and metabonomics, present in the international arena. It is generally believed that metabolomics is a technique to study the metabolic pathways of biological systems by examining the changes in their metabolites after stimulation or perturbation or over time. In contrast, metabonomics is the study of the qualitative and quantitative dynamic changes of metabolites produced by an organism in response to pathophysiological stimuli or genetic modifications. While the former typically focuses on cells as the object of study, the latter places greater emphasis on body fluids and tissues derived from animals. The term metabolomics is commonly used in botanical and microbiological sciences, while metabonomics is typically used in pharmaceutical research and disease diagnosis. The boundaries between these two definitions have been blurred in such a way that there is no longer any particular distinction between the two.

Metabolomics is characterized by:

1) Focus on endogenous compounds.

2) Qualitative and quantitative study of small molecule compounds in biological systems.

3) The up-regulation and down-regulation of the above compounds are indicative of the effect of disease, toxicity, genetic modification, or environmental factors.

4) The above endogenous compounds can be used for disease diagnosis and drug screening.

Metabolomics has the following advantages over transcriptomics and proteomics:

1) Small changes in the expression of genes and proteins are metabolically amplified by the complex biochemical processes of the body, making them relatively easier to detect.

2) Metabolomics research does not require the establishment of specialized databases for whole genome sequencing and transcript expression.

3) The variety of metabolites is much smaller than the number of genes and proteins (even the smallest bacterial genome has several thousand genes).

4) The metabolite species do not vary much across tissues, allowing for greater versatility in the techniques used in the study.

5) Instruments used in metabolomics are quite versatile, and most of the instruments currently used for toxicology (drug) analysis can be applied to metabolomics, facilitating the research.

6) Compared with other “omics”, metabolomics is affected by more factors, and the body's metabolomics response is more sensitive to external influences, making metabolomics research more informative.

The Evolution of Metabolomics: Tracing Its Path to Advancement

Metabolomics is a new discipline developed in the mid-1990s, but the application of metabolism has a long history. As early as ancient times, Chinese doctors used ants to assess glucose in urine for the diagnosis of diabetes. In the Middle Ages, doctors identified the metabolism of different individuals by mapping their urine by different colors, smells, and tastes, all of which were, in fact, the epitome of metabolomics. It was not until the late 1940s that metabolic profiling was first proposed by Roger Williams and others, who used paper chromatography to study the relationship between metabolites in urine and schizophrenia. In the early 1970s, Greg Horning et al. introduced metabolic profiling for the first time using gas chromatography, enabling the qualitative and quantitative determination of metabolites in urine. In the 1980s, high-performance liquid chromatography (HPLC) and nuclear magnetic resonance (NMR) were used for metabolic profiling. For example, in 1983, Sadler, Buckingham, and Nicholson published the first 1HNMR spectra on whole blood and plasma. In 1997, liquid chromatography coupled with mass spectrometry (LC-MS/MS) technology was used for pharmacokinetic high-throughput screening. In 1999, J. Nicholson et al. proposed the concept of metabonomics and did a lot of fruitful work in disease diagnosis, drug screening, etc., and is thus regarded as the “father of metabolomics”. In 2013, Metware Bio developed the widely-targeted metabolomics technology, allowing metabolite detection to enter the era of high-throughput detection; at the same time, Metware Bio developed the concept of metabolome + genome research, and their results were published in Nature Genetics. Subsequently, it has continued to enrich metabolomics techniques with the introduction of high-throughput assays such as TM widely-targeted metabolome detection and full-spectrum metabolome detection.

Metabolomics Today: Trends and Biomarker Insights

Biological organisms in a normal state are in internal homeostasis. When the organism is subjected to external interference, the internal homeostasis will be broken, which leads to the up-regulation or down-regulation of certain metabolic pathways. Metabolomics is a discipline that employs high-throughput, high-precision instrumental measurements and advanced data analysis methods to capture changes in metabolic pathways in organisms and combines them with other means to search for relevant specific biomarkers.

The human metabolome blueprint was initially determined in 2007, which includes 2,500 metabolites, 1,200 drugs, and 3,500 food components found in the human body. The metabolome is part of the Human Metabolome Database, which will help researchers discover the location of metabolites, normal and abnormal concentrations of metabolites, and the links between metabolites and genes. Only 2,500 metabolites have been identified as biomarkers of human metabolism compared to 25,000 genes and about 1 million proteins. The limited number of metabolite biomarkers allows them to be analyzed in an easier and more quantitative way.

Metabolomics has undergone rapid development in recent years, with major advances in both detection and identification of unknowns. However, the metabolome is easily affected by food, lifestyle, environment, season, and even emotions. Changes in a single base of DNA can lead to 100,000-fold changes at the metabolite level. As a result, the identification of metabolomics biomarkers remains sluggish, with less than one percent of known metabolites currently used in routine clinical testing.

Metabolomics Biomarker Research: Standard Processes and Protocols

The typical process of metabolomics-based biomarker research includes biological sample collection and preprocessing, metabolite detection, data preprocessing, data analysis, marker screening, and performance validation.

1 Sample collection and preprocessing

Metabolic analysis samples are divided into biological fluids, cells, tissues, feces, etc. Current research has largely focused on biological fluids such as serum, plasma, urine, bile, and cerebrospinal fluid. When the samples are selected, it should be noted that the parameters between the research populations should be as close as possible; the time, site, and type of samples collected should be consistent; and the influence of external factors such as living habits, day and night rhythms, medications, mental status, and geographical areas should be considered. The samples are generally stored at -80°C after collection, and the transportation process is kept at a low temperature with dry ice. The sample should be processed in such a way as to ensure the stability of the components in the sample and to avoid introducing changes.

2 Platforms for metabolite detection

(1) Nuclear magnetic resonance (NMR) platform

NMR requires no or little preprocessing of the sample, allowing for non-destructive testing of liquid or solid samples, such as tissues, cell extracts, or whole organs, yielding substantial structural information, in addition to its high stability. Therefore, it is widely used in metabolomics research. However, the NMR method also suffers from the disadvantage of being less sensitive and unable to detect some compounds with lower content. In some cases, it may be the changes in trace amounts of certain substances that predict the significant progress of the disease, thereby limiting the application of NMR in metabolic marker research.

(2) Mass spectrometery (MS) platform

MS-based metabolomics provides the ability to quantitatively analyze and identify metabolites with high selectivity and sensitivity. In addition, the combination with various separation techniques reduces the complexity of MS in the time dimension and offers additional information regarding the physicochemical properties of the metabolite, facilitating metabolite identification. Some such examples include gas chromatography/time-of-flight mass spectrometry, liquid chromatography/time-of-flight mass spectrometry, capillary electrophoresis/time-of-flight mass spectrometry, and liquid chromatography/ion trap mass spectrometry, among others. Further details of metabolomics detection technologies based on mass spectrometry platforms will be discussed in subsequent chapters.

3 Data preprocessing

The raw metabolomics data contains quality control (QC) samples and test samples. In order to better analyze the data, a series of preprocessing of the raw data is required, which mainly includes normalization of the raw data, screening of outliers, processing of outlier samples, and filling of missing values. Data preprocessing can reduce the impact of variants in the data that are not relevant to the purpose of the study on data analysis, facilitating the screening and analysis of potential target differential metabolites.

4 Data analysis

Metabolomics research generates a large amount of data, which have complex characteristics such as high dimensionality and high noise. How to extract valuable information from complex metabolomics data and screen out potential biomarkers has become a hotspot and difficulty in metabolomics research in recent years. The basic process of metabolomics data analysis includes overall metabolite analysis (such as PCA, OPLSDA, and cluster analysis), differential metabolite screening based on FC, p-value, or VIP value, and differential metabolic function analysis based on KEGG and HMDB databases.

5 Marker screening and performance validation

Screening of clinical biomarkers and the construction of an optimized diagnostic panel are the pre-requisite basis for translating into clinical applications. One of the main challenges in marker screening is how to efficiently obtain potential biomarkers with high sensitivity, stability, and accuracy from massive metabolomics data. Advanced machine learning has been widely applied to medical biomarker screening over the past few years, addressing these challenges to a large extent, which makes it a must-have tool for marker screening. Therefore, we will provide a detailed review of several machine learning algorithms, such as logistic regression, Lasso regression, elastic network regression, support vector machines, etc., which are commonly used in biomarker screening for metabolomics, as well as multiple methods for validating the performance of the markers in the subsequent chapters.