Workflow in Metabolomics
A metabolomics experiment starts with an appropriate design ensuring that the data will be relevant and statistical significant for further biological interpretation.
Biological matrices are very complex, and thus, a sample preparation step must be carried out prior to any metabolomic study. Metabolite extraction strongly depends on the type of biological and also on the chemical structures of the metabolites to be preferably detected. In an ideal metabolomic non-targeted approach, metabolite extraction method should not be biased towards any group of molecules.
Two different analytical approaches can be followed in non-targeted Metabolomics. While “metabolic profiling” is referred to analysing a subclass of metabolites (family or metabolic pathway), “metabolic fingerprinting” has been proposed as a means of analysing the total set of metabolites in a given sample. In the latter case the identity of the metabolites of interest is established after statistical data analysis of metabolic fingerprints. High- and ultra-high-resolution mass spectrometers are becoming increasingly popular in the field of metabolite profiling/fingerprinting because they provide accurate mass measurements which are useful for the discrimination between isobaric ions, and even isomers if their fragmentation patterns are different.
When the objective is to detect as many metabolites as possible in a complex sample matrix, and the number of samples is high, raw data processing is a key step in data analysis. High-throughput analysis of biological fluids, especially those which are obtained in a minimally invasive manner, will give a massive production of data. Thus, metabolomic analysis processes wil generate, especially at the output of automatic ion detection, a large matrix of data containing tens of thousands of variables (m/z, retention time, intensity). Bioinformatic tools play here an important role in order to develop strategies to convert the complex raw data obtained into useful information. Statistical analysis and data mining must be carried out to allow the identification of significant metabolites that capture the bulk of variation between datasets and that represent candidates that may serve as biomarkers.
Another challenge in Metabolomics is the metabolite identification process in a high-throughput manner. Today, many metabolites in complex biological samples are still not-annotated in databases. Due to the lack of comprehensive databases and the chemical complexity, accurate mass spectra must be exploited for structural elucidation of compounds. An accurate mass provides, at best, a molecular formula, not a chemical structure. Further experiments (MS/MS and standard coinjection) are necessary for confident identification. Tentative identification of metabolites is usually carried out by matching the obtained accurate m/z values and theoretical m/z values contained in different free available databases, such as Human Metabolome Database (HMDB) and Metlin.
To complete the metabolomic study, a pathway analysis of the identified molecules can be performed by means of some of the databases mentioned above. The identification of altered pathway after a intervention provides important information about the biochemical processes and possible consequences at a molecular level, enhancing hypothesis creation and new studies design.
Integration with other “omics”
To understand the organization of cellular functions at different levels (genes, metabolites, proteins) and link them to a particular phenotype, an integrative approach between different “omics” (Transcriptomics, Proteomics, Metabolomics) information is needed. Major challenges include interpretation and integration of large datasets to understand the principles underlying the regulation of genes, metabolites and proteins and also how their combined interactions associate with variation in phenotype. Bioinformatic tools play again an important role to integrate multiple “omics” data sets.