Recent advances and prospects of computational methods for metabolite identification: a review with emphasis on machine learning approaches

Abstract

Metabolomics involves studies of a great number of metabolites, which are small molecules present in biological systems. They play a lot of important functions such as energy transport, signaling, building block of cells and inhibition/catalysis. Understanding biochemical characteristics of the metabolites is an essential and significant part of metabolomics to enlarge the knowledge of biological systems. It is also the key to the development of many applications and areas such as biotechnology, biomedicine or pharmaceuticals. However, the identification of the metabolites remains a challenging task in metabolomics with a huge number of potentially interesting but unknown metabolites. The standard method for identifying metabolites is based on the mass spectrometry (MS) preceded by a separation technique. Over many decades, many techniques with different approaches have been proposed for MS-based metabolite identification task, which can be divided into the following four groups: mass spectra database, in silico fragmentation, fragmentation tree and machine learning. In this review paper, we thoroughly survey currently available tools for metabolite identification with the focus on in silico fragmentation, and machine learning-based approaches. We also give an intensive discussion on advanced machine learning methods, which can lead to further improvement on this task.

From Samples to Insights into Metabolism: Uncovering Biologically Relevant Information in LC-HRMS Metabolomics Data

“…To address this problem, recently, the computational metabolomics community has grown to develop and improve computational approaches for known and unknown metabolite identification (Table 3). These computational metabolomic approaches employ two main strategies: (1) In silico prediction of fragmentation MS/MS spectra from chemical structures of known compounds, and (2) in silico prediction of molecular substructures (i.e., molecular fingerprints or feature vectors that encode the structure of a molecule) and general chemical properties of the unknowns from experimentally acquired MS/MS spectra [112]. With the in silico fragmentation methods, the experimentally acquired spectra of an unknown metabolite (for which reference spectra are not available) can be matched against in silico theoretically predicted spectra simulated on known candidate structures retrieved from databases (Human Metabolome Database (HMDB), PubChem, KEGG, etc.)…”

“…To learn the mapping of an MS/MS spectrum to a molecule structure, these methods need to be trained on spectral databases of known metabolites. In general, machine learning methods can be divided in two groups, supervised learning for substructure prediction (e.g., CSI:FingerID) and unsupervised learning for substructure annotation and grouping of metabolites based on shared, biochemically relevant substructures (e.g., MS2LDA) [112,[114][115][116]. The main objective of supervised methods, such as CSI:FingerID integrated in Sirius tool, is to determine, using a database of molecular structures, the structure that best fits the experimental data.…”

MESSAR: Automated recommendation of metabolite substructures from tandem mass spectra

“…Structure elucidation from MS/MS data has always been a challenging and time-consuming task with a vast number of potentially interesting metabolites that are still unknowns. The main reason is that current MS/MS databases (spectral libraries) only contain a limited number of historical spectra, far below the number of metabolites in reality [3,4]. Advances in computational tools have led to a considerable extension of the search space that can be examined and have resulted in an improvement of the identification accuracy by using massive molecular databases (for example, PubChem currently contains about 100 million compounds [5]).…”

Probabilistic Framework for Integration of Mass Spectrum and Retention Time Information in Small Molecule Identification

“…In recent years numerous powerful approaches (Nguyen et al, 2018a;Schymanski et al, 2017) for annotating MS 2 spectra with a predicted molecular structure have been developed (Ruttkies et al, 2016(Ruttkies et al, , 2019Dührkop et al, 2015;Brouard et al, 2016;Allen et al, 2014;Nguyen et al, 2018bNguyen et al, , 2019Dührkop et al, 2019). Typically, these methods output a ranked list of molecular structure candidates, that can be shown to human experts, or further post-processed, e.g.…”