MESSAR: Automated recommendation of metabolite substructures from tandem mass spectra

Abstract

Despite the increasing importance of non-targeted metabolomics to answer various life science questions, extracting biochemically relevant information from metabolomics spectral data is still an incompletely solved problem. Most computational tools to identify tandem mass spectra focus on a limited set of molecules of interest. However, such tools are typically constrained by the availability of reference spectra or molecular databases, limiting their applicability of generating structural hypotheses for unknown metabolites. In contrast, recent advances in the field illustrate the possibility to expose the underlying biochemistry without relying on metabolite identification, in particular via substructure prediction. We describe an automated method for substructure recommendation motivated by association rule mining. Our framework captures potential relationships between spectral features and substructures learned from public spectral libraries. These associations are used to recommend substructures for any unknown mass spectrum. Our method does not require any predefined metabolite candidates, and therefore it can be used for the hypothesis generation or partial identification of unknown unknowns. The method is called MESSAR (MEtabolite Sub-Structure Auto-Recommender) and is implemented in a free online web service available at messar.biodatamining.be. OPEN ACCESS Citation: Liu Y, Mrzic A, Meysman P, De Vijlder T, Romijn EP, Valkenborg D, et al. (2020) MESSAR: Automated recommendation of metabolite substructures from tandem mass spectra. PLoS ONE 15(1): e0226770. https://doi.

Supervised topic modeling for predicting molecular substructure from mass spectrometry

“…With this in mind, we developed, described, and open-sourced a supervised topic modeling method for identifying chemical substructures in tandem mass spectrometry data via LLDA 17 . In a series of empirical studies, this supervised topic model was trained and tested on publicly available benchmark data and substructures, and LLDA was compared to an alternative method, MESSAR 20 . A k-nearest neighbors (k-NN) was also implemented as a means of testing spectral library matching to predict substructure labels based on neighbor averages.…”

Advances in decomposing complex metabolite mixtures using substructure- and network-based computational metabolomics approaches

“…MESSAR was developed to recommend substructures that are likely to be present in unlabeled (unannotated) MS/MS spectra. 23 This is inspired by recommendation services that suggest purchases or services based upon individuals' previous behaviour and choices. To train MESSAR in recognizing the potential relationships between spectral features and substructures, GNPS public spectral libraries 21 were used.…”

“…Here, the more efficient mining for unique substructure motifs will facilitate (i) the prioritization of novel chemistry in complex metabolite mixtures, and (ii) the structural elucidation procedure as such substructure motifs can often be related to biosynthetic or chemical building blocks of the metabolites. In mass spectrometry, tools like MS2LDA 16 and MESSAR 23 have started to recognize substructure motifs based on spectral data, and MotifDB 20 is able to store annotated substructure motifs facilitating their reuse for future structural annotation purposes; however, in many cases, the structural annotation and verification of the substructure chemistry still relies on analytical experts that need to identify the structural motifs by linking them to elemental formulas and structures or chemical compound classes. In NMR, 2D-NMR spectra have started to be exploited to recognize substructure motifs, 125 but here the automated recognition of NMR signals may hamper progress in this area.…”

New software tools, databases, and resources in metabolomics: updates from 2020

“…MEtabolite SubStructure Auto-Recommender (MESSAR), is a web-based tool that provides an automated method for substructure recommendation guided by association rule mining, captures potential relationships between spectral features and substructures as learned from public spectral libraries for suggesting substructures for any unknown mass spectrum (Y. Liu, Mrzic, et al, 2020 ; Liu, Nellis, et al, 2020 ). Though the interface does not perform batch processing currently, it provides an open-source approach to annotate substructures.…”

“…A buffer modification workflow (BMW) in which the same sample is run by LC–MS in both liquid chromatography solvent with 14 NH 3 –acetate buffer and in solvent with the buffer modified with 15 NH 3 –formate, resulted in characteristic mass and signal intensity changes for adduct peaks, facilitating their annotation (Lu et al 2020 ). Towards reference materials standardization, quantitative measures of approximately 200 metabolites for each of three pooled reference materials (220 metabolites for Qstd3, 211 metabolites for CHEAR, 204 metabolites for NIST1950) were obtained and supported harmonization of metabolomics data collected from 3677 human samples in 17 separate studies analyzed by two complementary HRMS methods (K. H. Liu, Mrzic, et al, 2020 ; Liu, Nellis, et al, 2020 ). Another review highlighted the recent progresses (since 2016) in the field of chemical derivatization LC–MS for both targeted and untargeted metabolome analysis (Zhao & Li, 2020 ).…”