Comprehensive chemotaxonomy: mining data from tandem mass spectrometry of lichens

From as early as the 19th century, it was suggested that the « chemical nature » of plants would have value for their taxonomy [1], although the means of verifying this postulate were scarce at the time. The recent development of untargeted metabolomics offers new tools to address this problem.

Computational solutions have been developed to optimally exploit such untargeted tandem mass spectrometry data. Spectral similarity, deduced from the cosine score, and substructure-linked spectral motifs [2] are notably focusing on metabolite organization and annotation.

In this study, we take advantage of these tools to obtain structural and spectral relationships, which will allow us to classify species (considered here as complex ensembles of molecules). We postulate that the biosynthetic relationships between the chemical features can be inferred from the cosine scores and spectral motifs, which in turn help to classify the producing organisms. We use spectral motifs to compute a modified Junker’s BioSynthetically Informed Distance expanding its scope to non-annotated features [3]. Furthermore, we combine it with the weighted Chemical Structural and Compositional Similarity distance [4], [5], introducing a new Structural and Substructural Similarity Informed Distance. Results produced by this method on 32 specimens of the genus Usnea (lichenized Ascomycetes) seem to efficiently reflect the phylogeny established by multilocus analysis. We further applied this metric to classify 169 specimens of Usnea lacking genetic information.

These tools should efficiently complement the set of approaches currently used to classify living organisms, which mainly so far rely on morphological and genetic characters.

• References

• 1 de Candolle Augustin Pyramus. Essai sur les propriétés médicales des plantes, comparées avec leurs formes extérieures et leur classification naturelle. Paris: Chez Crochard, Libraire; Available from: https://www.biodiversitylibrary.org/item/196369
• 2 Hooft JJJ van der, Wandy J, Barrett MP, Burgess KEV, Rogers S. Topic modeling for untargeted substructure exploration in metabolomics. PNAS 2016; 113: 13738-13743
• 3 Junker RR. A biosynthetically informed distance measure to compare secondary metabolite profiles. Chemoecology 2018; 28: 29-37
• 4 Sedio BE, Echeverri JCR, P CAB. Wright SJ. Sources of variation in foliar secondary chemistry in a tropical forest tree community. Ecology 2017; 98: 616-623
• 5 Brejnrod AD, Ernst M, Dworzynski P, Rasmussen LB, Dorrestein P, Hooft J van der, Arumugam M. Implementations of the chemical structural and compositional similarity metric in R and Python. bioRxiv 2019; 546150