Untargeted In Silico Compound Classification—A Novel Metabolomics Method to Assess the Chemodiversity in Bryophytes

In plant ecology, biochemical analyses of bryophytes and vascular plants are often conducted on dried herbarium specimen as species typically grow in distant and inaccessible locations. Here, we present an automated in silico compound classification framework to annotate metabolites using an untargeted data independent acquisition (DIA)–LC/MS–QToF-sequential windowed acquisition of all theoretical fragment ion mass spectra (SWATH) ecometabolomics analytical method. We perform a comparative investigation of the chemical diversity at the global level and the composition of metabolite families in ten different species of bryophytes using fresh samples collected on-site and dried specimen stored in a herbarium for half a year. Shannon and Pielou’s diversity indices, hierarchical clustering analysis (HCA), sparse partial least squares discriminant analysis (sPLS-DA), distance-based redundancy analysis (dbRDA), ANOVA with post-hoc Tukey honestly significant difference (HSD) test, and the Fisher’s exact test were used to determine differences in the richness and composition of metabolite families, with regard to herbarium conditions, ecological characteristics, and species. We functionally annotated metabolite families to biochemical processes related to the structural integrity of membranes and cell walls (proto-lignin, glycerophospholipids, carbohydrates), chemical defense (polyphenols, steroids), reactive oxygen species (ROS) protection (alkaloids, amino acids, flavonoids), nutrition (nitrogen- and phosphate-containing glycerophospholipids), and photosynthesis. Changes in the composition of metabolite families also explained variance related to ecological functioning like physiological adaptations of bryophytes to dry environments (proteins, peptides, flavonoids, terpenes), light availability (flavonoids, terpenes, carbohydrates), temperature (flavonoids), and biotic interactions (steroids, terpenes). The results from this study allow to construct chemical traits that can be attributed to biogeochemistry, habitat conditions, environmental changes and biotic interactions. Our classification framework accelerates the complex annotation process in metabolomics and can be used to simplify biochemical patterns. We show that compound classification is a powerful tool that allows to explore relationships in both molecular biology by “zooming in” and in ecology by “zooming out”. The insights revealed by our framework allow to construct new research hypotheses and to enable detailed follow-up studies.

Partition Dimension of Complete Multipartite Graph

Determining a resolving partition of a graph is an interesting study in graph theory due to many applications like censor design, compound classification in chemistry, robotic navigation and internet network. Let and , the distance between an is . For an ordered partition of , the representation of with respect to is . The partition is called a resolving partition of if all representation of vertices are distinct. The partition dimension of graph is the smallest integer such that has a resolving partition with element.In this thesis, we determine the partition dimension of complete multipartite graph , which is limited by , with and . We found that , , and , .

APPLICATION OF THE MACHINE AND DEEP LEARNING METHODS FOR THE CLASSIFICATION OF CANNABINOID- AND CATHINONE-DERIVED COMPOUNDS

Objective: New psychoactive substances (NPS) have been rapidly developed to avoid legal entanglement. In 2013–2018, the number of cathinonederivedcompounds increased from 30 to 89. In 2016, of 56 NPS compounds, 21 were identified as cannabinoid-derived; only 43 were regulated inthe narcotics law. Artificial intelligence, such as machine and deep learning, is a method of data processing and object recognition, including humanposes and image classifications.Methods: Herein, the machine and deep learning methods for cathinone- and cannabinoid-derived compound classification were compared usingpharmacophore modeling as the reference method. For classifying cathinone-derived compounds, the structure was transformed into fingerprints,which was used as a learning parameter for the machine and deep learning methods. Contrarily, the physicochemical properties and fingerprint shapewere utilized as learning materials for the deep learning method to classify the cannabinoid-derived substances.Results: Consequently, in the cathinone-derived compound classification, the deep learning method produced the accuracy and Cohen kappa valuesof 0.9932 and 0.992, respectively. Furthermore, such values in the pharmacophore modeling method were higher than those in the machine learningmethod (0.911 and 0.708 vs. 0.718 and 0.673, respectively). In the cannabinoid-derived compound classification, the deep learning method with thefingerprint form had the highest accuracy and Cohen kappa values (0.9904 and 0.9876). Such values in this method with the descriptor form werehigher than those in the pharmacophore modeling method (0.8958 and 0.8622 vs. 0.68 and 0.396, respectively).Conclusion: The deep learning method has the potential in the NPS classification.

Chemical Diversity and Classification of Secondary Metabolites in Nine Bryophyte Species

The central aim in ecometabolomics and chemical ecology is to pinpoint chemical features that explain molecular functioning. The greatest challenge is the identification of compounds due to the lack of constitutive reference spectra, the large number of completely unknown compounds, and bioinformatic methods to analyze the big data. In this study we present an interdisciplinary methodological framework that extends ultra-performance liquid chromatography coupled to electrospray ionization quadrupole time-of-flight mass spectrometry (UPLC/ESI-QTOF-MS) with data-dependent acquisition (DDA-MS) and the automated in silico classification of fragment peaks into compound classes. We synthesize findings from a prior study that explored the influence of seasonal variations on the chemodiversity of secondary metabolites in nine bryophyte species. Here we reuse and extend the representative dataset with DDA-MS data. Hierarchical clustering, heatmaps, dbRDA, and ANOVA with post-hoc Tukey HSD were used to determine relationships of the study factors species, seasons, and ecological characteristics. The tested bryophytes showed species-specific metabolic responses to seasonal variations (50% vs. 5% of explained variation). Marchantia polymorpha, Plagiomnium undulatum, and Polytrichum strictum were biochemically most diverse and unique. Flavonoids and sesquiterpenoids were upregulated in all bryophytes in the growing seasons. We identified ecological functioning of compound classes indicating light protection (flavonoids), biotic and pathogen interactions (sesquiterpenoids, flavonoids), low temperature and desiccation tolerance (glycosides, sesquiterpenoids, anthocyanins, lactones), and moss growth supporting anatomic structures (few methoxyphenols and cinnamic acids as part of proto-lignin constituents). The reusable bioinformatic framework of this study can differentiate species based on automated compound classification. Our study allows detailed insights into the ecological roles of biochemical constituents of bryophytes with regard to seasonal variations. We demonstrate that compound classification can be improved with adding constitutive reference spectra to existing spectral libraries. We also show that generalization on compound classes improves our understanding of molecular ecological functioning and can be used to generate new research hypotheses.

A novel graph kernel on chemical compound classification

Considering the classification of compounds as a nonlinear problem, the use of kernel methods is a good choice. Graph kernels provide a nice framework combining machine learning methods with graph theory, whereas the essence of graph kernels is to compare the substructures of two graphs, how to extract the substructures is a question. In this paper, we propose a novel graph kernel based on matrix named the local block kernel, which can compare the similarity of partial substructures that contain any number of vertexes. The paper finally tests the efficacy of this novel graph kernel in comparison with a number of published mainstream methods and results with two datasets: NCI1 and NCI109 for the convenience of comparison.