Enhancement of plant metabolite fingerprinting by machine learning

Integrated Study on Methane Activation: Exploring Main Group Frustrated Lewis Pairs through Density Functional Theory, Machine Learning, and Machine-Learned Force Fields

Frustrated Lewis Pairs (FLP) are an important advance in metal-free catalysis due to their ability to activate a variety of small molecules. Many studies have focused on a very limited sample of Lewis acids and bases. Herein, we disclose an automated exploration algorithm using density functional methods, artificial neural networks (ANNs), and a molecule builder that incentivizes the exploration of favorable FLP space for the activation of methane via two mechanisms: deprotonation and hydride abstraction. The exploration algorithm creates FLPs with different Lewis acids (LA), Lewis bases (LB), and their substituents (LA/LB), which proved successful in quickly converging in the favorable chemical space, suggesting chemically sound structures, and generating thousands of potential candidates for methane activating FLPs. By modeling thousands of reactions, an FLP database of methane activation was created, allowing one to data mine properties, e.g., adduct bond length, highest occupied molecular orbital-lowest-unoccupied molecular orbital (HOMO-LUMO) gap, global electrophilicity index, favored Lewis acids/bases/substituents, and substituent steric volume. These properties not only successfully narrow the FLP chemical space but also provide meaningful insight into the chemical nature of competent methane activators. The machine learning discovery strategy disclosed here is general enough to be applicable to many chemical optimization tasks. This study also investigates the efficacy of a Machine-Learned Force Field (MLFF) in predicting the formation energies of Frustrated Lewis Pairs (FLPs). Our model, exhibiting a test error of ±10 kcal/mol, highlighted impressive computational efficiency by enabling the calculation of all possible FLP permutations within our chemical space. The MLFF demonstrated proficiency in predicting energies, providing a significant acceleration compared to quantum mechanics methods. However, challenges emerged in accurately capturing forces, necessitating recourse to classical force fields for reliable structure relaxation. The present study sheds light on the MLFF's potential as a tool for rapid energy predictions, emphasizing the need for further refinement to enhance its accuracy, particularly in force predictions, to expand its utility in chemical simulations.

Metabolomics-Guided Elucidation of Plant Abiotic Stress Responses in the 4IR Era: An Overview

Plants are constantly challenged by changing environmental conditions that include abiotic stresses. These are limiting their development and productivity and are subsequently threatening our food security, especially when considering the pressure of the increasing global population. Thus, there is an urgent need for the next generation of crops with high productivity and resilience to climate change. The dawn of a new era characterized by the emergence of fourth industrial revolution (4IR) technologies has redefined the ideological boundaries of research and applications in plant sciences. Recent technological advances and machine learning (ML)-based computational tools and omics data analysis approaches are allowing scientists to derive comprehensive metabolic descriptions and models for the target plant species under specific conditions. Such accurate metabolic descriptions are imperatively essential for devising a roadmap for the next generation of crops that are resilient to environmental deterioration. By synthesizing the recent literature and collating data on metabolomics studies on plant responses to abiotic stresses, in the context of the 4IR era, we point out the opportunities and challenges offered by omics science, analytical intelligence, computational tools and big data analytics. Specifically, we highlight technological advancements in (plant) metabolomics workflows and the use of machine learning and computational tools to decipher the dynamics in the chemical space that define plant responses to abiotic stress conditions.

Natural variation in wild tomato trichomes; selecting metabolites that contribute to insect resistance using a random forest approach

BACKGROUND

Plant-produced specialised metabolites are a powerful part of a plant's first line of defence against herbivorous insects, bacteria and fungi. Wild ancestors of present-day cultivated tomato produce a plethora of acylsugars in their type-I/IV trichomes and volatiles in their type-VI trichomes that have a potential role in plant resistance against insects. However, metabolic profiles are often complex mixtures making identification of the functionally interesting metabolites challenging. Here, we aimed to identify specialised metabolites from a wide range of wild tomato genotypes that could explain resistance to vector insects whitefly (Bemisia tabaci) and Western flower thrips (Frankliniella occidentalis). We evaluated plant resistance, determined trichome density and obtained metabolite profiles of the glandular trichomes by LC-MS (acylsugars) and GC-MS (volatiles). Using a customised Random Forest learning algorithm, we determined the contribution of specific specialised metabolites to the resistance phenotypes observed.

RESULTS

The selected wild tomato accessions showed different levels of resistance to both whiteflies and thrips. Accessions resistant to one insect can be susceptible to another. Glandular trichome density is not necessarily a good predictor for plant resistance although the density of type-I/IV trichomes, related to the production of acylsugars, appears to correlate with whitefly resistance. For type VI-trichomes, however, it seems resistance is determined by the specific content of the glands. There is a strong qualitative and quantitative variation in the metabolite profiles between different accessions, even when they are from the same species. Out of 76 acylsugars found, the random forest algorithm linked two acylsugars (S3:15 and S3:21) to whitefly resistance, but none to thrips resistance. Out of 86 volatiles detected, the sesquiterpene α-humulene was linked to whitefly susceptible accessions instead. The algorithm did not link any specific metabolite to resistance against thrips, but monoterpenes α-phellandrene, α-terpinene and β-phellandrene/D-limonene were significantly associated with susceptible tomato accessions.

CONCLUSIONS

Whiteflies and thrips are distinctly targeted by certain specialised metabolites found in wild tomatoes. The machine learning approach presented helped to identify features with efficacy toward the insect species studied. These acylsugar metabolites can be targets for breeding efforts towards the selection of insect-resistant cultivars.

You Are What You Eat: Application of Metabolomics Approaches to Advance Nutrition Research

A healthy condition is defined by complex human metabolic pathways that only function properly when fully satisfied by nutritional inputs. Poor nutritional intakes are associated with a number of metabolic diseases, such as diabetes, obesity, atherosclerosis, hypertension, and osteoporosis. In recent years, nutrition science has undergone an extraordinary transformation driven by the development of innovative software and analytical platforms. However, the complexity and variety of the chemical components present in different food types, and the diversity of interactions in the biochemical networks and biological systems, makes nutrition research a complicated field. Metabolomics science is an "-omic", joining proteomics, transcriptomics, and genomics in affording a global understanding of biological systems. In this review, we present the main metabolomics approaches, and highlight the applications and the potential for metabolomics approaches in advancing nutritional food research.

Predicting future biomass yield in Miscanthus using the carbohydrate metabolic profile as a biomarker

In perennial energy crop breeding programmes, it can take several years before a mature yield is reached when potential new varieties can be scored. Modern plant breeding technologies have focussed on molecular markers, but for many crop species, this technology is unavailable. Therefore, prematurity predictors of harvestable yield would accelerate the release of new varieties. Metabolic biomarkers are routinely used in medicine, but they have been largely overlooked as predictive tools in plant science. We aimed to identify biomarkers of productivity in the bioenergy crop, Miscanthus, that could be used prognostically to predict future yields. This study identified a metabolic profile reflecting productivity in Miscanthus by correlating the summer carbohydrate composition of multiple genotypes with final yield 6 months later. Consistent and strong, significant correlations were observed between carbohydrate metrics and biomass traits at two separate field sites over 2 years. Machine-learning feature selection was used to optimize carbohydrate metrics for support vector regression models, which were able to predict interyear biomass traits with a correlation (R) of >0.67 between predicted and actual values. To identify a causal basis for the relationships between the glycome profile and biomass, a 13C-labelling experiment compared carbohydrate partitioning between high- and low-yielding genotypes. A lower yielding and slower growing genotype partitioned a greater percentage of the 13C pulse into starch compared to a faster growing genotype where a greater percentage was located in the structural biomass. These results supported a link between plant performance and carbon flow through two rival pathways (starch vs. sucrose), with higher yielding plants exhibiting greater partitioning into structural biomass, via sucrose metabolism, rather than starch. Our results demonstrate that the plant metabolome can be used prognostically to anticipate future yields and this is a method that could be used to accelerate selection in perennial energy crop breeding programmes.

Serum Metabonomics of Mild Acute Pancreatitis

BACKGROUND

Mild acute pancreatitis (MAP) is a common acute abdominal disease, and exhibits rising incidence in recent decades. As an important component of systemic biology, metabonomics is a new discipline developed following genomics and proteomics. In this study, the objective was to analyze the serum metabonomics of patients with MAP, aiming to screen metabolic markers with potential diagnostic values.

METHODS

An analysis platform with ultra performance liquid chromatography-high-resolution mass spectrometry was used to screen the difference metabolites related to MAP diagnosis and disease course monitoring.

RESULTS

A total of 432 endogenous metabolites were screened out from 122 serum samples, and 49 difference metabolites were verified, among which 12 difference metabolites were identified by nonparametric test. After material identification, eight metabolites exhibited reliable results, and their levels in MAP serum were higher than those in healthy serum. Four metabolites exhibited gradual downward trend with treatment process going on, and the differences were statistically significant (P < 0.05).

CONCLUSION

Metabonomic analysis has revealed eight metabolites with potential diagnostic values toward MAP, among which four metabolites can be used to monitor the disease course.

Opposite variations in fumarate and malate dominate metabolic phenotypes of Arabidopsis salicylate mutants with abnormal biomass under chilling

In chilling conditions (5°C), salicylic acid (SA)-deficient mutants (sid2, eds5 and NahG) of Arabidopsis thaliana produced more biomass than wild type (Col-0), whereas the SA overproducer cpr1 was extremely stunted. The hypothesis that these phenotypes were reflected in metabolism was explored using 600 MHz (1) H nuclear magnetic resonance (NMR) analysis of unfractionated polar shoot extracts. Biomass-related metabolic phenotypes were identified as multivariate data models of these NMR 'fingerprints'. These included principal components that correlated with biomass. Also, partial least squares-regression models were found to predict the relative size of plants in previously unseen experiments in different light intensities, or relative size of one genotype from the others. The dominant signal in these models was fumarate, which was high in SA-deficient mutants, intermediate in Col-0 and low in cpr1 at 5°C. Among signals negatively correlated with biomass, malate was prominent. Abundance of transcripts of the FUM2 cytosolic fumarase (At5g50950) showed strong positive correlation with fumarate levels and with biomass, whereas no significant differences were found for the FUM1 mitochondrial fumarase (At2g47510). It was confirmed that the morphological effects of SA under chilling find expression in the metabolome, with a role of fumarate highlighted.

Merits of random forests emerge in evaluation of chemometric classifiers by external validation

Real-world applications will inevitably entail divergence between samples on which chemometric classifiers are trained and the unknowns requiring classification. This has long been recognized, but there is a shortage of empirical studies on which classifiers perform best in 'external validation' (EV), where the unknown samples are subject to sources of variation relative to the population used to train the classifier. Survey of 286 classification studies in analytical chemistry found only 6.6% that stated elements of variance between training and test samples. Instead, most tested classifiers using hold-outs or resampling (usually cross-validation) from the same population used in training. The present study evaluated a wide range of classifiers on NMR and mass spectra of plant and food materials, from four projects with different data properties (e.g., different numbers and prevalence of classes) and classification objectives. Use of cross-validation was found to be optimistic relative to EV on samples of different provenance to the training set (e.g., different genotypes, different growth conditions, different seasons of crop harvest). For classifier evaluations across the diverse tasks, we used ranks-based non-parametric comparisons, and permutation-based significance tests. Although latent variable methods (e.g., PLSDA) were used in 64% of the surveyed papers, they were among the less successful classifiers in EV, and orthogonal signal correction was counterproductive. Instead, the best EV performances were obtained with machine learning schemes that coped with the high dimensionality (914-1898 features). Random forests confirmed their resilience to high dimensionality, as best overall performers on the full data, despite being used in only 4.5% of the surveyed papers. Most other machine learning classifiers were improved by a feature selection filter (ReliefF), but still did not out-perform random forests.

Recent advances in metabolomics in oncology

The burden of cancer is growing worldwide and with it a more desperate need for better tools to detect, diagnose and monitor the disease is required. It is well recognized that cancer cells are characterized by distinct metabolic perturbations. The metabolomics approach involves the comprehensive profiling of the full complement of low MW compounds in a biological system. By applying advanced analytical and statistical tools, the 'metabolome' is mined for biomarkers that are associated with the state of cancer. This review presents an introduction to the main analytical platforms used in metabolomics analyses, such as NMR spectroscopy and MS, as well as the statistical tools used to mine these datasets. The discussion focuses on 'state-of-the-art' investigations on the four cancer types that have received the most study by metabolomics, namely breast, prostate, colorectal and liver cancer.

Metabolomics in agriculture

Metabolome refers to the complete set of metabolites synthesized through a series of multiple enzymatic steps from various biochemical pathways processing the information encrypted in the plant genome. Knowledge about synthesis and regulation of various plant metabolic substances has improved substantially with availability of Omics data originating from sequencing of plant genomes. Metabolic profiling of crops is increasingly becoming popular in assessing plant phenotypes and genetic diversity. Metabolic compositional changes vividly reflect the changes occurring during plant growth, development, and in response to stress. Hence, study of plant metabolic pathways, the interconnections between them in context of systems biology is increasingly becoming popular in identification of candidate genes. The present article reviews recent developments in analysis of plant metabolomics, available bioinformatics techniques and databases employed for comparative pathway analysis, metabolic QTLs, and their application in plants.

Plantmetabolomics.org: mass spectrometry-based Arabidopsis metabolomics--database and tools update

The PlantMetabolomics (PM) database (http://www.plantmetabolomics.org) contains comprehensive targeted and untargeted mass spectrum metabolomics data for Arabidopsis mutants across a variety of metabolomics platforms. The database allows users to generate hypotheses about the changes in metabolism for mutants with genes of unknown function. Version 2.0 of PlantMetabolomics.org currently contains data for 140 mutant lines along with the morphological data. A web-based data analysis wizard allows researchers to select preprocessing and data-mining procedures to discover differences between mutants. This community resource enables researchers to formulate models of the metabolic network of Arabidopsis and enhances the research community's ability to formulate testable hypotheses concerning gene functions. PM features new web-based tools for data-mining analysis, visualization tools and enhanced cross links to other databases. The database is publicly available. PM aims to provide a hypothesis building platform for the researchers interested in any of the mutant lines or metabolites.