A Phytochemical Network Framework for Plant Defense Theory

Thesis Abstract

Plant defense in terrestrial ecosystems is governed by complex interactions among phytochemicals that form dynamic networks influencing herbivore resistance, pathogen deterrence, and abiotic stress tolerance. This study addresses the fragmentation in understanding how biochemical networks translate to adaptive defense phenotypes, proposing a Phytochemical Network Framework (PNF) that integrates metabolomic, genomic, and ecological data to predict plant defensive outcomes. The aims are to (i) elucidate the architecture of phytochemical networks across representative taxa, (ii) identify key network motifs associated with robust defense under biotic and abiotic stress, and (iii) develop a predictive model linking network properties to defense efficacy. Specific objectives include cataloguing secondary metabolite profiles from 30 plant species spanning three families, measuring induced defensive responses following controlled herbivory and pathogen challenge, and validating a network-based predictive model using cross-validation and independent test sets. A mixed-methods design is employed. The population comprises perennial model and crop species grown under controlled greenhouse conditions and field trials. A stratified sampling scheme yields 10 individuals per species for metabolomic analyses, with additional replicate plants subjected to standardized stress treatments mechanical wounding with jasmonic acid elicitation, and inoculation with a common pathogen consortium. Data collection instruments include high-resolution mass spectrometry (LC-MS/MS) for untargeted and targeted metabolomics, transcriptomic sequencing (RNA-Seq) to capture gene expression linked to defense pathways, and ecological assays quantifying herbivory resistance, pathogen incidence, and fitness metrics. Analytical techniques combine network science with multivariate statistics construction of weighted phytochemical interaction networks using correlation and mutual information measures, community detection to identify modules, centrality analyses to pinpoint hub metabolites, followed by regression analyses (random forest, elastic net) to relate network features to phenotypic defense outcomes. The theoretical anchors include the Optimal Defense Theory and the Network Pharmacology framework, extended to a Phytochemical Network Framework (PNF) that operationalizes module-phenotype mappings and path-analytic pathways from gene regulation to metabolite fluxes and ecological performance. Expected findings indicate that defense-relevant modules are conserved across taxa but exhibit lineage- and environment-specific reconfigurations under stress. Hub metabolites within glucosinolate, phenolic, and alkaloid modules are anticipated to show strong predictive power for resistance indices, with network entropy and modular coherence positively correlating with defense efficacy under multifactorial stress. The study is likely to reveal cross-talk between primary metabolism and specialized metabolite networks, with jasmonate- and salicylic-acid–associated modules driving distinct but interlinked defense responses. A validated predictive model is expected to achieve robust out-of-sample accuracy (R2 > 0.65) in forecasting defense outcomes from metabolomic and transcriptomic network features. The study contributes to knowledge by formalizing a transferable, network-based theory of plant defense that connects molecular networks to ecological performance, enabling cross-species generalization and practical applications in breeding and crop protection. It advances methodological frontiers by integrating untargeted metabolomics, transcriptomics, and network analysis into a coherent framework capable of generating testable hypotheses about defense strategy evolution and plasticity. The anticipated implications include identifying metabolite hubs as candidates for bioengineering and selecting varieties with network configurations conferring durable resistance under climate-change–driven stress regimes. The conclusion emphasizes that phytochemical network topology, rather than isolated metabolites, governs defense robustness, and recommendations highlight the integration of PNFs into breeding pipelines, standardization of network metrics for cross-study comparability, and future work to incorporate microbiome interactions as modulators of phytochemical networks.