Ai-driven lipid network analysis platform for soybean root rot disease resistance
An integrated machine learning and lipid network mapping platform that identifies novel lipid biomarkers mediating soybean tolerance to Phytophthora root rot. Combines mass spectrometry-based lipidomics with AI-driven predictive modeling and biochemical network analysis to uncover causal mechanisms of disease resistance, enabling development of plant protection solutions for a disease causing approximately $2B in annual global crop losses.
Overview
This research program delivers an integrated computational and analytical platform combining machine learning with lipid network mapping to identify novel biomarkers that mediate soybean tolerance to Phytophthora root and stem rot, a destructive oomycete disease responsible for approximately $2B in annual global crop losses. By bridging predictive AI descriptors with causal biochemical context, the platform enables robust identification of lipid-based biomarkers suitable for future plant protection solutions. The work addresses a critical gap in plant pathology: while ML approaches can generate predictive descriptors of plant-pathogen interactions, methods for translating predictions into causal biochemical mechanisms have been lacking.
Technical specifications
Key features:
- Mass spectrometry-based lipidomics for tissue-specific measurement of structural, neutral, and oxygenated lipid responses to pathogen inoculation in root and stem tissues
- ML-based predictive modeling integrated with network mapping to infer biochemical modules driving soybean defense and adaptation
- Lipid network analysis using structural similarity and regularized causal networks to overcome challenges posed by ambiguous lipid-lipid interaction databases and enzyme substrate promiscuity
- Comparative cultivar analysis between wildtype (Conrad) and resistant (OX760-6) soybean genotypes to characterize time-dependent lipid responses
- Visualization of biomarkers embedded in lipid biochemical relationships, including stigmasterol, phospholipids, and glycerolipids implicated in successful host-pathogen interactions
- Anatomical and chemical characterization of soybean root suberin and mechanisms of partial resistance to P. sojae
Technology readiness level
The platform builds on established experimental protocols, mass spectrometry measurement methods, and lipid modeling and network mapping approaches that have been successfully applied to interrogate a variety of biochemical systems. The team has already developed capabilities to determine anatomical distribution and chemical composition of soybean root suberin and identified alterations in the lipidome and oxylipin cascades through tissue-specific analyses. Preliminary partial analysis of the data has identified some unique lipid biomarkers. The next phase will extend this work by incorporating state-of-the-art ML and network mapping techniques to validate lipid-signaling mechanisms conferring resistance to P. sojae and develop a soybean lipid metabolic map for biomarker discovery.
About Western University
Western University is a comprehensive public research university in London, Ontario, with a full spectrum of professional and research programs and a strong culture of partnership with industry. Companies engage through co-located labs and shared core facilities, a research and technology park with industry tenancy, and advanced prototyping testbeds that enable scale-up. Integration with major hospital systems and affiliated research institutes supports clinical translation and trials, while co-op and internship pathways provide reliable talent pipelines. Research is backed by competitive funding from NSERC, CIHR, SSHRC, and the Canada Foundation for Innovation, and a dedicated technology transfer office offers IP strategy, licensing, and startup support.
delivered to your inbox.