Digital Daphnia: Unravelling Multi-Omics Mechanisms of Environmental Chemical Stress Using Graph Representation Learning

Secondary Supervisor(s): Professor Luisa Orsini

University of Registration: University of Birmingham

BBSRC Research Themes: Understanding the Rules of Life (Systems Biology)

Project Outline

Chemical pollution poses a significant threat to ecosystems and human health. Traditionally, assessments of toxic effects have relied on animal testing, which is ethically contentious, time-consuming, and expensive. High-throughput omics technologies, such as transcriptomics and metabolomics, enable comprehensive analysis of biological samples, providing multi-omics data that reflect biomolecular responses to environmental chemical stress. This data is information-rich and well-suited for integrative in-silico modelling, offering the potential to accurately simulate an organism’s biological processes while reducing reliance on animal testing. However, the complexity of multi-omics data, characterised by small sample sizes, high dimensionality, and heterogeneity in data acquisition and processing, presents significant challenges for computational modelling and biological interpretation. Cutting-edge artificial intelligence approaches, such as graph representation learning and graph neural networks, can tackle these challenges by effectively integrating multi-omics data into comprehensive, coherent models.

Daphnia is a non-sentient model organism widely used in ecotoxicology research. It serves as an early warning indicator for environmental health due to its sensitivity to pollutants. This project aims to holistically model environmental multi-omics data using graph representation learning to create a digitalised Daphnia, unveiling the biomolecular mechanisms responding to environmental chemical pollution. By integrating experimental data with advanced computational technologies, this digital twin model will simulate biological processes across multiple levels of organisation.

Particularly, the PhD student will focus on three research objectives:

OB 1. Develop Integrative Multi-Omics Models: Construct multi-layer graph models based on existing Daphnia magna multi-omics data. Develop graph neural networks to capture complex interactions within and between different omics levels. Deconvolute bulk omics data to obtain cell type-specific information, enhancing resolution at the cellular level.

OB 2. Verify Multi-omics Models with Experimental Data: Collect new experimental multi-omics data from Daphnia magna exposed to environmental chemicals. Validate the accuracy of the model’s simulations, ensuring computational insights correspond to biological reality. Iteratively refine the models based on experimental findings.

OB 3. Create a Digitalised Daphnia for Environmental Chemical Stress: Develop a digital twin model of Daphnia based on the verified graph models that simulate the organism’s biological systems under chemical stress, unveiling the underlying biomolecular mechanisms. Create an interface that enables stakeholders to predict the effects of chemical toxicity, facilitating practical applications in environmental monitoring and policy-making.

This PhD project synergies with the EU H2020 project PrecisionTox, which is co-led by both supervisors at the University of Birmingham. The student will focus on multi-omics data from Daphnia magna, with the scope potentially broadening to include the other four model organisms utilised in the PrecisionTox project.

The PhD student will work at the interface between computational modelling (primary supervisor) and biological interpretation (secondary supervisor), acquiring a unique multidisciplinary profile with expertise in AI, systems biology, and environmental science. The student will also be embedded in two large pan-European initiatives, the Horizon 2020 consortium PrecisionTox and the Horizon Europe PARC. These initiatives provide an extensive multidisciplinary network of experts in diverse fields.

References

[1]. P. Barbiero, R. Viñas Torné, and P. Lió, Graph Representation Forecasting of Patient’s Medical Conditions: Toward a Digital Twin, Front. Genet., 2021.

[2]. W. Ju et al., A Comprehensive Survey on Deep Graph Representation Learning, Neural Networks, 2024.