Design novel computational methods to understand ageing (Professor Kazem Rahimi)
PROJECT TITLE
Design novel computational methods to understand ageing
SUPERVISORS
DESCRIPTION OF PROJECT
Ageing is a key determinant of chronic disease burden and mortality, yet our ability to measure and monitor ageing in clinical settings remains limited. Current approaches to ageing biomarkers, such as telomere length, epigenetic clocks, or frailty indices, have provided valuable insights but face challenges in terms of scalability, cost, invasiveness, and clinical applicability. With the increasing availability of large-scale, longitudinal health data and advances in artificial intelligence (AI), there is now a unique opportunity to develop next-generation, data-driven biomarkers of ageing that are scalable, dynamic, and clinically actionable.
This DPhil project, based in the Deep Medicine research group led by Professor Kazem Rahimi, aims to design, develop, and validate novel ageing biomarkers using machine learning applied to routinely collected health data, such as electronic health records (EHRs). By leveraging recent advances in deep learning architectures, temporal modelling, and multimodal data integration, the project will explore how real-world health trajectories can be transformed into interpretable and predictive markers of ageing.
The student will begin by reviewing existing approaches to ageing measurement and identifying their limitations. Key research questions may include how to model cumulative physiological decline using longitudinal data, how to distinguish between intrinsic and modifiable components of ageing, and how to ensure biomarkers are robust across populations and health systems.
Potential methodological areas include:
- Development of computational models to capture temporal ageing trajectories
- Identification of modifiable vs non-modifiable contributors to ageing acceleration
- Methods for validating generalisability across health systems and demographic groups
- Approaches for integrating biomarkers into risk prediction tools or clinical workflows
The project will offer opportunities to work with large and richly phenotype datasets from UK and international sources, and to collaborate with experts in machine learning, epidemiology, and geroscience. It will combine technical innovation with clinical relevance, ensuring that proposed solutions are both scientifically rigorous and practically useful.
Candidates should have a strong background in statistics, computer science, or related quantitative disciplines. Prior experience with healthcare data or ageing research is welcomed but not essential. Proficiency in programming and enthusiasm for translating advanced methods into real-world clinical applications are crucial.
TRAINING OPPORTUNITIES
The DPhil student will receive comprehensive training in developing data-driven ageing measurements using advanced machine learning and epidemiological methods. This will include:
- Deep learning techniques for dealing with longitudinal health data
- Temporal data analysis and multimodal data integration from EHRs
- Explainable AI approaches for understanding model drivers and outputs
- Statistical methods for biomarker validation, risk prediction, and outcome modelling
- Ethical and privacy considerations in real-world health data research
The student will work within the interdisciplinary Deep Medicine group, gaining exposure to experts in AI, geroscience, clinical epidemiology, and public health.
They will also receive training in scientific communication, including academic writing, presenting at conferences, and publishing in peer-reviewed journals. The programme includes opportunities to conduct systematic literature reviews, design robust research protocols, and collaborate on joint projects with clinicians, statisticians, and international research teams.
Funding Information
Unfunded, however, the Deep Medicine group will assist the DPhil candidate in securing funding.
HOW TO APPLY
To apply for this research degree, please click here.