Doctoral Fellow: Cracking the genetic code of weight management with AI in Cambridge

Weight management - 3 year fixed term contract

Location: The Discovery Centre, Cambridge Biomedical Campus, Cambridge, UK

Salary: £40,000 gross (subject to deductions in line with UK policy) plus benefit fund and bonus.

AstraZeneca UK has received funding from the Marie Skłodowska‑Curie Actions programme through the EU and is now pleased to offer this position.

Project 101226456 — MLCARE — HORIZON-MSCA-2024-DN-01

Weight management is a complex challenge—but what if AI could help us understand who is at risk of weight‑related issues, why, and what interventions would be most effective? This PhD project will harness the power of deep learning and multi‑omics to uncover the hidden genetic and biological drivers of weight regulation and associated conditions. By enhancing genome‑wide association studies (GWAS) with cutting‑edge machine learning, the project aims to identify genetic variants and effector transcripts that influence body weight, metabolism, and individual responses to treatment. The models will integrate genomic insights with behavioural and clinical data to develop personalized, precision‑driven strategies—redefining how weight is monitored, managed, and improved over time.

PhD award entity: Universidad Carlos III Madrid. Signal Processing and Comm. Engineering department.

Secondments:

• AstraZeneca España – AZ (Centre for Artificial Intelligence) (ES): developing big‑data methods for enhanced GWAS with omics (Potentially June - Aug. 2027).
• University of Copenhagen - UCPH (Section for Computational and RNA Biology) (DK): incorporate omic FMs into enhanced (Potentially June - Aug. 2028).
• Institut Pasteur – IP (Computational Biology, Statistical Genetics group) (FR): multi‑trait obesity GWAS (March - May 2029).

Supervisors:

• Dr Tom Diethe (AstraZeneca UK)
• Dr Dimitrios Athanasakis (AstraZeneca España)
• Dr. Pablo M. Olmos (Universidad Carlos III de Madrid - UC3M)
• Dr. Ole Winther (University of Copenhagen)
• Dr. Hanna Julienne (Institut Pasteur)

Project Objectives and Tasks:

• Build biologically inspired, hierarchical discrete deep generative models to integrate multi‑omics with behavioural and clinical data for weight regulation.
• Enhance GWAS with deep learning to identify causal variants, effector transcripts, and pathways affecting body weight, metabolic rate, adiposity, and treatment response.
• Incorporate pathway‑based priors, regulatory networks, and tissue‑specific annotations into modelling for interpretability and robustness.
• Develop uncertainty‑aware inference, quantization, and error‑correcting strategies to manage missingness, heterogeneity, and batch effects across data sources.
• Construct multi‑domain foundation models for behavioural data (sleep, mobility, smartphone usage) and EHR, with multi‑modal tokenization and autoregressive/multiresolution backbones.
• Detect behavioural and biological change‑points that signal risk of weight‑related deterioration, relapse after weight‑loss interventions, or metabolic decompensation.
• Validate models in clinical settings and independent cohorts; derive personalized risk scores and adaptive intervention policies for weight management.
• Collaborate within a multidisciplinary network of machine learning researchers, bioinformaticians, endocrinologists, psychiatrists, and industry partners.

Expected Results:

• Methods for learning hierarchical discrete deep generative models that fuse GWAS/TWAS with multi‑omics and behavioural data to produce interpretable embeddings and causal signals.
• Identification of genetic variants, effector transcripts, and pathways linked to body weight regulation and differential treatment outcomes.
• A behavioural foundation model and change‑point detection framework for early warning of weight‑related relapse or metabolic complications.
• Personalized strategies for precision weight management, including risk stratification and intervention timing.

Essential criteria:

• Study records, including Bachelor in the areas of Computer Science, Maths, Physics, or a related quantitative field.
• Master’s degree in the area of AI or Machine Learning within Biology as the preferred area, but not essential.
• Minimum total of 300 ECTS credits at the time of application.
• Previous work & research experience.
• Positive attitude, good communication skills.
• English proficiency.

Candidates must:

• Be - at the date of recruitment - a doctoral candidate (i.e., not already in possession of a doctoral degree).
• Be - at the date of recruitment - formally admitted to a PhD programme leading to the award of a degree in at least one EU Member State or Horizon Europe associated country.
• Not have resided or carried out their main activity (work, studies, etc.) in the UK for more than 12 months in the 36 months immediately before the recruitment date.
• Be working exclusively for the action.

Our mission is to build an inclusive and equitable environment. We welcome and consider applications from all qualified candidates, regardless of characteristics.