Machine learning methods to analyse and integrate spatially resolved transcriptomic profiles of prostate cancer patients across ethnic groups

Applications for this project are now open. Please complete your application on The University of Manchester website.

Project Outline

The risk and prognosis of prostate cancer (PCa) varies between different ethnic groups across the globe. However, the reasons for this disparity have so far remained challenging to determine. With the development of novel spatial transcriptomics (ST) technologies, it may be possible to explain these disparities. However, there is a need for development of new computational tools and methods that can be applied to multiple ST datasets from different clinical/biological conditions and from patients of different ethnicities.

Within this project, we will develop a rigorous method that is capable of performing data integration across multiple ST datasets by estimating the shared embeddings of biological effects, aligning the shared embeddings that capture cellular biological variation and clustering the aligned embeddings to obtain domain clusters across different ethnic datasets.

This will provide metrics that will aid in providing insights into the underlying differences in progression of the tumours and tumour aggression between ethnicities and improve our understanding of what type of patient profiles eventually transform PCa into a lethal disease.

This project will obtain ST data for training and validation of the models and employ probabilistic embeddings and clustering of gene expression at the single cell level.

Objectives:

• Generate spatial transcriptomic atlases across patients from different ethnic groups during tumour progression
• Associate differences in spatial biology of prostate cancer to ethnicity
• Develop machine learning methodology to analyse and integrate spatial transcriptome profiles of prostate cancer across ethnic groups
• Validate the method through comprehensive simulations and through testing with real world datasets
• Identify evolutionary subtypes (‘evotypes’) of prostate cancer in Kenya, to enable the prediction of future aggressiveness of individual prostate cancers at an early stage

ACED Non-clinical PhD research project in cancer early detection at The University of Manchester

The International Alliance for Cancer Early Detection (ACED) is a partnership between Cancer Research UK, UCL, Canary Center at Stanford University, University of Cambridge, Knight Cancer Institute at OHSU, and The University of Manchester.

Earlier detection of cancer offers the greatest opportunity to deliver improvements in successful outcomes for patients. ACED aims to accelerate cancer early detection research by bringing together the best researchers across the UK and US.

We are looking for talented and motivated clinical doctors who have a strong academic background and are passionate about research in cancer early detection.

Funding Information

ACED Non-Clinical Studentships are funded for four years and cover:

• project running costs
• university tuition fees (at the UK rate – please see below)
• a stipend of £21,000 per year

The University of Manchester offers a limited number of bursaries enabling full studentships for international applicants. These studentships are awarded to exceptional quality candidates, due to the competitive nature of this scheme.

International applicants (including EU nationals) must meet the required academic eligibility criteria (including English Language). International candidates also need to be aware of additional costs, including visa costs and the UK health surcharge.

Applications Timeline

Before submitting an application, we strongly advise you to contact the project supervisor to discuss the project and any particulars.

Please also make sure you have reviewed and meet the eligibility criteria for the project. Further information is available on the Non-Clinical PhD Studentships webpage. Information relating to CRUK ACED Manchester is available on their website.

Applications close on Friday 2nd June 2023.