Translational pathway to implementation of polygenic risk scores for prostate cancer in Scotland

Precision Medicine Project - Translational pathway to implementation of polygenic risk scores for prostate cancer in Scotland

Supervisor(s): Prof Jim Wilson, Prof Riccardo Marioni & Dr Xia Shen (Karolinska Institutet)
Centre/Institute: Centre for Global Health Research, Usher Institute

Background

The new paradigm of preventative medicine seeks to identify individuals who are at higher risk of specific outcomes before the onset of clinical disease. This “precision medicine” can take many forms, but one of the most promising avenues is using genetic information to predict disease risk. In the two decades since the deciphering of the human genome, more and more of our DNA variation is finally being understood and most diseases have been shown to be due to both genetics and the environment. In these cases, disease risk arises from many different risk-increasing or -decreasing variants, together with lifestyle factors and sometimes also rare pathogenic (disease-causing) variants. This complexity means that polygenic scores need to be deployed with realistic expectations (Sud 2023), however the latest approaches have improved accuracy (Zeng & Visscher 2025).

Such advances will have the greatest clinical impact for diseases for which there are no good biomarkers of risk. One of the most common diseases in this category is prostate cancer, which kills twelve thousand men in the UK each year. Prostate-specific antigen (PSA) has clinical utility for monitoring progression, but suffers from a high rate of false positive results when used for screening. Survival varies dramatically according to how early prostate cancer is detected, reaching almost 100% among men diagnosed at stage I or II, but decreasing to a 5-year survival of 50% for those diagnosed at stage IV (Cancer Research UK 2024). Effective screening tools to detect early-stage, clinically significant prostate cancer are therefore urgently required. 

Earlier this year, McHugh et al. (2025) demonstrated for the first time that the method of summing the risks across over 100 different risk variants – a polygenic score – was very predictive of disease. In testing of 6393 men aged 55-69 years, 745 (11.7%) had a polygenic score in the 90th percentile or higher. Of these, 468 went forward for MRI and prostate biopsy, and incredibly prostate cancer was detected in 187 participants (40%), more than half of whom were classified as intermediate or higher risk, thus requiring treatment. Cancer would not have been detected in 72% of these individuals according to the prostate cancer diagnostic pathway currently used in the NHS.

Aims

We propose to calculate polygenic risk scores (PGS) for prostate cancer using the same approach as McHugh et al., but also including more recent data from the Polygenic Score Catalog (https://www.pgscatalog.org/). The scores will be validated in Generation Scotland, where genome-wide genetic data and electronic health record linkage are available for 20,000 individuals. Summaries of incident case numbers over the first ten years since baseline show 148 cases of prostate cancer. We will also investigate the utility of polygenic scores for three of the other most common cancers, namely colorectal, breast and lung, to allow comparison of the genetic architecture and positive predictive values across the different diseases. PGS will then be calculated for the 10,000 participants in Viking Genes (40% male), who also have genome-wide genetic and EHR data available.

After assessment of the portability and efficacy of the PGS in Scottish populations, we will investigate how best to integrate information from rare highly penetrant pathogenic variants, such as BRCA2 variants for prostate and breast cancer and mismatch repair gene variants for colorectal cancer, to enhance prediction (e.g. Bolze 2023). Males over 55 years of age in Viking Genes will be invited to receive their PGS and results will be returned to all consenting participants. Over 100 volunteers in Viking Genes have already received information about their actionable genetic findings, such as Lynch syndrome, hereditary breast and ovarian cancer and Long QT syndrome variants. Identified individuals will be able to access a different pathway of care, involving PSA tests and MRI. Uptake and outcomes will be quantified, including numbers undergoing MRI, biopsy, diagnosis and treatments.

Training outcomes

-Quantitative skills, e.g., programing in bash and python/R, using specialist statistical genetics tools, such as bcftools, PLINK, building pipelines, using high-performance compute clusters

- Experience in genomic data analyses, including sequencing quality control and genome imputation, exome/genome-wide association analyses, polygenic risk scores and working with electronic health records

- Familiarity with various publicly available resources, e.g. ClinVar, gnomAD, OMIM, PGS catalog

- Knowledge of research governance, including data protection, ethics, privacy and experience in translational approaches to increase impact

References 

  1. Bolze A, et al. (2023) Combining rare and common genetic variants improve population risk stratification for breast cancer. MedRxiv: https://doi.org/10.1101/2023.05.17.23290132.
  2. Cancer Research UK (2024) Prostate Cancer Statistics. https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/prostate-cancer.
  3. McHugh JK, et al. (2025) Assessment of a polygenic risk score in screening for prostate cancer. New Engl J Med 392, 1406-17.
  4. Sud A, et al. (2023) Realistic expectations are key to realising the benefits of polygenic scores. BMJ 380, e073149.
  5. Zeng J & Visscher PM (2025) Harnessing functional annotation to improve the accuracy and transferability of polygenic scores. Nature Rev Genet. https://doi.org/10.1038/s41576-025-00893-4.

Apply Now

Click here to Apply Now

  • The deadline for 26/27 applications is Monday 12th January 2026
  • Applicants must apply to a specific project. Please ensure you include details of the project on the Recruitment Form below, which you must submit to the research proposal section of your EUCLID application.
  • Please ensure you upload as many of the requested documents as possible, including a CV, at the time of submitting your EUCLID application.  

Q&A Sessions

Supervisor(s) of each project will be holding a 30 minute Q&A session in the first two weeks of December. 

If you have any questions regarding this project, you are invited to attend the session on TBC via Microsoft Teams. Click here to join the session.

This article was published on