In collaboration with the CRUK Scotland Institute, DiRAC is pleased to invite applications for three 6-month internships to join a highly interdisciplinary and collaborative team focused on using high performance computing to analyse cancer genomics and imaging data. The goal is to improve our understanding of the processes that occur as tumours develop.
The proposed start date for the placement is May/June 2025.
Applications are now closed.
The CRUK Scotland Institute is a world leading cancer research institute focused on developing a better understanding of how cancers develop, grow and spread, as well as how the immune system can be harnessed to combat the disease. Research within the Institute makes extensive use of cutting-edge scientific technologies include Xenium and CosMX spatial transcriptomics, imaging mass spectrometry, single cell sequencing, and advanced imaging and microscopy to develop a more holistic view of a tumour. Fundamental to this research is the application of computational and mathematical techniques to these data.
Computation is supported by a local GPU-accelerated High Performance Computing (HPC) cluster, and through collaborations with DiRAC to access these technologies at scale.
As a student or Early Career Stage Researcher (ECR) you will gain experience in:
You will be embedded within Professor Miller’s group, which includes mathematics, computer science and research software engineering, alongside computational biology and bioinformatics. The team interacts closely with bench and clinical-scientists as part of a highly collaborative research programme across the institute and beyond. These internships will provide an excellent opportunity not only to work with cutting edge computing and mathematical techniques, but also to develop an understanding of fundamental questions in biology.
Approaches developed during these internships are expected to contribute towards scientific publications, and to make a significant impact on how we think about, visualise, and interpret molecular data.
We are using our data to understand how different sets of genes behave in individual cells within a tissue, and how these patterns of activity vary according to the local microenvironment. We are generating integrated datasets that combine enumeration of different cancer-causing mutations, with measurements of RNA abundance, and microscopy images identifying where proteins are found within sets cells. A major challenge is the joint analysis of these different modalities.
We have three internships focusing on different aspects of gene expression:
We have recently been able to exploit the substantial GPU capabilities of DiRAC to generate Large Language Models (LLMs) or RNA sequences. This is part of a wider collaboration involving research groups led by Crispin Miller, Ke Yuan (at the CRUK Scotland Institute) and David Robertson (University of Glasgow Centre for Viral Research), and DiRAC. We are already using these to develop new insights into cancer biology. We are offering an internship to explore how information captured within our DL networks can be used to better understand how regulatory information is encoded within DNA sequences. The project will involve working with DL models and libraries such as PyTorch to generate and interrogate sequence-derived embeddings. The internship will interact closely with biologists to ask how patterns identified in these analyses map onto real world data derived from tumours.
Since gene expression data can describe the combined expression profiles of many thousands of genes, it is typical to use dimensionality reduction, before projecting the data into 2D or 3D. As datasets become increasingly complex, dynamic visualisation tools and animations are increasingly important to help illustrate how, for example, different patterns of gene expression occur in different cell types. The project will focus on the development of visualisation approaches using tools such as Blender to support the presentation and interpretation of our cancer data and to ask how abstract relationships between sequence and gene expression profile map to physical relationships between cells in tissue.
Our increasing ability to probe the same tissue using multiple technologies has the potential to generate a more complete view of a tumour. We are interested how this can help us understand why different patients respond better to different therapies, and why different tumours progress in different ways. This requires multiple high dimensional datasets to be integrated and analysed for recurring patterns, similarities and differences. The project will focus on developing approaches to perform these analyses at scale across imaging and spatial transcriptomics datasets. This may involve developing GPU-accelerated implementations to support data integration.
We welcome applications from candidates who have:
All applicants will receive consideration without regard to race, national origin, gender, age, religion, disability, or any other category protected by law.
Placements are open to PhD students and early career researchers, and are fully funded but you must get your supervisor or PI’s permission before applying – under UKRI rules participation in the scheme is only allowed with their consent.
Applications are now closed.
The successful candidate will remain based at their home university. We do our best to offer flexibility; part-time working can be arranged as long as the placement does not exceed 1 year.
If you have any questions, please email them to DiRAC_placements@leicester.ac.uk
For more information about how an Innovation Placement can benefit your career, as well as testimonials from previous placements students, see our Innovation Placements landing page.
Banner image: A cell of fire and ice, by Nikki Paul. This image shows in vitro migration of a melanoma skin cancer cell over time. The actin cytoskeleton at different time points is shown in different colours.
