Job details

Job title

PhD Studentship (Crispin Miller)

Job reference

REQ00312

Date posted

10/11/2023

Application closing date

15/12/2023

Location

Glasgow

Salary

£21,000 stipend.

Package

All tuition fees will be paid.

Contractual hours

Blank

Basis

Full time

Job category/type

PhD Students

Attachments

Blank

Job description

Deep Tissue Phenotyping and Explainable AI

As life sciences becomes increasingly data-centric, AI techniques including deep learning neural networks and classical machine learning algorithms are increasingly being used to characterise, classify, and make predictions from biological datasets. A challenge with many of these algorithms is that they can be both complex and difficult to interpret.

Explainable AI aims to improve the interpretability of machine learning approaches, with the potential both to increase confidence in the underlying basis of a prediction, and to uncover new patterns and relationships in the underlying data – therefore advancing our understanding of basic biology.

At the CRUK Scotland Institute we make extensive use of different technology platforms including spatial transcriptomics (Xenium, CosMX), single cell RNA-seq, MS-based proteomics and metabolomics, and whole exome and genome sequencing to capture a multimodal view of the tumour environment. We then use computational approaches to integrate and interrogate these datasets and generate a more holistic understanding of oncogenic transformation, tumour growth and maintenance. These analyses are performed on our dedicated HPC cluster.

The goal of this project is to develop Explainable AI techniques to analyse multimodal data arising from over 900 colorectal and pancreatic ductal adenoncarcinoma (PDAC) patients. These include, but are not limited to, extensive profiling using spatial transcriptomics, H&E images, high-plex immunofluorescence, genomics and transcriptomics data, and detailed clinical metadata. We will combine these datatypes with H&E images processed using AI computational pathology techniques to develop a better understanding of the molecular underpinnings of different tumour subtypes, to identify possible therapeutic targets, and to further advance our understanding of the changing landscape of tumour host interactions and tumour evolution.

The studentship will be embedded within a highly collaborative team comprising computer scientists, mathematicians, software engineers, bench-scientists and clinicians.

For informal enquires please contact Prof Crispin Miller ([email protected])