*4 Year MRC PhD Programme* Deep learning exploration of large bioresources for the discovery of retinal biomarkers

Big Data analytics bears huge promises for improving healthcare. Dundee has all the elements needed to play a significant role in the delivery of this promise: large collections of cross-linked clinical and research data including genomics, an advanced infrastructure for making such data readily available (including a safe haven) for research, and a solid history of clinical informatics and computing research.

This project will take forward existing collaborative, interdisciplinary projects involving machine learning, deep learning biomedical informatics and statistics. These projects largely focus on extracting latent information from routine clinical imaging such as photographs of the retina (the VAMPIRE initiative) and scans of the brain and combining this with clinical and genomic data to investigate how this information may improve understanding of susceptibility to disease and therefore better clinical management.

The target questions: Can deep learning approaches to Big Data replicate known relevant associations in large data sets? Can deep learning be used for novel concept discovery that is clinically relevant and useful? Can deep learning outpower more conventional analytical approaches to discovery of novel and clinically informative concepts?

The student will become part of a truly interdisciplinary and highly active team ranging from computer scientists to bioinformaticians and clinicians, and learn about machine learning techniques, deep learning, clinical statistics, healthcare data storage and management, and medical image analysis.

As such the student will achieve a broad and comprehensive training in the use of computational algorithms to support advances in clinical medicine by exploiting Big Data.

The project builds on the extensive experience of the supervisory team in machine learning for medical image and data analysis and retinal biomarkers for systemic conditions accessing large bioresources, e.g. GoDARTS (n~= 15,000) and UKBioBank Eye and Vision (n >50,000).

The project is proposed under the auspices of the Computing-Ninewells iCIGHT initiative (informatics, Clinical Imaging, Genomics and Healthcare Technology).