AI for precision medicine: Subgroup discovery in survival analysis with high-dimensional data

Subgroup discovery is a statistical methodology aimed at identifying promising subgroups of individuals that respond to treatment or have a higher susceptibility to disease risk factors. These methods help inform the design of future research studies and clinical trials to confirm and characterise these individuals, with impact on the prescription of treatment and government policy.

This project will develop novel probabilistic models for subgroup discovery in survival analysis based on the fusion of Bayesian statistics and deep learning architectures and apply them to real data to identify risk factors influencing survival or disease onset. We intend to develop statistical software for the application of these models in the Comprehensive R Archive Network and make it publicly available.

The specific focus areas will be:

·         Methodologies for subgroup discovery in survival analysis based on mixture models.

·         Model averaging and variable selection methodologies for multi-dimensional data

·         Deep learning architectures and their application to high dimensional data in survival analysis

·         Development of an R statistical package to implement these methodologies

Training

The successful candidate will receive comprehensive research training including machine learning, deep neural networks, bioinformatics and statistical genetics, hands-on practice with several clinical trial and epidemiological studies, in addition to an extensive training programme in technical, personal, and professional skills at the Doctoral College of the University of Surrey. The candidate will have multiple opportunities to present his work in conferences, workshops, and seminars.

The Research Team

Dr Alex Couto Alves is Assistant Professor (lecturer) in Bioinformatics and Statistical Genomics and Head of the Bioinformatics Core Facility. He supervises a team of bioinformatics research officers and PhD Students interested in biostatistical models of longitudinal data¹. Dr Zhenhua Feng is Assistant Professor (lecturer) in Computer Vision and Machine Learning at the Department of Computer Science. Dr Feng supervises a team of PhD Students focused on deep learning architectures for high dimensional data^2,3,4. Prof. Simon Skene is Professor of Medical Statistics and Director of Clinical Trials Unit and provides leadership in statistics and research methodologies to support decision-based practice in health and medicine⁵.

The Research Environment

Surrey University hosts the largest GPU facility in the UK supporting world-leading research in AI, machine learning and data sciences as part of the Joint Academic Data Science Endeavour (JADE). The research in this project is affiliated with Surrey’s i) AI Institute, ii) The Centre for Vision, Speech and Signal Processing, and iii) the Centre for Mathematical and Computational Biology. Besides providing a stimulating environment for cross-disciplinary collaborative research, these groups will foster multiple training opportunities for the student.

A 3.5-year fully funded studentship open to applicants worldwide starting in October 2022. Funding includes stipend, full fees and a research grant.

More information on the School of Biosciences and Medicine.

Entry requirements

A First or Upper Second-Class Honours degree from the UK (or equivalent qualification from international Institutions) or Masters degree in a relevant subject area.

English language requirements: An IELTS Academic of 6.5 or above with 6 in each individual category (or equivalent qualification from other agencies).

How to apply

Applications should be submitted via the online application portal for Biosciences and Medicine PhDs.

This project is part of the Faculty-funded studentship scheme and you can express interest in one or two of the projects available via this scheme. When completing your application, in place of a research proposal, please provide a 1-page (maximum) document containing the reference numbers(s), project title(s) and supervisor name(s) of the project or two projects you have selected, together with an explanation of your motivations for wanting to study for a PhD and your reasons for selecting the project(s) you have chosen.

The reference number for this project is FHMS PL-BM-16 .

For those interested in the project described above, we strongly encourage informal enquiries to be sent to Dr Alex Couto Alves ([Email Address Removed]).