(MRC DTP) Statistical machine learning approaches cross-cohort multi-omics data

Multimorbidity refers to the co-existence of two or more chronic conditions in an individual and its prevalence is increasing due to aging populations. The management of multimorbidity is imposing an increasing burden on health systems and places significant strains on clinical practices that are traditionally adapted to deal with single disease conditions. Furthermore, a deep understanding of the causes of multimorbid conditions is made challenging by its complex association with factors such as socioeconomic deprivation, lifestyle, etc.

A large number of population cohorts are now available in the UK that have been screened using multiple molecular technologies. The availability of different types of molecular information allows a more comprehensive study of the mechanisms underlying certain physical traits as well as potentially disease aetiology. These cohorts therefore provide a valuable resource for the study of multimorbidity but typically the only molecular information that is common across all cohorts is genetic with each cohort surveying different combinations of molecular properties using different properties.

In this project, we will develop statistical machine methodologies that will allow us to integrate clinical and molecular information effectively from across different population cohorts. We will use genetic similarity as a common anchor by which to match individuals across cohorts and then develop new techniques that will allow us to leverage the comprehensive clinical profiling and multiomics information. Our goal is to combine information across the cohorts to allow us to identify molecular features that are associated with individual with multimorbid conditions. The project is likely to require developing competence in multi-view and transfer learning as well as flexible model construction using Gaussian Processes or deep neural networks.

This work is in collaboration with the University of Cambridge and supported by Health Data Research UK and the Alan Turing Institute.

Group website: http://cwcyau.github.io

Entry Requirements:
Applications are invited from UK/EU nationals only. Applicants must have obtained, or be about to obtain, at least an upper second class honours degree (or equivalent) in a relevant subject.