Computational phenotyping to understand disease progression across the human lifespan
In the last century improvement in general societal conditions and technological advance in medical care has increased life expectancy dramatically. This has resulted in longer lives but also lengthening the proportion of life which is characterised by chronic health conditions. It is thus a fundamental goal for health research today to not only further increase life expectancy, but to improve quality of health and thus life quality in the last part of our lives. This goal needs much better prevention interventions and medications to be achieved, which require a better understanding of both the biological and non-biological factors which lead to a decreased health status before the onset of disease.
UK Biobank is unique in the sense that is had deeply phenotyped individuals (thousands of recorded measurements, diagnoses, medications, etc) and large-scale multi-modal imaging (MRIs, retinal fundus scans, Optical coherence tomography, etc). Imaging data is a very complicated and rich source of information which hasn't been fully utilised to date. We will develop machine learning algorithms to automatically extract phenotypes from these images, that can help us characterise the chronic conditions of ageing.
Our project has the bold aim of combining statistical and bioinformatics approaches to dissect the complex relationships which lead to a reduced quality of life, understanding which factors are driving complex disease burden and how. Furthermore, a primary focus will be to examine the role that inflammation plays in the healthy aging process and how this relates to the measures derived from the imaging data.
Initially, we set the duration of the project at 36 months.