Improving estimation and prediction of common complex disease risk

Despite decades of study, there is generally a poor understanding of the modifiable risk factors for common disease (lifestyle, diet, environmental exposure), with a limited number identified for each disorder. Genetic association studies on the other hand suggest that most common diseases are underlain by many genomic regions, each of small effect, and that many of these disorders share a genetic basis. There is now substantial evidence that there are many different underlying reasons why patients display symptoms of a disease, some of which are interlinked. This proposal focuses on the development of statistical approaches that can be applied to the study of obesity, type-2 diabetes (T2D), asthma, cardiovascular disease (CVD, including blood pressure, stroke, cardiac arrest, angina) and psychiatric disease to better understand the underlying causes of disease and the relationships among them. Our approach provides improved prediction and understanding disease risk as it enables novel biomarkers and environmental and genetic risk factors to be identified for a range of complex diseases. We aim to determine which risk factors are unique to a disease outcome (e.g. T2D) and which risk factors act indirectly through an intermediate underlying drivers of disease risk (e.g. BMI, hypertension, glucose tolerance, etc.). This will provide better modelling of the consequences of clinical intervention, as we know which factors are independent, giving improved prediction and stratification of patients as they will be scored not only on their overall risk of disease outcome, but also for different intervention strategies.