Approved research

Information-Theoretic Inference of Complex Multivariable Associations: Improving the Genetic Discrimination of Type 1/Type 2 Diabetes

Lay summary

In our work we are focusing our attention on one of the aims of UK Biobank: to help scientists discover why some people develop particular diseases and others do not. In particular, we will focus on the problem of distinguishing Type 1 and Type 2 Diabetes, and non-diabetic subjects. Genetics play a large role in both of these diseases with multiple genes affecting the disease in a complex way. We have developed powerful, new computational methods to find complex genetic contributors to diabetes. Using these new methods, we will work on find multiple genes and their relationships, which distinguish participants with Type 1 Diabetes (T1D), Type 2 Diabetes (T2D) or neither. Careful analysis of the function and relationships of these genes in each of these three classifications will allow us to build an efficient computational approach to distinguishing Type 1/Type 2 Diabetes. This will enable us to use genetic data to make reliable predictions of disease occurrence. To improve the statistical power of our analysis we will use the full cohort. It is important to note that our computational methods are general, and can be further extended to detect the complex relationships among genes and phenotypes in any other disease or pathology.