Genotype-based clustering, prediction, and functional annotation across autoimmune diseases
Approved Research ID: 91246
Approval date: January 12th 2023
Autoimmune diseases are caused by auto-antigens harming the body's cells and tissues. Still, they show considerable heterogeneity concerning disease presentation and outcome, with genetic and environmental factors contributing to their onset and progress.
Genetic factors underlying autoimmunity have been identified, but predicting personal disease risk based on genetics is challenging any by no means standard in medical care yet. This is because both common and rare private mutations contribute to individual diseases, whose investigation requires a large enough and diverse autoimmunity patient cohort for which comprehensive genetic data is available. The UK Biobank is currently the only feasible database for such analyses.
Here we will follow a novel strategy to stratify autoimmune disease patients into more homogenous subgroups based on common genetic and molecular features. Instead of focusing on a single autoimmune disease, we will use all patients within UK Biobank with any autoimmune disease. Altogether there are about 28,000 cases of 20 different autoimmune diseases. We will first search for shared genetic factors and affected molecular networks using unsupervised clustering techniques. In the second step, we will use deep learning in combination with the patient's clinical data to predict autoimmune disease and to identify the molecular mechanisms that drive the transition and progress from health to autoimmune disease in an individual. Altogether, the project will show if there are individual-level genetic characteristics across autoimmune diseases that can be exploited for personalized disease prediction, management, and treatment