Statistical methods development for integrating genomic and multi-scale data
Principal Investigator: Dr Quan Long
Approved Research ID: 45132
Approval date: May 29th 2019
The effectiveness of precision medicine relies on healthcare plans that can be tailored to an individual's traits. To plan the most helpful course of treatment of action, an individual's traits needs to be accurately predicted from their unique combination of genetic material. Two recent technological advancements are bringing this predictive pipeline into the realm of clinical possibility. The first is the increased availability of linked genetic and clinical information. Bringing these variables together facilitates cross-disciplinary insight, where the predictions of one can inform the other. The second is the application of machine learning to healthcare problems, leveraging next-generation computing techniques to make full use of high-dimensional biological data. However, there is a gap between theoretical predictive performance and the complex reality of patient outcomes. In this project, we will study the integration of genomics with other data such as imaging files and health outcomes to augment experimental design. By pre-identifying associations of biological significance, our tool will optimize the research process by reducing the time and cost of making clinical interesting discoveries. The ultimate goal is to predict various medically important traits, including risk of diseases, responses to potential medication, etc.