This $138,428 project grant from the National Science Foundation's Computer and Information Science and Engineering program (CFDA 47.070) will support the development of computationally tractable structured hierarchical models to identify complex genetic associations hidden from current methods. The grantee, J. David Gladstone Institutes, will analyze genomic and clinical trait data using machine learning to build predictive models for individualized disease risk assessment and personalized medicine.
Specifically, the grantee will pursue three integrated projects. The first will develop models and robust inference methods to exploit correlations among multiple traits to improve power in finding relationships between high-dimensional genetic and clinical data sets. The second will create a predictive model regularizing both additive and nonadditive effects to recover variants with complex predictive impacts. The third will develop a hierarchical causal analysis model to pinpoint cellular disease mechanisms by moving beyond unrealistic assumptions in current genomic causal inference approaches. Outcomes will enable more complete understanding and individual risk prediction of genetic disease regulation.
Generated 1/6/24, 7:01 PM