Understanding how genotype (variation in DNA) relates to phenotype (observed charactereistics of an individual) is at the foundation of modern biomedical science. Genome-wide association studies and quantitative trait locus mapping have been successful in linking genetic variation to phenotypes of interest, for example disease status or gene expression in specific tissues and cell types. However, with datasets growing in size and complexity, new approaches are needed. Graph foundation models offer a new way to integrate relationships between observations and data types. This project will explore the use of these methods to develop new approaches to mapping genotype to multi-omics phenotypes in diseases like pulmonary fibrosis.
Supervised by
