Research and Software
Resources for cancer modeling and high-dimensional inference
The fundamental assumption in all our models is that high-dimensional (i.e., genomic) data are concentrated on a low-dimensional manifold.
We develop statistical models that exploit this assumption for dimension reduction and inference of graphical models. We also use ideas from computational algebraic topology to infer higher order dependence structures - we call these simplicial models.
We apply these models to different types of genomic data (i.e., copy number variation, gene expression data, or SNP data) to infer mechanisms through which complex disease traits arise (for example, tumor progression, as depicted in the picture above).
This webpage contains information and links to our ongoing research projects and related software. Select a specific topic from the menu to the right.