Research Areas

Medical Genetics

alt text 

There are estimated to be a total of over 8,000 rare Mendelian disease while less than 5,000 diseases have been definitively diagnosed thus far. Hundreds of novel gene-disease-phenotype relationships are discovered every year resulting in a continously growing knowledgebase in literature. It is a time consuming process for clinicians to diagnose patients with rare diseases (estimated 1hr/gene). I am developing computational methods and tools to improve the disease diagnosis process.

Variant Classification

alt text 

A typical individual has over 4 million single nucleotide variants (SNVs). Only a small fraction of these variations are thought to likely be involved in or related to disease. I am developing machine learning methods to model signatures of evolutionary importance to better predict the likelihood that a given variant is functionally important and disease-causing.

Genome Privacy

alt text 

The human genome uniquely identifies each individual revealing much about one's gender, ethnicity, susceptibility to complex diseases and much more. Sharing a genome can lead to unintended consequences such as discrimination and identification. I am developing methods to protect an indivdual and their relatives privacy while still providing the flexibility to successfully identify the causative variant in a patient.