Research Overview
We develop methods and code to extract knowledge about cancer origins and druggable vulnerabilities from patterns of somatic mutations.
Mutational signatures decoded from cancer structural variants
Structural variants (SVs) are the most cryptic type of somatic mutation in cancer. They make many cancers more aggressive, yet the origin of even simple SVs is poorly understood. We are developing algorithms to decode the origin of common types of SVs, in collaboration with leading experts in DNA repair.
Resolving mutational signatures in time
For therapeutic purposes, ongoing DNA repair deficiencies predict response to cancer drugs that target them as a vulnerability. Current methods cannot distinguish ongoing mutational processes from ones that have ceased — for example, due to acquired resistance to PARP inhibitor (PARPi) therapy. Together with clinicians, we are developing algorithms and assays that can make this distinction.
Understanding and detecting early cancer
We use mutational signatures to quantify the trajectory to cancer. This is made possible by highly specific mutational signatures found in individuals with a heightened predisposition to cancer. We want to understand which mutational signatures are “normal” during aging, and which ones are cancer-specific.
Synthetic lethality discovery
New and effective cancer drugs, like PARP inhibitors, exploit the principle of synthetic lethality. We use mutational signatures and AI to discover new drug targets and identify the patient populations most likely to benefit and respond.