Cancer Pharmacogenomics

The Mayo Clinic BEAUTY breast cancer clinical trial applies Next Generation sequencing to guide the therapy of women with high-risk breast cancer. Mayo also has generated a genomic data-rich cell line model system that has proved to be a powerful tool for generating and testing pharmacogenomic hypotheses. We will apply the KnowEnG framework to both of these data sets.

Research Plan

1. Predicting Drug Response

The BEAUTY clinical trial provides data on the response of patients to neoadjuvant chemotherapy, the LCL model system on cytotoxicity to a panel of ~25 anti-cancer treatments, along with molecular profiling of patients and LCLs, respectively. The goal of collecting such data is to predict drug response from a molecular profile: a classification problem. The next step is to distill the complex molecular determinants of drug response to the most essential sub-components shared across many patients: feature selection. This allows accurate targeted therapy at a fraction of the cost of complete molecular profiling.

2. Molecular Stratification

Molecular profiling of patients is already used to determine appropriate therapy, but for only a few diseases. In those cases, the targeting is based on simple biomarkers, and many patients do not respond. It is possible that the known biomarker is part of a more complex molecular signature or one of multiple signatures. Identifying sub-groups of patients carrying distinct, potentially complex signatures is called molecular stratification and will be one of our research goals.