Integration of a prognostic gene module with a drug sensitivity module to identify drugs that could be repurposed for breast cancer therapy

Abstract

BACKGROUND Efficiently discovering low risk drugs is important for drug development. However, the heterogeneity in patient population complicates the prediction of the therapeutic efficiency. Drug repositioning aiming to discover new indications of known drugs provides a possible gateway. METHOD We introduce a novel computational method to identify suitable drugs by using prognosis information of patients. First, we identify prognostic related gene modules, Prognostic Gene Ontology Module (PGOMs), by incorporating multiple functional annotations. Then, we build the drug sensitivity modules based on gene expressions and drug activity patterns. Finally, we analyze the potential effects of drugs on prognostic gene modules and establish the links between PGOMs and drugs. RESULT AND DISCUSSION With PGOMs generated based on the patient outcome, FDA approved drugs for breast cancer treatment have been successfully identified on one hand; several drugs that have not been approved by FDA, such as Etoposide, have found to strongly associate with the outcome on the other hand. With PGOMs generated based on the patient ER status, Tamoxifen and Exemestane rank at the top of the drug list, suggesting that they may be more specific to ER status of breast cancer. Especially, the rank difference of Exemestane in ER+ group and ER- group is very large, demonstrating that Exemestane may be more specific to ER+ breast cancer and would cause side-effect to ER- breast cancer patients. Our method can not only identify the drugs that could be repurposed for breast cancer therapy, but also can reveal their effective pharmacological mechanisms.

DOI: 10.1016/j.compbiomed.2014.12.019

Cite this paper

@article{Zhu2015IntegrationOA, title={Integration of a prognostic gene module with a drug sensitivity module to identify drugs that could be repurposed for breast cancer therapy}, author={Lida Zhu and Juan Liu}, journal={Computers in biology and medicine}, year={2015}, volume={61}, pages={163-71} }