Predictive models for cancer therapy can one day support the treatment choices a physician and patient make toward achieving the best possible clinical outcome. A team of collaborators from the DOE and NCI aims to accelerate this by using advanced computation to rapidly develop, test, and validate predictive pre-clinical models for precision oncology. They used observational data from experiments with cancer cell lines and animal models to do statistical analysis, deep learning, modeling, and simulation on Theta. This project has developed a suite of machine learning models for both single drugs and drug combinations, and became the first group to apply deep residual networks (DRN) to predict drug response both for single drugs and for multiple drugs.
Search results are limited to members of the NCI Pilot Group.Request Access to NCI Pilot Group