This short note describes how data and computing could help find cancer cures. It discusses three projects supported by the National Cancer Institute (NCI) and the Department of Energy (DOE); each uses machine learning to predict which tumors will respond to which drugs. The pilot projects described include one at Oak Ridge National Laboratory, which is mining 40 years of patient records; another at Argonne National Laboratory, which is building models to predict the reactions of new tumors to drugs; and the last at Lawrence Livermore National Laboratory, studying one critical cancer gene and its mutations.

The article explains that research is combining better sensor technology with larger data records of patient experiences, and exascale computing. The author hopes that good predictive models can catch up to rapidly mutating tumor cells and predict which drugs to use. This is not a research paper; it is an introduction, but has a few references for those who wish to learn about the efforts being made.