The dwindling drug pipeline is driving increased interest in the use of genome datasets to inform drug treatment. In particular, networks based on transcript data and protein-protein interactions have been used to design therapies that employ drug combinations. But there has been less focus on employing human genetic interaction networks constructed from copy number alterations (CNAs). These networks can be charted with sensitivity and precision by seeking gene pairs that tend to be amplified and/or deleted in tandem, even when they are located at a distance on the genome. Our experience with radiation hybrid (RH) panels, a library of cell clones that have been used for genetic mapping, have shown this tool can pinpoint statistically significant patterns of co-inherited gene pairs. In fact, we were able to identify gene pairs specifically associated with the mechanism of cell survival at single gene resolution. The strategy of seeking correlated CNAs can also be used to map survival networks for cancer. Although the cancer networks have lower resolution, the RH network can be leveraged to provide single gene specificity in the tumor networks. In a survival network for glioblastoma possessing single gene resolution, we found that the epidermal growth factor receptor (EGFR) oncogene interacted with 46 genes. Of these genes, ten (22%) happened to be targets for existing drugs. Here, we briefly review the previous use of molecular networks to design novel therapies. We then highlight the potential of using correlated CNAs to guide combinatorial drug treatment in common medical conditions. We focus on therapeutic opportunities in cancer, but also offer examples from autoimmune disorders and atherosclerosis.