In three new studies that could redefine how cancer is viewed, researched, and treated, scientists have created a detailed map of the genetic mutations that underlie two of the deadliest forms of the disease: pancreatic cancer and glioblastoma, the type of brain tumor that Senator Edward Kennedy was diagnosed with this past spring. The new findings are the first steps in the huge task of mapping the genomes of cancer, as researchers work to learn about cancers from the ground up.

Scientists have known for decades that cancer develops in response to genetic changes that cause cells to grow and divide uncontrollably. But uncovering each of these changes, and understanding how they lead to disease, is a Herculean task--one that involves sequencing and analyzing upward of 100 different kind of tumors, with hundreds of different patient samples of each. And while some believe that systematically cataloging the mutations could provide unprecedented insight into fighting or even preventing cancers, others believe that the high cost of such research might not be worth the rewards. These papers provide the first glimpse at what the rewards could be.

One paper, published online in Nature, is the first study born from data gathered by the publicly funded Cancer Genome Atlas (TCGA), an initiative created to use large-scale genome sequencing to find and map different cancers' genetic aberrations. Lynda Chin and Matthew Meyerson, both at the Dana-Farber Cancer Institute, in Boston, analyzed more than 200 glioblastoma tumors for genetic changes (such as the number of copies of each protein-coding gene present in the sample, and whether these genes have been turned off through a process called methylation), and they also analyzed 600 genes already implicated in the disease. Their results confirmed known culprits and revealed previously unknown changes in three major genes: two known tumor suppressors (NF1 and ERBB2), and one that is newly associated with cancer (PIK3R1) and could potentially be targeted by drugs already in development.

The other two studies--the fruits of a private cancer genome project headed by a trio of researchers at Johns Hopkins University, in Baltimore--analyzed far fewer tumors at a far greater level of detail. Published online in Science, these papers examine 22 pancreatic tumors and 24 glioblastomas for gene copy number and gene expression, as well as the sequences of just about every single one of their more than 20,000 protein-encoding genes. The researchers found an average of around 60 genetic changes per tumor, but they also discovered that most of those mutations acted on a core set of just 12 cellular pathways.

These pathways may be central to future drug development. "It may be more productive to screen for drugs that act against the core pathways," says Bert Vogelstein, one of the project heads at Johns Hopkins. "By targeting the pathways, it's possible that new drugs could be effective against a much greater fraction of tumors."