Blueprint Medicines, a startup based in Cambridge, Massachusetts, plans to use the growing amount of genomic information about cancer to create new drugs targeted at the mechanisms that drive specific subtypes of the disease. The company, which announced its creation last week with $40 million in funding from Third Rock Ventures, reflects a growing trend toward defining cancers not by their location in the body but by the particular collections of genetic mistakes that enable tumor cells to grow out of control.

“How the community, both academic and industry, understands cancer is really shifting from largely a pathology-centric viewpoint—classifying cancers based on what we can see in the microscope—to a molecular viewpoint,” says Chris Varma, the company’s president and cofounder. “What we find is that the mechanisms that are driving cancer come up again and again in different environments. The same mechanism responsible for some breast cancers could be driving a subset of brain cancer or melanoma.”

The growing understanding of cancer genetics has enabled researchers to develop an increasing number of drugs designed to zero in on cancer cells. Varma says Blueprint will use genomics and novel chemistry to develop such drugs in a more systematic way, targeting a broader range of molecular mistakes.

In addition to Varma, cofounders include Nicholas Lydon and Brian Druker, who led the development of Gleevec, one of the first targeted cancer drugs; Scott Lowe, deputy director of the Cold Spring Harbor Laboratory Cancer Center; and David Armistead, a biotech entrepreneur.

Genomic data on cancer has been pouring in to public databases over the last few years, thanks to projects such as the Cancer Genome Atlas, an NIH-funded effort to read the entire DNA sequence of various cancers. “There are now literally hundreds of genome efforts on various types of cancer,” says Lowe, “and all of this data is in the public domain and can be readily accessed.”

One of Blueprint’s goals will be to use this information to identify new drug targets. “Right now, we have lot of experience in identifying these alterations; we can catalogue and sequence genomes and identify changes present in particular cancers,” says Michael Hemann, a cancer researcher at MIT who isn’t involved with Blueprint. “The problem is, what do we do with that information? Which of these changes are relevant, and what do we do about it? For example, cancer cells can accumulate many mutations, some that drive abnormal growth and some that have no effect. And scientists need to identify which is which.”

Blueprint is building a platform that will initially use computational algorithms to analyze genomic data for potential molecular mistakes linked to cancer. For example, they would detect whether multiple data sets drawn from cancers of different types reflect mutations in the same enzymes. Researchers would then run tests on cells or animals to determine which of the mutations identified actually drive the growth of cancer cells and are thus the most promising targets for cancer drugs.

The company is also building a proprietary library of chemical compounds called kinase inhibitors—molecules that block activity of kinase enzymes, which have been linked with cancer. Armistead says the researchers will use computational modeling and x-ray crystallography to create a library of structures and then test the compounds’ ability to inhibit different kinases. They plan to generate and screen the library within the next 12 to 18 months. “We also have a short list of targets we already have interest in, which we will begin working on immediately,” says Armistead.