School President and Faculty Member Harold Varmus

In 2005, 15 years after the start of the Human Genome Project and two years after the full human genome sequence was completed, the National Cancer Institute and the National Human Genome Research Institute (two components of the National Institutes of Health) announced the launch of the pilot phase of The Cancer Genome Atlas. Known as TCGA -- the initials also represent the four chemical building blocks in DNA (thymine, cytosine, guanine, and adenine) -- the project seeks to accelerate the understanding of the molecular basis of cancer through the application of a variety of genome analysis technologies.

Memorial Sloan-Kettering Cancer Center President Harold Varmus was director of the NIH during much of the time that the Human Genome Project was under way. More recently, he also was a member of the working group that recommended the formation of a project to analyze the human genome comprehensively in many types of cancer.

"The cancer genome project was conceived to create a catalog of every type of genetic change that can lead to cancer, and to link these changes to clinical data," Dr. Varmus said. "As with the Human Genome Project, all of the information is being placed in a free public data repository so that any researcher worldwide can access it."

TCGA is initially focusing on three types of cancer that are especially difficult to treat: ovarian cancer; squamous cell non-small cell lung cancer; and glioblastoma, the most common and aggressive type of primary brain tumor. The goal is to study 500 samples from each tumor type. The first phase of the pilot project must meet clear milestones and goals before a large-scale effort for additional types of cancer is initiated.

Cancer Genome Characterization Centers (CGCCs) were established at seven institutions around the country to study different types of genetic changes in the same tumor samples, and Memorial Sloan-Kettering was funded to house one of these CGCCs, under the leadership of molecular pathologist Marc Ladanyi. The CGCC is located in Memorial Sloan-Kettering's Genomics Core Laboratory.

"The idea is to do a definitive, comprehensive, fully integrated profiling of all the genomic alterations in these samples," Dr. Ladanyi said. "We can't anticipate what genomic profiling might look like in the future, but today this would be the state-of-the-art, most detailed, exhaustive view of all the genetic alterations that are present in these three cancers."

The CGCC at Memorial Sloan-Kettering is one of four centers that are studying changes in the number of copies of genes -- whether particular bits of the genome are gained (have extra copies) or have been deleted (have lost copies) in a given tumor sample. "When you do this in hundreds of samples of the same cancer," Dr. Ladanyi explained, "you see that there are characteristic genetic changes that occur over and over again, in tumors from different patients." The other centers looking at copy numbers are using different approaches and in some cases different technologies. "The idea is that each technology has its own advantages and disadvantages, and by using several methods to measure the same thing, you can get the most accurate picture of this class of genetic abnormalities," Dr. Ladanyi said.

Other characterization centers are looking at additional aspects of genetic changes in tumors, including patterns of gene expression, changes in microRNAs (which regulate gene expression), and alterations called DNA methylation (a modification that does not change the gene sequence but alters expression of neighboring genes). There are also three Genome Sequencing Centers that use high-throughput methods similar to those used for the Human Genome Project, and a bioinformatics center to analyze all of the data that is being generated.

Some members of the research community have criticized TCGA's approach, questioning whether it is the most efficient way to find genetic changes related to cancer and expressing concern about the cost of the project in relation to the amount of data it is likely to generate. One alternative that's been suggested is a so-called functional genomics approach, which looks for normal cellular pathways that play a role in cancer by using molecules called short hairpin RNAs (which are engineered to target and suppress specific genes) and studying the response of cancer cells.

"The various approaches to learning more about cancer genes complement each other," said Dr. Varmus. "Ultimately, the more data we have about all the genes that play a role in cancer, the faster we will be able to develop diagnostic procedures and drugs that will benefit patients."