The Cancer Genome Atlas launches

Historical milestone — this page describes an event dated 2006. It is not current breaking news.

A pilot, not a promise

On October 16, 2006, the National Cancer Institute and the National Human Genome Research Institute announced two more components of The Cancer Genome Atlas Pilot Project.

Read the wording of the notice and the modesty stands out. It calls TCGA a three-year, $100 million project. Its stated job was to test whether large-scale genome analysis could reliably find the genetic changes involved in cancer.

Nobody was promising treatments. The question on the table was narrower: can this even be done reliably at scale?

Three cancers to start with

The pilot picked lung cancer, brain cancer in the form of glioblastoma, and ovarian cancer.

NCI put up $11.7 million a year for seven new labs in five states. They were called Cancer Genome Characterization Centers. The named sites included the Broad Institute, Harvard Medical School, Lawrence Berkeley National Laboratory, Memorial Sloan-Kettering, Johns Hopkins, Stanford, and the University of North Carolina.

Each lab got a different job. Some looked at gene expression, which means the genes a tumor switches on. Some counted copy number, which means how many spare copies of a chunk of chromosome a tumor holds. One pair looked at methylation. Those are chemical tags that mute a gene without changing the letters of the code.

That split of work mattered. A cancer genome is not one thing to read. It has several layers, and each layer needs its own tool.

What it eventually produced

NCI's own summary of the program records the scale. TCGA molecularly characterized over 20,000 primary cancer samples, with matched normal tissue, spanning 33 cancer types.

Over the next dozen years it produced more than 2.5 petabytes of data. That covered the genome, the chemical tags on it, the RNA read from it, and the proteins built from it. All of it stays public for any researcher to use.

The program closed in 2018 with the Pan-Cancer Atlas. That was a set of studies looking across tumor types rather than within one. It covered cell-of-origin patterns, cancer-driving processes, and signaling pathways.

Why "matched normal" is the important phrase

This is the technical detail that makes the whole thing work, and it is rarely explained.

Every person's genome is full of harmless variation. Sequence a tumor on its own and you cannot tell which changes drive the cancer. Some of them are just that person's normal genetic makeup.

Sequencing normal tissue from the same person gives you a baseline. Subtract one from the other and what remains are the changes the cancer acquired.

Scale that to 20,000 pairs and a pattern appears. Changes that keep recurring across many tumors are likely drivers. Changes that show up once probably mean nothing. That split is routine in research now. Before TCGA it had no dataset behind it.

What it changed for patients, eventually

NCI says the data has already led to improvements in the ability to diagnose, treat, and prevent cancer. That improvement arrived slowly and indirectly.

The clearest effect is on how cancers are sorted. NCI describes cancer as a group of more than 100 diseases, defined by where they start and how the cells behave. TCGA showed two things that complicate that. Tumors from the same organ can look quite different at the molecular level. Tumors from different organs can share one driver. So some treatment choices now turn on a molecular finding rather than a location.

That is why a modern pathology report often carries a list of tested genes, and why a lung tumor and a colon tumor can occasionally be offered the same drug.

Where this touches an appointment

Nothing on this page is a test to ask for. But two things follow from it that are worth raising:

• After a new cancer diagnosis, ask what molecular or biomarker testing has been ordered, and when the results are expected.
• If a result names a gene, ask what it changes: eligibility for a drug, eligibility for a trial, information about prognosis, or nothing yet.
• If cancer runs in your family, ask about a referral for genetic counseling. That is a different test from tumor sequencing. It looks at inherited risk, not at the tumor.

What to keep in perspective

• TCGA was a research resource. It was not a treatment, and nobody in it was treated because of it.
• The samples came mostly from untreated primary tumors at a limited set of centers. The collection does not represent every patient. It also under-represents cancers that had already been treated or had spread.
• Finding a mutation is not the same as having a drug for it. Most of the alterations catalogued have no targeted therapy.
• Twenty thousand samples across 33 cancer types leaves rare cancers thinly covered.
• The pilot asked whether the work could be done at all. It answered that. Turning the answer into care has been slow and uneven, and it continues.
• This is a summary of a historical event, not advice about anyone's own care.

Sources

• NHGRI: NIH Announces Two Integral Components of The Cancer Genome Atlas Pilot Project (October 16, 2006)
• NCI Center for Cancer Genomics: The Cancer Genome Atlas Program (TCGA)
• NCI: What Is Cancer?

How this article was prepared

An AI-assisted editorial system helped prepare this page. No named medical reviewer has reviewed it unless one is listed.

Cancer Explained is published by the National Cancer Information Foundation. It is not medical advice and does not suggest a test or treatment.

See an error, old source, or unclear wording? Tell us.