Since the discovery of the first cancer-causing genes in the 1970s, researchers have been eager to catalogue the mutations that can cause cancer. Each mutated gene holds the potential to expand our understanding of what causes the disease — and how to treat it.

The latest progress towards that goal was on display last week, when 18,400 people descended on San Diego in California to attend the annual American Association for Cancer Research meeting. Researchers showed how patterns of mutation can be used to track down the agent that caused them — sunlight, for example, leaves a footprint that differs from a cancer-causing viral infection. Another team had catalogued cancer-associated mutations in patients with advanced melanoma, hoping to use the information to tailor immune cells to destroy tumours. And promising initial results were unveiled on targeting a protein called IDH2, mutations in which crop up in many different tumour types (see Nature 508, 158–159; 2014).

It has taken a massive effort to make such achievements possible. Seventeen countries have invested in sequencing cancer genomes through the International Cancer Genome Consortium (ICGC), which aims to sequence more than 25,000 samples. The largest and oldest component of that project is The Cancer Genome Atlas (TCGA) at the US National Cancer Institute (NCI) in Bethesda, Maryland, which intends to characterize 10,000 tumours.

TCGA was initially controversial, because researchers worried that the project would direct funds away from grants to individual investigators. Early results — which showed that cancer mutations were much more abundant and diverse than expected — even evoked schadenfreude in some circles (see Nature 455, 148; 2008). Criticism died down as the project bore fruit.

But TCGA is now winding down: the project ceased collecting new tissue samples last December. The ICGC, too, has virtually stopped accepting proposals for new projects. TCGA aims to complete sequencing and further characterization of its cancer samples by the end of the year. After that, a few groups will receive funding to analyse the data for the next two years. But the programme, as it existed, will cease.

The end of the older projects should herald a needed transformation in the field.

Some cancer researchers have advocated that the programme should continue. Stopping now would be premature, their argument goes, because we have yet to achieve a comprehensive catalogue of cancer-causing mutations. A study published earlier this year determined that compiling a list of mutations present in at least 2% of cancers would require sequencing of about 2,000 tumours in each of at least 50 tumour types (M. S. Lawrence et al. Nature 505, 495–501; 2014). For most cancers, we are still far short of that goal.

On a more optimistic note, the end of the older projects should herald a needed transformation in the field. When TCGA and the ICGC were launched, the technology dictated that only fresh tumour samples could be sequenced. That restriction limited researchers’ ability to link sequencing data with clinical outcomes because that information might not be available until years after the sample was taken. Also, at that time, oncologists did not take more than one biopsy of a tumour from a patient, which limited studies of how tumours changed during and after treatment, and how metastases differed from primary tumours.

Those hurdles have now been largely surmounted. Improved techniques allow researchers to sequence DNA collected from tissue preserved in formaldehyde and embedded in paraffin, opening the door to using banked samples with linked clinical data. And although still an uncommon practice, several clinical trials have shown that many patients are willing to submit to extra biopsies.

To replace TCGA, the NCI intends to sequence tumours from patients enlisted in some of its clinical trials. Other teams will no doubt do the same, allowing researchers to learn more about the importance of a given mutation by associating it with the response to therapy or to overall prognosis.

These changes require a new mindset. Clinical researchers will need to change consent forms for donation of tissue samples, to allow the association of clinical data with the sample. They will have to collect their samples using protocols that ensure utility not just for classical pathology but also for sequencing. Data security, always a concern when dealing with patient information, will need to be bolstered.

Nevertheless, to continue the work is a worthy undertaking. The end of TCGA also represents an opportunity for the field to balance its cancer-genomics projects more evenly between cataloguing mutations and studying their functional significance. Functional studies have had short shrift, whereas sequencing — a simple concept, and easier to communicate to policy-makers and the public — has taken the lead. Correcting that imbalance will lead to exciting discoveries for science and for patients.