The study was funded by the Lorraine Dell Program in Bioinformatics for Personalized Cancer Medicine; an MD Anderson Faculty Scholar Award; the Institute of Information & Communications Technology (B0101-15-0104); the Development of Supercomputing System for the Genome Analysis; the Korea Health Industry Development Institute (KHIDI); the Ministry of Health & Welfare, Republic of Korea (H114C0072 and H117C1836); and the Korean National Research Foundation (NRF-2016R1D1A1B03934001 and NRF-2017R1A2B2012796). The cancer types range from common forms like colorectal and breast cancers, to rare cancer types including pancreatic and brain cancers.
Altogether, the project has sequenced 2,658 whole-cancer genomes from 38 different tumor types.
WATCH: Professor Sean Grimmond discusses the work behind the Global Cancer Atlas.
The genomic data-more than 800 terabytes' worth-have been shared with the research community to continue combing for new clues about cancer's inner workings.
In the current study, the researchers applied seven different pathway and network analysis methods to noncoding regions of whole genome sequences from more than 2,500 cancer patients - the largest collection to date of uniformly processed cancer genomes.
Dr Brewer said: "We found that the presence of HPV is mutually exclusive to some of the important changes seen in head and neck cancer, which means that the HPV virus can help the cancer develop without additional changes to its genome". Of the mutations, about 100 fell outside the protein-coding areas of the genome. "For more than 30 cancers, we now know what specific genetic changes are likely to happen, and when these are likely to take place".
To begin the project, every cancer tumour was described by a pathologist including its tissue of origin and cancer stage.
"Cancer genome sequencing is beginning to enter the clinic", Campbell said.
The first signs of cancer can appear years if not decades before diagnosis, scientists have discovered, raising hopes for new tests to detect the disease much earlier.
A massive decade-long study has sequenced the genomes of 38 types of cancer revealing secrets about how tumours form
Scientists also developed a way of "carbon dating" mutations.
"This is particularly exciting as it provides a rationale for future testing of drugs known to be successful in one cancer type in patients with very different tumours that bear the same driver mutation profile".
The worldwide team also discovered that DNA might have many different mutations at the same time.
Some of the most surprising results were finding the sheer number of genetic changes, what we call mutation signatures.
The Pan-Cancer Project has made available a comprehensive resource for cancer genomics research, including the raw genome sequencing data, software for cancer genome analysis, and multiple interactive websites exploring various aspects of the Pan-Cancer Project data.
For example, the work found the early development of some cancers can occur decades before diagnosis, sometimes even in childhood. This uniqueness is a huge advantage when a patient presents with a cancer that has moved or metastasized.
Better understanding the type and number of driver mutations that define someone's cancer could also let physicians compare any given patient's tumor to a larger database of similar cases.
"If we don't understand where a cancer comes from, we can't even rely on traditional clinical approaches to treatment", Professor Grimmond says.
This unprecedented level of data crowd-sourcing means researchers, clinicians and industry can now work together, anywhere across the globe.
Perhaps the most interesting noncoding mutations found in the study were located near genes involved in RNA splicing, a key intermediate step that allows genes' instructions to be translated into the proteins that carry out cellular activities.
The new study provides a new way to detect cancer earlier and providing the needed and appropriate treatment for the cancer type.