In August 1976, scientists first proposed that the development of cancer follows an evolutionary process. Since then, researchers have characterized the evolution of cancer from the perspective of random mutations and natural selection. If the mutations carried by cancer cells are highly adaptive, such cells will rapidly multiply and become the largest number of cell clones in the cell population. This phenomenon is called clonal sweep, which occurs repeatedly during tumor growth. Sequencing multiple regions of the same tumor over time is the most effective way to study the evolution of cancer; however, researchers can also reconstruct the evolutionary process with a single biopsy- Gerstung et al. in the fifth paper took this approach.
The author introduces the concept of "molecular time" here to classify cloned and subcloned mutations. They reasoned that subclonal mutations, which are only present in a portion of tumor cells, should have occurred late in the evolution of cancer. For clonal mutations present in all tumor cells, the authors classified the cloned mutations as early or late based on whether the mutation occurred before or after the copy number gain of the clone (an increase in the copy number of a gene or chromosomal region). The researchers summarized the evolutionary data of various tumors to determine some common mutation trajectories, such as APC-KRAS-TP539, which is a typical mutation sequence in colorectal cancer.
Gerstung et al.'S study found that the most common driver mutations in certain types of cancer often appear earliest. Similarly, if copy number gain occurs repeatedly in some cancers, it occurs earlier. For example, in clear-cell renal cancer, local copy number gains on chromosome 5 are common, and most often occur early in the development of kidney cancer. In contrast, genome-wide replication occurs later in this type of kidney cancer. Finally, researchers found that mutational characteristics change over time in at least 40% of tumors. These changes mean that as the disease progresses, the effects of environmental exposures gradually diminish, and the frequency and severity of DNA repair defects gradually increase. All in all, the team's findings suggest that driver mutations can occur years before a cancer is diagnosed, which is important for early disease detection and the development of biomarkers.
In the final paper, the PCAWG Transcriptome Core Team and their colleagues analyzed the transcriptome and whole-genome sequencing data of 1,188 tumors, and established a functional link between DNA mutations and RNA mutations. The team found a link between hundreds of single-nucleotide DNA mutations and expression of nearby genes. However, larger copy number variation is the main factor that drives gene expression changes in cancer cells. In addition, mutations are also related to structural changes in transcripts, such as the formation of new protein-coding regions (exons) in non-coding regions (introns).
The researchers also described how often "bridged fusion" occurs. Bridge fusion refers to the phenomenon that two genes are fused due to the insertion of a third DNA fragment. In the end, of the 1,188 samples analyzed, although 87 did not drive mutations at the DNA level, changes in RNA levels were found in each sample. In conclusion, it can be seen from these results that the integrated analysis of RNA and DNA sequencing results has an important role in cancer research.
These 6 papers and related papers published by other journals (see go.nature.com/3boajsm) can be regarded as milestones in the field of cancer and cloud genomics. Through analysis and inference, the Alliance has successfully advanced observation-based cancer sequencing research in the past decade. It is worth noting that although inferential analysis improves our understanding of cancer compared to descriptive research, its results are also more uncertain.
The openness and high quality of the PCAWG dataset will bring a new round of biological insights and promote the development of methodologies. Integrating it with other functional genomic data sets, such as detecting the 3D architecture of the genome, is bound to expand our understanding of the causes and consequences of genetic abnormalities.
The biggest limitation of current research is the lack of clinical data on patient treatment and outcome. These data can help researchers identify genetic changes that predict clinical outcomes. Fortunately, a project called the International Cancer Genome Alliance-Accelerated Genomic Oncology Research (ICGC-ARGO) is underway, which will build such a resource bank for more than 100,000 cancer patients.
The PCAWG brings together the strength of thousands of scientists to accomplish this mission together. The long-term impact of these collaborations not only comes from the research results published today, but also from the global collaboration of researchers and the knowledge exchange between members.
Original source: Marcin Cieslik & Arul M. Chinnaiyan. Global genomics project unravels cancer ’s complexity at unprecedented scale. Nature 578, 39-40 (2020)