The EACR’s ‘Highlights in Cancer Research’ is a regular summary of the most interesting and impactful recent papers in cancer research. Previously known as our Top 10 Cancer Research Publications, it is curated by the Board of the European Association for Cancer Research (EACR).
The list below appears in no particular order, and the summary information has been provided by the authors unless otherwise indicated.
Use the dropdown menu or ‘Previous’ and ‘Next’ buttons to navigate the list.
Sherman, M.A., Yaari, A.U., Priebe, O. et al. Nat Biotechnol (2022).
doi: 10.1038/s41587-022-01353-8.
Summary of findings
The search for mutations that cause cancer has focused primarily on the 2% of the genome that codes for proteins. We developed a computational tool, Dig, that enables the entire genome – coding and noncoding alike – to be searched for mutations that may drive cancer.
Dig uses deep learning to map cancer-specific somatic mutation rates across the whole genome based on the epigenetic organization of DNA. The number of mutations observed in a cohort of tumors can then be compared against Dig’s maps to pinpoint regions of the genome with unexpected mutational patterns, a crucial signal that the region may be driving cancer.
We created mutation rate maps for 37 types of cancer and applied them to find potential drivers. Across all types of cancer, we found that mutations within gene introns likely account for 5% of single nucleotide driver mutations in tumor suppressor genes by inducing cryptic alternative splicing. We also found that tumor suppressor genes may be inhibited by rare mutations in 5’ UTRs. Finally, by examining large cohorts profiled with targeted sequencing panels, we found that genes that commonly drive one type of cancer can infrequently drive other types of cancer as well.
Future impact
The discovery of cancer-causing mutations within protein-coding DNA paved the way for targeted therapies. Now, our method together with the application of whole genome sequencing to tumor samples may open up therapeutic avenues within the noncoding genome. For example, antisense oligonucleotides, which can reverse cryptic splicing, may be effective treatment options for some patients with intronic driver mutations. More broadly, our work demonstrates the power of deep-learning to provide insights into cancer biology. As the volume of genomics data continues to increase, we believe these techniques will be important tools to unravel the complexities of cancer.
