Skip to main content
Premium Trial:

Request an Annual Quote

Researchers Generate Catalog of Noncoding Somatic Mutations in Cancer Genomes

NEW YORK — Researchers led by a team at Dana-Farber Cancer Institute have compiled a catalog of noncoding somatic mutations that arise in different cancer types.

The development of somatic mutations is a hallmark of cancer, though most analyses have focused on those affecting protein-coding regions of the genome. In a new study appearing Thursday in Science, researchers from Dana-Farber and their colleagues focused instead on somatic mutations that crop up in noncoding regions.

"The opportunity to study noncoding mutations is relatively recent," first author Felix Dietlein, now an assistant professor at Boston Children's Hospital, said in an email.

Using a combination of three different approaches, the researchers searched through whole-genome sequencing data of nearly 4,000 people representing 19 different cancer types. Their analysis uncovered not only coding events near known driver genes but also noncoding events near tissue-specific genes and in the regulatory or enhancer regions of cancer-linked genes that could be driver events themselves. Such regions point to potential therapies and diagnostics, according to the researchers.

To detect the somatic mutations, the researchers tiled the genome at 1-kilobase, 10-kilobase, and 100-kilobase intervals and conducted three significance tests in each interval. Those tests gauged whether the regions harbored more mutations than expected based on their epigenomic signals — as mutation rates vary between heterochromatic and euchromatic regions — whether the number of mutations in the region varied by cancer type, and whether the mutations clustered more than expected. This approach, Dietlein noted, builds on previous work by the Pan-Cancer Analysis of Whole Genomes consortium.

In all, there were 142 events in coding regions across the 19 cancer types, but also 73 events in regulatory regions, 70 events near tissue-specific genes, and 87 other types of events.

Of the 142 events in coding regions, 93 percent involved known cancer-driver genes and 96.5 percent matched results from MutSigCV and dNdScv, both established approaches for identifying coding drivers. These results, the researchers said, support the robustness of their approach at identifying somatic mutations.

The events uncovered in regulatory regions were enriched for canonical cancer genes. For instance, there were mutations in the TERT promoter region within bladder, brain, head and neck, and other cancer samples as well as noncoding mutations upstream of FOXA1 in breast cancer but downstream of FOXA1 in prostate cancer.

At the same time, the researchers identified 46 additional events in the promoters and enhancers of cancer-associated genes, such as within the promoter of FGFR2 in bladder and lung cancer and the promoters of B2M, KLF6, and SRCAP in lung cancer.

Other events, though, fell near tissue-specific genes and could represent the expression programs of the tissues of origin rather than reflect candidate driver mutations.

A third class of noncoding mutations fell into neither of the previous categories or had conflicting results. Some of these events, though, were still enriched for canonical cancer genes, such as the mitotic spindle assembly checkpoint genes MAD1L1 and MAD2L1 in brain and ovarian tumors and the tumor suppressor gene NF1 in breast cancer.

One event in particular stood out to the researchers: In breast cancer, they noted a number of mutations near XBP1 that fell in a regulatory region outside its promoter. Using both CRISPR-interference screening and luciferase reporting, they found that these mutations increased XBP1 expression. This, Dietlein noted, "demonstrates that mutations outside the traditional regulatory regions can be functionally relevant and that examining noncoding mutations can implicate new genes in tumor development."

According to Dietlein, this analysis and other studies are beginning to tease out what roles noncoding mutations might have in cancer development. "Moreover, mutations have been a major propeller to inspire the development of new cancer therapies," he added, and their work may help advance these.