Only a small proportion of our DNA contains nucleotide sequences used to make proteins. Much of the remainder is devoted to specifying how, when and where those proteins are made. These rules are encoded in our DNA as regulatory elements, and they're what makes one cell type different from another, and keep them from running wild like children in an unattended classroom. When things go awry, the consequences (like rampant growth and cancers) can be severe.
Geneticist Michael Snyder, PhD, and postdoctoral scholar Collin Melton, PhD, recently combined information from The Cancer Genome Atlas, a national effort to sequence and identify mutations in the genomes of many different types of cancers, with data from the national ENCODE Project, which serves as an encyclopedia of DNA functional regions, or elements. Their aim was to better understand the roles that mutations in regulatory regions may play in cancer development.
Snyder and Melton found that fewer than one of every thousand mutations in each cancer type occurs in the coding region of a gene. In contrast, more than 30 percent of the mutations occur in regulatory regions. The study was published this morning in Nature Genetics.
As Snyder explained to me:
Until recently, many mutations outside the coding regions of genes have been mostly invisible to us. Cancer researchers largely focused on identifying changes within coding regions. Using ENCODE data, we’ve been able to define some important regions of the genome and found that certain regulatory regions are often enriched for mutations. This opens up a whole new window for this type of research.
Snyder, who leads Stanford’s genetics department and directs the Stanford Center for Genomics and Personalized Medicine, likens looking for cancer-causing mutations only in coding regions as "looking under the lamppost" for keys lost at night. Until recently, the coding regions of genes were the most well-studied, and unexpected mutations stood out like a sore thumb. We've known there's a lot more of the genome outside the coding regions, but until the ENCODE project was largely completed in 2012, researchers were often in the dark as to where, or even how, they should look.
The researchers in the current study compared the genome sequences of eight types of cancers from 436 people, focusing on the regulatory regions as defined by ENCODE as well as the coding sequences of the genes they control. Some regulatory mutation hotspots were already known, some were found in the regulatory regions of genes known to be associated with cancer, and some appear to control the expression of genes not previously associated with cancer. In many cases, however, the mutations seemed to affect the ability of proteins called transcription factors to bind to DNA and trigger or suppress the expression of nearby genes.
It makes sense that tinkering with a gene's regulation might be a more effective way to subvert the rules in the long run than changing the coding sequence itself. As Synder told me, "Some genes you probably can’t mutate without killing their ability to make protein at all. In contrast, a regulatory mutation stands a better chance of producing a functional, but mis-regulated protein. And cancers are all about mis-regulation."
Previously: Scientists announce the completion of the ENCODE project, a massive genome encyclopedia, Flies, worms and humans – and the modENCODE Project and Gene regulation controls identity – and health
Photo by Steve Calcott