Getz uncovers exciting new trends in cancer genomes
By Jeffrey Stumpf
Gad Getz, Ph.D., provided a glimpse of the exciting new world of identifying and treating cancer, in his talk Oct. 5 at NIEHS. As director of Cancer Genome Computational Analysis at the Broad Institute of Harvard and the Massachusetts Institute of Technology, Getz discussed the difficulties and potential of analyzing genome and exon sequencing of around 6,000 tumor samples.
Getz’s research conducts an orchestra of various disciplines, including clinicians, computational biologists, software engineers, statisticians, data analysts, and project managers, all with the common goal that Getz stated clearly.
“We want to kill cancer,” Getz boldly said.
Teaming up with The Cancer Genome Atlas, (http://cancergenome.nih.gov/) Getz’s group wants to understand the genetics of tumor development. Using the vast array of DNA sequence data, Getz outlined a two-step approach.
- Characterize genes and pathways that are different between normal and tumor tissues.
- Interpret those differences in the context of a large population of cancer patients to find out if these mutations are occurring by chance.
Mutations along for the ride
Within a quagmire of mutations and genome rearrangements in cancer cells, it is important to know what mutations are necessary and sufficient for development of a tumor. Getz said that his dream would be to get the genome sequence of every cell in a tumor.
“It is kind of like archaeology, in that you have to take the status today and infer what was in the past,” Getz explained.
Mutations that drive cells to tumor formation, or drivers, need to be distinguished from so-called passenger mutations that coincide with the driver mutations during tumorigenesis. Getz searched for mutations that are common among tumors and identified mutations in 428 genes.
Some genes, such as those that encode olfactory receptors, are unlikely to be driver mutations. So, Getz used an algorithm to account for different mutation rates that occur at different regions of the genome. For instance, mutation rates increase in regions that replicate in the late part of the S-phase and are reduced in genes with higher gene expression. Accounting for heterogeneous mutation rates reduced the number of driver mutations to 13 and uncovered new pathways, such as RNA splicing, chromatin modeling, and ubiquitination, as possible pathways in driving tumor formation.
A flood of data
Getz explained that the cost of sequencing one million bases has declined from $30,000 to 10 cents in a little over a decade, and this drop in cost has allowed sequencing of protein-coding regions to identify thousands of pairs of tumor and adjacent normal cells. The ambitious sequencing project has caused a problem that most scientists would love to have — too much data.
Regardless, Getz envisions a future of personalized medicine where genome sequencing occurs routinely at birth as a reference point for later tumor formations. Providing a secure computational infrastructure will be the next challenge.
“Medicine, in general, is becoming an information-rich discipline. We need to build a system that integrates all this information for doctors to make the next choice in treatment,” said Getz.
(Jeffrey Stumpf, Ph.D., is a research fellow in the NIEHS Laboratory of Molecular Genetics Mitochondrial DNA Replication Group and a frequent contributor to the Environmental Factor.)