Posted: May 20, 2011

Researchers Use TCGA Data to Develop an Approach for Identifying Important Gene Alterations in Glioblastoma Multiforme

Catherine Evans

Researchers in the field of cancer genomics are exploring new ways to sort through the mountains of data generated by The Cancer Genome Atlas (TCGA) on gene changes from cancers like glioblastoma multiforme (GBM). GBM is a lethal adult brain cancer with mean survival time of little more than a year. In the September 2010 issue of PLoS Genetics, a research group describes their approach to identifying gene alterations that are especially relevant to GBM development.

The researchers developed a statistical approach to examine both the germline and somatic genomes of GBM patients. The germline genome is the full set of chromosomes we inherit from our parents. The somatic genome is the full set of DNA that acquires changes through our lifetime and that will not be passed on to our offspring. Researchers can compare a person’s normal tissue to their tumor tissue to determine which changes are specific to the tumor (somatic) and which were inherited (germline). 

Single Nucleotide Polymorphisms May Provide Clues for Finding Altered GBM Genes

The scientists were interested in looking at single nucleotide polymorphisms (SNPs) in GBM patients’ germline DNA. SNPs are single base pair changes in a sequence of DNA (for example, a change from an A to a G), and they occur commonly throughout the population. Most SNPs have no known biological effect. However, some SNPs are associated with an increased risk in certain diseases, including cancer. This means that people with specific SNPs may be more likely to develop the disease than individuals with other variations.

The researchers thought that the SNPs that are likely to increase risk for GBM might be found near genes that are known to be altered in these tumors. The current study aimed to develop a method to find those SNPs that may have a functional role in GBM development. They used TCGA’s data from patients’ normal tissue and GBM tumors to perform their analyses.

They statistically determined which SNPs in the TCGA germline data (from their normal tissue) were likely to play a role in gene alterations in GBM. From a list of 44,132 SNPs, they identified 139 SNPs that appeared significant. They found that many of these SNPs were in areas of the genome that coded for genes involved in cell growth and communication between brain cells. This finding indicated that they were on the right path to honing in on SNPs that might be involved with GBM development. 

A Target Gene is Identified

The group next tried to determine if any of their candidate SNPs were in altered genes whose levels of expression actually increased in the GBM tumors. One target was immediately clear. DOCK4 is a gene that is commonly amplified, meaning that several more copies of DOCK4 than are usually observed exist in some tumors. Of the patients whose DOCK4 gene was amplified, most also had an amplified version of a SNP within the gene. Patients who had this amplified SNP version within the amplified DOCK4 gene also had higher expression levels of DOCK4 in their tumors. The same pattern occurred in patients who had an amplified version of a SNP in another cancer-associated gene, EGFR. These results give statistical support to the notion that the presence of certain risky SNPs, which are inherited, indicates the genes they are in may be more likely to acquire alterations that lead to cancer.

The authors say that their results call for further study of the EGFR and DOCK4 genes and their SNPs. They also suggest that their method for identifying SNPs involved in tumor growth is a potential way to validate and identify important genes involved in other cancers. According to the authors, this approach is a useful way to make sense of large amounts of genomic data generated from large-scale projects like TCGA. The method may also identify subgroups of people who are at risk for developing a certain type of cancer.

LaFramboise, T., Dewal, N., Wilkins, K., Pe'er, I. and Freedman, M.L. (2010) Allelic selection of amplicons in glioblastoma revealed by combining somatic and germline analysis. PLoS Genet. 6(9):e1001086. Read the full article.