Posted: August 31, 2011
Researchers use TCGA GBM data to develop a new data analysis tool
Catherine Evans
As The Cancer Genome Atlas (TCGA) continues to generate comprehensive datasets from the genomes of more than 20 different cancer types, researchers are scrambling to develop computer-based tools to manage and make sense of it all. Scientists’ ability to make sense of the data will be one of the most important challenges on the path to a completed cancer genome atlas. Several research groups have begun to undertake this problem using the publicly available data from TCGA’s completed characterization of glioblastoma multiforme (GBM). GBM is a highly lethal form of brain cancer, with median survival just over a year. In the July 1, 2011 issue of Cancer Research, Johns Hopkins University researchers reported a new data analysis program that they suggest can find important patterns in cancerous gene changes.
Cancer genomics research so far has shown that each person’s cancer, even of the same type, is genetically distinct. An important goal, therefore, is to identify groups of gene alterations that drive tumor development in each case. This approach will allow scientists to categorize tumors according to the particular combination of gene alterations they harbor. Patients could then be given the appropriate treatment and prognosis based on their specific tumor.
Program Finds Gene Mutations That Correspond to Altered Gene Expression
The authors of the current study developed an analysis method using TCGA data from 149 GBM samples. The method determines the degree to which mutated genes co-occur with greatly over- or under-expressed genes in the GBM tumor samples. This approach, they hope, can identify gene combinations that are important to tumor development and progression. The genes identified by the method may also represent potential drug targets. The analysis can also give insight into combinations of altered genes that affect the varying lethality of tumor subtypes. The program, according to the authors, is unbiased because it does not take into account pre-existing knowledge about the genes’ cell signaling networks.
Of the 307 mutated genes found by their analysis of the 149 GBM samples, 13 percent were associated with significant over- or under-expression of the 11,828 genes whose expression information was available. One group that stood out contained 38 genes whose over- or under-expression correlated with TP53 gene mutation. TP53 is a well-known tumor-suppressor in many cancers, including GBM. The analysis confirmed previous study findings by showing that in tumors with TP53 mutation, genes like PLK1, TCB1, and DBF4 are overexpressed.
Tool Reveals New Mutations and Relationships Among Mutated Genes in GBM
Importantly, the group’s analysis also revealed 12 overexpressed genes newly linked to TP53 mutation. These 12 genes are known to be overexpressed in various cancers, but this study is the first to link their overexpression to TP53 mutation in GBM. In previous studies, drugs blocking the activity of the overexpressed genes blunted cancer growth. The authors suggest that similar drugs could be used in cancers with mutated TP53.
The largest network of correlated mutation and expression in genes was that of the IDH1 gene. IDH1 has been found to be frequently mutated in GBM. It is thought to be a driver gene in GBM development. Therefore, the current study’s finding that IDH1 mutation was associated with the over- and under-expression of 1,001 genes is consistent with previous research. Among the overexpressed genes in this network are several known to be crucial to tumor survival, as well as many known oncogenes, or genes that are involved in tumor growth and progression. Many of the underexpressed genes were previously found to be tumor suppressors. The authors suggest this knowledge could lead to the development of drug cocktails, in which drugs inhibiting each of the overexpressed genes are combined into one treatment.Finally, the analysis found that the second largest network of correlated gene changes were those expression changes associated with the mutated SYNE1 gene. Mutations in SYNE1 occurred frequently in the GBM tumors used in the current study. However, SYNE1 mutations have not previously been reported in GBM. Mutated SYNE1 was associated with overexpression of several known oncogenes. The authors suggest that their finding of this large SYNE1 network points to its potentially large role in GBM. SYNE1 and its associated genes may be new targets for future treatments.
In sum, the researcher’s new method confirmed previous findings and pointed to potentially new genetic culprits in GBM. The group points to its usefulness in identifying groups of genes that cooperatively drive cancer. As TCGA continues to crank out genomics data, additional approaches like that of the Johns Hopkins group will be necessary to translate raw data into biological insights. These developments will take the cancer genomics community one step closer to understanding tumorigenesis and ultimately developing improved treatments for cancer patients.
Masica, D. and Karchin, K. (2011) Correlation of Somatic Mutation and Expression Identifies Genes Important in Human Glioblastoma Progression and Survival. Cancer Res. 71:4550-4561. Read the full article.
