Identifying cancer genetic network signature from integrative microarray analysis
Mathematical Systems Biology of Cancer
May 05, 2006 01:30 PM to 02:30 PM
Speakers:
Zhou, Jasmine
|
 |
Abstract: |
Microarray gene expression profiling has been widely applied to cancer research. The
commonly used analysis approach is to identify genes differentially expressed in cancer versus
normal tissues or those in different cancer subtypes. However, it is known that phenotypes are
determined not only by genes, but also by the underlying structure of genetic networks. Often, it is
the interaction of many genes that causes phenotypic differences. In this work, we develop graphbased
methods to integrate multiple microarray data sets for cancer-related network module
discovery. By transforming each microarray dataset into a co-expression network, we perform
comparative network analysis to extract subnetworks which occur frequently in cancer datasets but
not in other datasets. In this way, we identify network modules that characterize commonalities and
variances of different types of cancer. Finally, we focus on dense co-expression networks which
represent cancer-specific co-expression clusters, and assign putative transcriptional regulator to
them. |
|
|
