Chen-Hsiang Yeang (Institute of Statistical Science, Academia Sinica)
Abstract. Cancer cells share certain common features with organisms in terms of evolution. On the one hand, both cancer and germ cells undergo molecular aberrations on DNAs and can pass these variations to their descendents. On the other hand, natural selection exerts pressure on both cancer cells and individual organisms and thus shapes their genotypes and phenotypes in a specific manner. Therefore, tools and discoveries in molecular evolution may help understanding the evolutionary processes of cancer and designing treatment strategies accordingly.
In this talk, I will cover my past and ongoing work in studying cancer genomes and network evolution. In cancer genomics, I will describe a data integration algorithm that associates molecular aberrations on DNAs (sequence mutations, copy number variations, DNA methylations, etc.) with molecular phenotypes of gene expressions. Using this algorithm we characterized molecular aberration patterns in glioblastoma and identified the gender and ethnic specific molecular signatures for prognosis in lung adenocarcinoma. In network evolution, I will describe a belief propagation that reconstructs the evolutionary history of domain-protein-reaction (DPR) networks. In a metabolic system, a DPR network specifies the domain architectures and catalytic functions of all enzymes. By applying this algorithm to the DPR networks across 12 species, we found that prokaryotes and eukaryotes followed quite distinct evolutionary patterns. Finally, I will discuss possibilities where cancer research and treatments can benefit from knowledge or tools in evolution.