As high throughput technologies continue to advance, various types of cancer genomics data quickly accumulate to an unprecedented scale. This includes large amount of tumor mutations, transcriptomics, epigenetics, proteomics, metabolomics data, and other more specific types such as signal cell or personalized cancer genomics information. The wealth of data ignites the hope to understand genetic mechanisms of cancer development and metastasis, and to identify driver mutations or drug targets for the treatment. However the pluralism of causes and effects makes the long expected understanding and cure of this disease most feasibly to be achieved through integrated and systematic research approaches. Such approaches typically reply on network representation considering all molecular components in a cell as nodes, and their direct or indirect functional relationships as edges, usually with weights assigned. The network representation allows observation of multiple components simultaneously, and by deploying mathematical models working on network topological structures and the dynamics, we could explore and validate hypotheses of cancer development mechanisms through in silico studies. In this talk, I will present several cancer genomics networks derived from The Cancer Genome Atlas (TCGA) database. Some examples are cancer gene co-expression and regulation networks, epigenetics networks, and their integrations, with highlight of the methodology and prominent results. I will discuss as well my work on bringing natural selective pressure into gene co-expression network analysis. Preliminary results using a bacterial model organism, E. coli, show that evolutionary persistence is the driving force of co-expression modules, and determines regulation pattern of co-expressed genes. Gene co-expression and co-evolution are highly coupled, and selective pressures act not on each node (gene or protein) individually, but on the edges, i.e. biological relationships between genes. I also applied evolutionary concept to the comparisons between naturally evolved biological networks and man-made systems. Specifically, the gene regulatory network of E. coli has been compared to the call graph network of the Linux operating system. This comparison has provided us clues as to why the naturally evolved network is more robust then artificial system, which underlying the importance of focusing on cancer network topologies in the fight against this disease.
Biography
Dr. Gang Fang is an assistant professor of Biology at NYU Shanghai. He is also affiliate assistant professor at the Department of Biology and Center for Genomics and Systems Biology at NYU's campus in New York City. Prior to joining NYU Shanghai in 2015, he was an associate research scientist at Yale University. His PhD degree is from Institute Pasteur, Paris of France. Dr. Fang’s research interests are evolution and comparative genomics, and biological network analysis. He has developed the concept of gene evolutionary persistence and employed this concept in the studies of genome organization, proteome evolution, and transcriptome and biology networks. His papers have been published in Molecular Biology and Evolution, Genome Research, Genome Biology, PLoS Computational Biology, PNAS, BMC genomics, Nature, Nature Genetics Reviews, and Trends in Genetics, among others. Fang’s current work focuses on large-scale cancer genomics data.