Science at the Shine Dome 2010
Symposium: Genomics and mathematics
Friday, 7 May 2009
Professor Simon Tavaré
Simon Tavaré has worked for many years on statistical and computational problems arising in molecular biology, human and population genetics, molecular evolution, bioinformatics and computational biology. He is a professor in the Department of Applied Mathematics and Theoretical Physics and Professor of Cancer Research (Bioinformatics) in the Oncology Department at the University of Cambridge. He is also a senior group leader in the Cancer Research UK Cambridge Research Institute. His group there focuses mainly on cancer genomics and evolutionary approaches to cancer, the themes of his talk.
In 2009 Simon was elected a Fellow of the Academy of Medical Sciences. He is also a research professor, and George and Louise Kawamoto Chair in Biological Sciences, at the University of Southern California. He is principal investigator of USC’s NIH Center of Excellence in Genomic Science, which is developing computational and experimental approaches for understanding how genotype relates to phenotype.
Combining genomics and mathematics to learn about cancer
It is difficult to follow the evolution of dividing cells in a human tumour over large numbers of divisions. Instead of using direct measurements, it should be possible to infer features of this evolution from indirect measurements such as patterns of molecular variation (eg, single nucleotide polymorphisms or CpG island methylation) measured in the cells. In this talk I will outline the experimental and mathematical underpinnings of our approach to inference about tumour evolution. ‘Genomics’ in my title refers to a novel high-throughput technology for obtaining molecular data from single cells, and ‘mathematics’ refers to a physically based model for tumour growth (a cellular Potts model that allows for cell proliferation, differentiation, migration, invasion and cell signalling) and the statistical inference method that connects the dynamics of the model to the variation data. This last step uses a Bayesian inference method known as Approximate Bayesian Computation, a likelihood-free method that has broad applicability in the analysis of complex stochastic processes.
