Inverse Problems in Systems Biology

Biology lacks the simple elegant laws of physics. Instead we have to work out the rules of life from data. We are developing tools that allow us to do so from a wide range of molecular, cellular, and phenotypic data.

We have made advances in learning the structure and dynamics of complex gene regulation networks; we are working on Bayesian methods that allow us to parameterise large mathematical models of biological systems; and we develop new statistical and computational approaches that will allow us to penetrate deeper into the dynamics and design principles of cellular systems. For us inference relies strongly on our ability to capture our uncertainty fairly and correctly. We use our inference approaches to design better experiments, and to choose from different hypotheses or models to determine which one offers the best mechanistic representation for a real-world biological system.

Network inference, parameter estimation, model selection, and robustness analysis have all become essential parts of our toolset. They allow us to create better models of biological systems, but they also allow us to work out how much we can learn from biological data.