Ageing is intrinsically complex, being driven by multiple causal mechanisms. Each mechanism tends to be partially supported by data indicating that it has a role in the overall cellular and molecular pathways underlying the ageing process. However, the magnitude of this role is usually modest. The systems biology approach combines (i) data-driven modelling, often using the large volumes of data generated by functional genomics technologies, and (ii) hypothesis-driven experimental studies to investigate causal pathways and identify their parameter values in an unusually quantitative manner, which enables the contributions of individual mechanisms and their interactions to be better understood, and allows for the design of experiments explicitly to test the complex predictions arising from such models. A clear example of the success of the systems biology approach in unravelling the complexity of ageing can be seen in recent studies on cell replicative senescence, revealing interactions between mitochondrial dysfunction, telomere erosion and DNA damage. An important challenge also exists in connecting the network of (random) damage-driven proximate mechanisms of ageing with the higher level (genetically specified) signalling pathways that influence longevity. This connection is informed by actions of natural selection on the determinants of ageing and longevity.