Personalised medicine is one of the challenges of systems biology. If some clinical trials have already explored the possibility to treat patients individually based on their genomic profiles, there are more and more attempts to use mathematical models as an aid to select therapies in the clinics. The purpose of these models is to describe, understand and predict cellular responses from a molecular perspective. They serve to explain complex biological phenomena, provide predictions that can be tested experimentally, and formulate plausible scenarios of a complex biological behaviour when intuition is not sufficient anymore.
The process of building these models ideally starts with the analyses of omics data and the gathering of prior information (from published experiments and databases) in the form of a network. The network is then translated into a mathematical model with the appropriate formalism that will depend on the initial biological question and the type of networks (chemical kinetics, logical approach, etc.). The model then aims at predicting and anticipating the effect of a perturbation, either from an intrinsic point of view (e.g., mutations) or from an extrinsic point of view (e.g. drug treatment). I will show some examples of such analyses with logical models of signalling pathways which are known to be altered in cancer.