On the personalised modelling of cancer signalling*

Dynamic modelling has long been used to understand fundamental principles of cell signalling and its dysregulation in cancer. More recently these models have also been used to understand the individual risks of cancer patients, and predict their survival probabilities. However, the current methodologies for integrating tumour data and generating patient-specific simulations suffer from the lack of general applicability; they only work for cell signalling models in which only posttranslational protein modifications are considered, so that the total protein concentrations are conserved. Here, we present novel, generally applicable method. The method is based on a simple theoretical framework for modelling gene-regulation, and the indirect estimation of patient-specific parameters from tumour data. Because our method does not require time-invariance of the total-protein concentrations, it can be applied to models of any nature, including the many cancer signalling models involving gene-regulation.