Abstract:

Computer simulation can be used to inform in vivo and in vitro experimentation, enabling rapid, low-cost hypothesis generation and directing experimental design in order to test those hypotheses. In this way, in silico models become a scientific instrument for investigation, and so should be developed to high standards, be carefully calibrated and their findings presented in such that they may be reproduced. Here, we outline a theoretical framework that supports developing simulations as scientific instruments, and we select cancer systems biology as an exemplar domain, with a particular focus on cellular signalling models. We consider the challenges of lack of data, incomplete knowledge and modelling in the context of a rapidly changing knowledge base. Our framework comprises a process to clearly separate scientific and engineering concerns in model and simulation development, and an argumentation approach to documenting models that provides a rigorous way of recording assumptions, knowledge gaps and the inclusion of new knowledge. We also propose the use of interactive, dynamic visualisation to enable the biological community to interact with cellular signalling models directly for experimental design. There is a recognised mismatch in scale between these cellular signalling models and the tissue structures that are affected by tumours, and bridging this gap requires substantial computational resource. We present concurrent programming and distributed simulation as a technology to link scales without losing important details through model simplification. Moreover, we discuss the particular value of combining this technology, interactive visualisation, argumentation and model separation to support the development of multi-scale models that represent biologically plausible cells arranged in biologically plausible structures that model cell behaviour, interactions and response to therapeutic interventions.

@article(SS-CDT12,
  author = "James Bown and Paul S. Andrews and Yusuf Deeni and Alexey Goltsov and Michael Idowu
           and Fiona A. C. Polack and Adam T. Sampson and Mark Shovman and Susan Stepney",
  title = "Engineering simulations for cancer systems biology",
  journal = "Current Drug Targets",
  volume = 13,
  number = 12,
  pages = "1560-1574",
  year = 2012
)