Forecasting in cancer cell simulations
Critical to the development of new cancer treatments is understanding the mechanisms underlying the emergence of cells resistant to different drug treatments, as well as discovering synergistic combinations of drugs that reduce the chance of this event. The emergence of drug resistance is determined by complex molecular mechanisms that ultimately lead to a reduction of the effectiveness of a particular drug, as well as various types of dynamic processes concerning large populations of cells. Multicellular system's dynamics such as tumour growth and evolution can only be understood by studying how the individual cells grow, divide and die, and their interactions at the population level.
To that end, in-silico simulation of tumour growth and evolution are employed using multi-scale models (MSM), bridging the gap between different levels of description, and connecting events that occur at different scales. MSM simulations represent a powerful approach to test alternative hypotheses about phenomena observed in cancer, enabling the prioritization of optimal drug treatments. One such tool is PhysiBoSS, that merges the agent-based simulator PhysiCell with the stochastic Boolean model simulator MaBoSS.
The study of complex biological systems relies on the execution of heuristic methods requiring very large numbers of simulations. Applying model exploration to MSM involves an iterative workflow, where simulations are run across a high dimensional parameter space. Complex event forecasting techniques may be employed to improve parameter exploration. In particular, our goal is to forecast online the simulations of treatments that do not reduce or stabilize tumour growth, in order to stop them, thus freeing high-performance computing resources for new simulations.
Alevizos E., Artikis A. and Paliouras G., Event Forecasting with Pattern Markov Chains.
In 11th International Conference on Distributed and Event-Based Systems (DEBS), pp. 146–157, 2017. PDF Slides BibTeX DOI