Fitting the LCMV CD8+ Data Based on Measurement and Process Noise Models Mathematical modeling of immunological data can be a very powerful tool for understanding the immune system dynamics. Although it is reasonable to assume that the nature of the inter-cellular interactions is stochastic, usually, deterministic models are fitted to experimental data. We analyze the experimental data of lymphocytic-chorimeningitis virus (LCMV) infected mice. To explain the stochastic variation of the data, we use different stochastic measurement error models and we find that the intensity of the errors is quite high regardless of the model. Using the Gillespie simulation, we are able to show that the intensity of the process noise is not negligible in comparison to the variation seen in the experimental data. This suggests that the process noise of the underlying dynamics must be included in parameter estimation procedures, and considered separately from stochastic measurement errors. This work is a part of a larger research project leading to the system theory approach for modeling immunological data and corresponding sources of uncertainties. This work is developed in collaboration with Prof. Rob De Boer.
| |

Copyright
© University of Notre DameLast Updated: Friday, November 4, 2005 |