Applications of sensitivity analysis to model calibration

Sensitivity analysis has important applications in model calibration.

One application of sensitivity analysis addresses the question of "What's important to model or system development?" One can seek to identify important connections between observations, model inputs, and predictions or forecasts. That is, one can seek to understand what observations (measurements of dependent variables) are most and least important to model inputs (parameters representing system characteristics or excitation), what model inputs are most and least important to predictions or forecasts, and what observations are most and least important to the predictions and forecasts. Often the results are surprising, lead to finding problems in the data or model development, and fixing the problems. This leads to better models.[1][2] In biomedical engineering, sensitivity analysis can be used to determine system dynamics in ODE-based kinetic models. Parameters corresponding to stages of differentiation can be varied to determine which parameter is most influential on cell fate. Therefore, the most limiting step can be identified and the cell state for most advantageous scale-up and expansion can be determined.[3] Additionally, complex networks in systems biology can be better understood through fitting mass-action kinetic models. Sensitivity analysis on rate coefficients can then be conducted to determine optimal therapeutic targets within the system of interest.[4]

References

  1. Hill, M.; Kavetski, D.; Clark, M.; Ye, M.; Arabi, M.; Lu, D.; Foglia, L.; Mehl, S. (2015). "Practical use of computationally frugal model analysis methods". Groundwater. 54 (2): 159–170. doi:10.1111/gwat.12330. OSTI 1286771. PMID 25810333.
  2. Hill, M.; Tiedeman, C. (2007). Effective Groundwater Model Calibration, with Analysis of Data, Sensitivities, Predictions, and Uncertainty. John Wiley & Sons.
  3. Selekman, JA; Das, A; Grundl, NJ; Palecek, SP (2013). "Improving efficiency of human pluripotent stem cell differentiation platforms using an integrated experimental and computational approach". Biotechnol Bioeng. 110 (11): 3024–37. doi:10.1002/bit.24968. PMC 3970199. PMID 23740478.
  4. Tian, D; Solodin, NM; Rajbhandari, P; Bjorklund, K; Alarid, ET; Kreeger, PK (2015). "A kinetic model identifies phosphorylated estrogen receptor-α (ERα) as a critical regulator of ERα dynamics in breast cancer". FASEB J. 29 (5): 2022–31. doi:10.1096/fj.14-265637. PMC 4415015. PMID 25648997.
This article is issued from Wikipedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.