Model-based identifiable parameter determination applied to a simultaneous saccharification and fermentation process...

Area of Science:

• Systems biology
• Biochemical engineering
• Computational modeling

Background:

• Biological reaction networks often involve over-parameterized, nonlinear models.
• Parameter identification in these systems presents ill-posed inverse problems, hindering accurate analysis.
• High parameter correlations complicate model development and prediction.

Purpose of the Study:

• To present a systematic methodology for model-based identifiable parameter determination (MBIPD).
• To address challenges in structure and parameter identification for nonlinear biological models.
• To reduce the parameter search space for reliable and efficient estimation.

Main Methods:

• The MBIPD methodology includes model selection, initial guess tracking, and iterative parameter estimation.
• An identifiable parameter subset selection (SsS) algorithm, based on sensitivity matrix analysis and rank revealing factorization, is employed.
• The approach was tested on the simultaneous saccharification and fermentation (SSF) process for bio-ethanol production.

Main Results:

• MBIPD successfully reduced the parameter space for the SSF process model.
• Cross-validation confirmed the model's ability to accurately predict experimental data despite parameter reduction.
• The methodology facilitated easy and efficient adaptation of the model to new process conditions.

Conclusions:

• The MBIPD methodology provides a robust approach for parameter identification in complex biological systems.
• It effectively tackles over-parameterization and ill-posed problems, leading to more manageable models.
• This method enhances model reliability and adaptability for applications like biofuel production.