Predicting biological system objectives de novo from internal state measurements

Area of Science:

• Systems Biology
• Metabolic Engineering
• Computational Biology

Background:

• Optimization theory is crucial for understanding complex biological systems and metabolic engineering.
• Flux Balance Analysis (FBA) predicts optimal metabolic fluxes but requires a biologically meaningful objective function.
• Defining accurate objective functions remains a significant challenge in FBA applications.

Purpose of the Study:

• To introduce a novel method, Biological Objective Solution Search (BOSS), for inferring biological system objectives.
• To enable the discovery of unknown stoichiometric objectives in biological networks.
• To enhance the biological relevance and predictive power of metabolic network models.

Main Methods:

• BOSS infers objectives by defining a putative 'objective reaction' and integrating it into the network stoichiometry.
• Linear programming (LP) is used to maximize the putative objective reaction.
• Minimizes the difference between in silico flux distributions and experimental data (e.g., isotopomer labeling) for objective validation.

Main Results:

• BOSS successfully infers objective functions with previously unknown stoichiometry.
• The method was validated on the central metabolic network of Saccharomyces cerevisiae.
• BOSS extends the biological relevance of objective inference beyond existing methods.

Conclusions:

• BOSS provides insights into the functional organization of biochemical networks.
• The method facilitates the interrogation of cellular design principles and aids in cellular engineering.
• For yeast metabolism, cellular growth was identified as the best-fit objective function based on experimental flux data.