Multi-Target Analysis and Design of Mitochondrial Metabolism
Fig 1
The BioCAD framework is able to perform multi-objective optimization on gene sets and protein abundances.
The optimization can be applied to the Boolean arrays of gene sets (simulating an on/off condition), to the real-valued fluxes, or to the real-valued arrays representing protein abundances. In both cases, we seek for the Pareto-optimal arrays to simultaneously optimize two or more objective functions. The optimization is augmented with sensitivity, identifiability, robustness and ϵ-dominance analysis. The sensitivity analysis quantifies the importance of the input variables in the model, the identifiability analysis infers functional relations between them, the robustness is used in combination with the sensitivity and quantifies if a solution is reachable even if small perturbations are applied to the system, while the ϵ-dominance analysis identifies sub-optimal points.