Chemical reaction network theory analysis
Repository: RuleWorld/PyBioNetGen
Title: Add chemical reaction network theory analysis for generated BNGL networks
Motivation
BNGsim already computes conservation laws from the stoichiometry matrix, so conservation-law detection itself is not missing.
However, generated BioNetGen networks could support a broader set of structural chemical-reaction-network analyses, including:
- reaction complexes;
- linkage classes;
- strong linkage classes;
- weak reversibility;
- stoichiometric rank;
- deficiency;
- terminal strong linkage classes;
- network decomposition.
These properties can reveal qualitative model structure without simulation and are useful for understanding multistability, complex balancing, and network organization.
Proposed API
analysis = model.crnt()
analysis.stoichiometric_rank
analysis.deficiency
analysis.linkage_classes
analysis.weakly_reversible
analysis.conservation_laws
For rule-based models, this analysis would apply to the finite generated network and should state that explicitly.
Implementation
This could reuse:
- BioNetGen's generated species/reactions;
- BNGsim's stoichiometry representation where available;
- existing conservation-law information rather than recomputing it separately.
The first version does not need to implement every theorem from CRNT. Exposing a reliable structural representation and the common network invariants would already be useful.
Chemical reaction network theory analysis
Repository:
RuleWorld/PyBioNetGenTitle: Add chemical reaction network theory analysis for generated BNGL networks
Motivation
BNGsim already computes conservation laws from the stoichiometry matrix, so conservation-law detection itself is not missing.
However, generated BioNetGen networks could support a broader set of structural chemical-reaction-network analyses, including:
These properties can reveal qualitative model structure without simulation and are useful for understanding multistability, complex balancing, and network organization.
Proposed API
For rule-based models, this analysis would apply to the finite generated network and should state that explicitly.
Implementation
This could reuse:
The first version does not need to implement every theorem from CRNT. Exposing a reliable structural representation and the common network invariants would already be useful.