Differences and similarities¶

RAVEN is available as two independent implementations:

RAVEN — the original MATLAB toolbox
raven-toolbox — the Python port, built on cobrapy

Both implement the same core methods for genome-scale metabolic modelling, but differ in language conventions, solver interface, and which features are available. See the function mapping table for a full side-by-side cross-reference.

Shared functionality¶

Homology-based reconstruction¶

Both toolboxes support building draft models by transferring reactions from template models based on protein sequence homology (BLAST+, DIAMOND, HMMER). The underlying algorithm and default parameters are consistent between implementations.

RAVEN (MATLAB)	raven-toolbox (Python)
`getModelFromHomology`	`get_model_from_homology`
`getBlast`	`get_blast`
`getDiamond`	`get_diamond`

KEGG-based reconstruction¶

Both support building draft metabolic models from KEGG using organism codes, and can download and cache KEGG data locally.

RAVEN (MATLAB)	raven-toolbox (Python)
`getKEGGModelForOrganism`	`get_kegg_model_for_organism`

Gap-filling¶

Both provide gap-filling by linear programming to identify the minimum reaction set needed to restore connectivity or growth.

RAVEN (MATLAB)	raven-toolbox (Python)
`fillGaps`	`fill_gaps`

Flux analysis¶

Flux balance analysis (FBA), gene knockouts, and flux variability analysis (FVA) are available in both, with equivalent solver interfaces.

Model import and export¶

Both can read and write SBML (.xml) and RAVEN's Excel format (.xlsx), making models fully portable between implementations.

RAVEN (MATLAB)	raven-toolbox (Python)
`importModel`	`import_model`
`exportModel`	`export_model`
`importExcelModel`	`import_excel_model`
`exportToExcelFormat`	`export_to_excel_format`

Python-only features¶

KEGG artefact generation¶

raven-toolbox includes tools to generate KEGG artefact databases — stoichiometric matrices and reaction lists for arbitrary KEGG organisms. This is used to produce updated KEGG reference data and is not available in the MATLAB toolbox.

cobrapy interoperability¶

raven-toolbox is built directly on cobrapy's cobra.Model, so any cobrapy- compatible tool works immediately. The MATLAB toolbox provides ravenCobraWrapper for conversion, but this is an explicit transformation step rather than native compatibility.

MATLAB-only features¶

COBRA Toolbox compatibility¶

ravenCobraWrapper converts between RAVEN and COBRA Toolbox model formats in both directions. raven-toolbox uses cobrapy natively and has no COBRA Toolbox dependency.

Visualisation¶

RAVEN's drawMap-family functions provide flux map visualisation on top of stoichiometric network maps. raven-toolbox currently has no built-in visualisation; use Escher or cobrapy-compatible tools instead.

Solver interface¶

	RAVEN (MATLAB)	raven-toolbox (Python)
Configuration	`setRavenSolver('gurobi')`	`cobra.Configuration().solver = 'gurobi'`
Recommended solver	Gurobi (free academic licence)	Gurobi (free academic licence)
Bundled open-source solver	GLPK (via COBRA Toolbox)	GLPK (bundled with cobrapy)

Naming conventions¶

MATLAB uses camelCase function names; Python uses snake_case. The conversion is mechanical: getModelFromHomology → get_model_from_homology. See the function mapping table for the full list.