Since I wasn't able to use the models distributed with the package (cf. #2), I tried to use another model but then discovered that the compounds in the thermodynamic database are all SEED identifiers. While this is not a problem, of course, using MetaNetX identifiers would be a great improvement as it easily maps to a whole range of databases.