Biomass equations should be irreversible

[mattb112885]

Existing biomass equations in the modelSEED database (which were imported from existing metabolic models) are currently labeled as reversible but should if they are to be left in the database of possible reactions, they should be irreversible (otherwise it is possible to trivially include a biomass equation from some organism in the reverse direction, and create almost all biomass components needed for any organism!).

This might not be a problem if all reactions in the database are mass balanced (and unbalanced reactions are thrown out) before making a model.

Example: rxn12821