Figure 1.

Existing methods for context-specific model extraction can be classified in three families: GIMME-, iMAT-, and MBA-like. This classification is based on the mathematical objective followed by the methods, i.e., similarity and operability, similarity or consistency. Moreover, methods can perform a model extraction and a flux prediction (GIMME- and iMAT-like families) or only a model extraction (MBA-like family). Although similar in formulation to INIT, the metabolic task constraint imposed by tINIT renders this algorithm close to the objective of the GIMME-like family (see main text).