| (1) Obtain all possible reactions |
| How? From database sources, a set of all possible reactions must be obtained. |
| Difficulties: Databases may not contain all the possible reactions involved. Biological knowledge is required. |
| (2) Collect experimental data |
| How? Experimental data can be obtained from the literature for similar systems or from laboratory experiments. |
| Difficulties: Hard to found literature data, or reported data very scarce. Laboratory experiments are |
| time and money consuming. |
| (3) Obtain the reactions that seems to occur |
| How? From experimental data, apply topological Analysis (TA) and Flux Balance Analysis (FBA), which demands |
| optimization but no parameters estimation. |
| Difficulties: The obtained results from the techniques TA and FBA above are not unique. Specific biological knowledge about |
| the system is required. |
| (4) Propose kinectic models |
| How? Typically, it is proposed expressions similar to Michaellis Menten kinetics or law of mass actions. Databases with typical |
| kinetic expressions can be useful. |
| Difficulties: Databases can not contain reactions involved. The modfications on Michalis Menten expressions can lead to |
| combinatorial explosion of possible models. |
| (5) Obtain initial parameters estimates |
| How? In the same databases from the previous step or from literature review for similar systems, an initial parameters |
| estimation can be obtained. |
| Difficulties: Here is one of the most difficult steps. In some cases, one can have no idea of possible parameters values. |
| (6) Employ parameters identification technique |
| Objective? Verify the most influence parameters. |
| Difficulties: Implement such methodologies. Also, different techniques can result in non unique results. Many can be very |
| influenced by the initial parameters estimation. |
| (7) Evaluate the results |
| If the results are satisfactory, then stop the procedure. Otherwise, perform Optimal Experimental Design. |
| Difficulties: It is very difficult to estabilish when is desirable to stop. Specially because the satisfactory parameters |
| uncerainties evaluation will demand an unfeasible number of experiments. |
| (8) Employ optimal experimental design |
| Objective? Obtain experimental regions for better parameters estimation or models discrimination. Return to Step (2). |
| Difficulties: Implement a methodology described in the literature. |
