Skip to main content

Advertisement

Table 4 Fitness functions used

From: Designing minimal microbial strains of desired functionality using a genetic algorithm

i Design objective Fitness function \(F_i\)
1 Ethanol production with minimal MCS size \(w_{1} \min Y_{Etoh} + w_{3} (1 - \vert C \vert /n)\)
2 Substrate specific productivity with minimal MCS size \(w_{2} \min \eta _{Etoh} + w_{3} (1 - \vert C \vert / n)\)
3 Growth coupled product yield with minimal MCS size and maximum number of surviving modes \(w_{1} \min Y_{Etoh} \times w_{2} \max \eta _{Etoh} + w_{3} (1 - \vert C \vert / n) + w_{4} \vert \mathbf D ^C \vert / \vert \mathbf E \vert\)
  1. Fitness functions used, where, \(w_{1}\), \(w_{2}\), \(w_{3}\) and \(w_{4}\) are weights associated with ethanol yield (\(Y_{Etoh}\)), ethanol efficiency (\(\eta _{Etoh}\)), MCS cardinality (\(\vert C \vert\)) and number of surviving modes (\(\vert \mathbf D ^C \vert\)) respectively. These weights are used primarily to ensure desired contribution of the different variables towards the fitness function. They can also be used to give higher preference to a particular variable. C is the MCS, n the total number of reactions and E the set of all EFMs in a network. All fitness functions were maximised