Figure 3. MTA workflow.
MTA performs knockouts for each of the reactions in the metabolic model at the AD state (source), and assigns a score for each gene knockout reflecting the predicted extent by which this knockout may transform the metabolic state back to the healthy (target) state. Next, we aggregate the gene/reaction level predictions to identify the pathways whose knockout is predicted to be most successful in transforming the metabolic state as close as possible back to the healthy state.