A comparison between predicted amino-acid biosensor designs and a list of known amino-acid auxotrophic strains.

For the first 10 amino-acids (rows in the table), the predicted knockouts match exactly those of the known auxotrophic strains. For the next 5, the known knockouts are correctly predicted as causing auxotrophy (when simulated via FBA in the model), but not to ultra-auxotrophy, while the biosensor design method yielded different knockouts that are predicted to lead to the desired ultra-auxotrophy. For the last 3 amino-acids in the table, the predicted and known knockouts differ.