In their correspondence relating to our review article in Nature Reviews Microbiology, Gruss et al. are surprised that we failed to mention respiration as an approach to improve on some desired properties of LAB, such as biomass yield and shelf life. It is clear that this approach has been important and is industrially relevant. However, the focus of this review was not to describe the full metabolic potential of LAB, nor all methods that have been used to improve desired properties of LAB — in contrast to what is suggested by Gruss et al., we were asked to specifically focus on modelling strategies, that is, on how modelling can help in suggesting ways to improve industrially relevant traits. We have therefore described the desired traits, outlined the modelling strategies used in biotechnology, and subsequently mapped these modelling technologies onto the desired properties that we identified.

We can add that we have recently introduced respiration in our Lactobacillus plantarum model (which required not only heme, but also menaquinone in the medium) and it did indeed increase biomass yield, as expected (unpublished results). If we had obtained this result before the experiments described in the 12 additional references, respiration would indeed have been a good example where a prediction derived from a model had suggested a way to improve industrially relevant properties.