Just enter your email
Hi Pedro et al.,
Here some feedback, I have tried to write this as I would write a "normal" review, maybe a bit less formal. I tried to be critical and constructive.
The paper is written in a very concise manner and I think it would be good to give a little more background to the study. For instant give a motivation for the study and phrase a scientific question that you address in the paper.
I would say that it is an important paper, mostly because so few accurate models and data are available for the pathway. In my opinion the model evaluation is not very strong, see below for details.
some more specific comments:
In the material and methods section "proteomics", some results are given and discussed, this seems not appropriate for the methods section.
It would have been good to show some of the enzyme kinetic characterisations. There are only few studies published on kinetics of the PPP pathway, (mostly for the NADPH coupled enzymes) and it would be good to show some of the saturation curves and give some more detail on how the rate equations were derived and how parameter values were obtained. Were all the parameters identifiable in the data sets?
I would not stress the goodness of fit and the small number of parameters that were fitted to obtain Fig. 2. In fact, the enzyme concentration (or kcat), for each of the reactions that react with the measured variables were fitted. In addition the k value for the NADPH oxidase of 1 /s, which most likely has a large effect on the NADPH concentration is not discussed at all in the manuscript.
In the MCA section it is stated that these four enzymes completely control the flux into the PPP, so using them as fitting parameters has large consequences.
To me this is not a model prediction, I suggest to not use that term for the model simulations in Fig. 2.
A value of 1 is chosen for the k NADPH oxidase and for the E4P and E5P sink reactions, which I assume is a guess. I know that these reactions are grouped reactions and that it is not possible to measure such a k value. However I would think that one might be able to estimate a value on the basis of flux estimations in genome scale model, i.e. estimate a total NADPH oxidase, E4P, and E5P flux in growing yeast cells.
In the last paragraph of the "oxidative stress" section it is stated that the flux ratio of PPP over glycolysis is consistent with experimentally measured flux ratios. But the model gives a 1/19 ratio and experimentally a 1/10 ratio was observed. Actually the model values for the oxidised conditions 1/10 seem to agree with the experimentally measured value for the non-oxidised state.
It would be good to give also the absolute flux values for the PPP pathway and glycolysis, both for the model and for experiment (if known).
We have made the models associated with the manuscript available for simulation in your browser via JWS Online with the following link:
The link brings you to a webpage where you can chose between four models each corresponding to a specific Figure or Table in the manuscript, (as indicated).
JWS Online simulations of MODEL1311290000.xml and MODEL1311290001.xml give identical results as obtained with Copasi.
Most of the results in the manuscript were reproducible, except for the simulation shown in Fig. 2. When we simulated the SBML model: MODEL1311290000.xml in Copasi we obtained a different result from what is displayed in the manuscript. For instance the NADPH concentration in the simulation peaks at 0.25 while in the manuscript the value is close to 0.22.