Reply to "Challenges in modeling the human gut microbiome"

VOLUME 36 NUMBER 8 AUGUST 2018 NATURE BIOTECHNOLOGY for growth of the AGORA models show a similar pattern to that of the reaction similarities (Supplementary Fig. 11). Use of the same set of growth-required metabolites is to be expected considering the high degree of similarity of the models. In conclusion, our analysis shows that the AGORA models can be divided into two groups based on their in silico growth rates. The first group consists of models that predict growth rates in biologically reasonable ranges and the second group consists of models that predict very high growth rates. The reason behind this behavior is the bounds used on the exchange reactions. Exchange reactions are important to the value of the growth rate because a large part of biomass metabolites is directly taken from medium and consumed by the biomass reaction. Although the models in the first group are exposed to very low bounds on their exchange reactions, the models in the second group are not constrained at all. One way to overcome this problem is to divide the growth rate by the total carbon influx, but from our analysis, we find that this resulted in very low biomass yields. Therefore, it is necessary to carefully curate these models before they are used for simulations of growth. It is a major effort to manually curate GEMs, but from our analysis of the model similarity, it is also clear that one may not need to actually curate all the 773 GEMs of AGORA, as one could select representative species from each of the functionally similar groups. It should, however, be mentioned that this similarity analysis may to some extent be driven by the fact that only nine different biomass reactions are used for the AGORA models. Thus, the similarity in biomass reactions have affected the gap-filling and curation process of the AGORA models and may therefore have resulted in the clustering of the models that we present here. To ensure progress in the field of modeling of gut microbiota using GEMs, we recommend performing a more detailed analysis of individual species with the aim of further expanding the set of biomass reactions, as well as manual curation of the GEMs with the objective of improving their oxygen and ATP metabolism. Considering the scale of such an effort, this will have to be done at the community level, where clear standards are being defined for the quality of individual models.

[1]  Tao Chen,et al.  Collaborative regulation of CO2 transport and fixation during succinate production in Escherichia coli , 2015, Scientific Reports.

[2]  Ines Thiele,et al.  Computationally efficient flux variability analysis , 2010, BMC Bioinformatics.

[3]  Jason A. Papin,et al.  Applications of genome-scale metabolic reconstructions , 2009, Molecular systems biology.

[4]  Florian Jarre,et al.  CycleFreeFlux: efficient removal of thermodynamically infeasible loops from flux distributions , 2015, Bioinform..

[5]  Ronan M. T. Fleming,et al.  Generation of genome-scale metabolic reconstructions for 773 members of the human gut microbiota , 2016, Nature Biotechnology.

[6]  H. Qian,et al.  Energy balance for analysis of complex metabolic networks. , 2002, Biophysical journal.

[7]  Ronan M. T. Fleming,et al.  Consistent Estimation of Gibbs Energy Using Component Contributions , 2013, PLoS Comput. Biol..

[8]  Tom M. Conrad,et al.  Omic data from evolved E. coli are consistent with computed optimal growth from genome-scale models , 2010, Molecular systems biology.

[9]  Karthik Raman,et al.  Critical assessment of genome-scale metabolic networks: the need for a unified standard , 2015, Briefings Bioinform..

[10]  Ronan M. T. Fleming,et al.  von Bertalanffy 1.0: a COBRA toolbox extension to thermodynamically constrain metabolic models , 2011, Bioinform..

[11]  W. Chesbro,et al.  Very slow growth of Escherichia coli , 1979, Journal of bacteriology.

[12]  Ronan M. T. Fleming,et al.  A variational principle for computing nonequilibrium fluxes and potentials in genome-scale biochemical networks. , 2011, Journal of theoretical biology.

[13]  Jeffrey D Orth,et al.  What is flux balance analysis? , 2010, Nature Biotechnology.

[14]  D. Fell,et al.  A general definition of metabolic pathways useful for systematic organization and analysis of complex metabolic networks , 2000, Nature Biotechnology.

[15]  B. Palsson,et al.  Elimination of thermodynamically infeasible loops in steady-state metabolic models. , 2011, Biophysical journal.

[16]  B. Palsson,et al.  A protocol for generating a high-quality genome-scale metabolic reconstruction , 2010 .

[17]  Bernhard O. Palsson,et al.  Optimizing genome-scale network reconstructions , 2014, Nature Biotechnology.

[18]  Jens Nielsen,et al.  Challenges in modeling the human gut microbiome , 2018, Nature Biotechnology.

[19]  Philip Miller,et al.  BiGG Models: A platform for integrating, standardizing and sharing genome-scale models , 2015, Nucleic Acids Res..

[20]  Jason A. Papin,et al.  Integration of expression data in genome-scale metabolic network reconstructions , 2012, Front. Physio..

[21]  Jason A. Papin,et al.  Comparison of network-based pathway analysis methods. , 2004, Trends in biotechnology.

[22]  Ronan M. T. Fleming,et al.  fastGapFill: efficient gap filling in metabolic networks , 2014, Bioinform..

[23]  B. Palsson,et al.  Systems approach to refining genome annotation , 2006, Proceedings of the National Academy of Sciences.

[24]  A. Wolfe.  The Acetate Switch , 2005, Microbiology and Molecular Biology Reviews.

[25]  Ronan M. T. Fleming,et al.  Conditions for duality between fluxes and concentrations in biochemical networks. , 2015, Journal of theoretical biology.

[26]  Vinay Satish Kumar,et al.  Optimization based automated curation of metabolic reconstructions , 2007, BMC Bioinformatics.