Additional file 1 of Predicting microbial interactions with approaches based on flux balance analysis: an evaluation

Additional file 1: Table S1. Overview of statistical values for growth rates predicted in monocultures. H: time step of 1 per hour, H/10: time step of 0.1 per hour. Table S2. Mean absolute difference per species across the four methods and for the different media. Table S3. Global Spearman correlation refers to the correlation across all methods and settings. Significant p values are highlighted in grey. GR: Interaction strength computed as the ratio of growth rates, H: time step of 1 per hour, H/10: time step of 0.1 per hour, MX: interaction strength computed as the ratio of maximal abundances, Pars: Parsimonious FBA. Table S4. Correlation between predicted and measured interaction strengths for each species, with those that have refined metabolic reconstructions highlighted in grey. Interaction strength is quantified by comparing the growth rate or maximal biomass in co- and monoculture. Table S5. Wilcoxon test value for monocultures and cocultures with refined versus AGORA Models. Table S6. Products returned from FBA using the AGORA and a refined model of Faecalibacterium prausnitzii using in vitro media and alterations of those. Table S7. Experimental growth rates for monoculture species and bacterial load in media. Growth rates for the ABB, YCAG, YCFA, YCGD and YCGMS media, were obtained manually. Growth rates are expressed in h−1. Table S8. Description of the GEM used for each species with its structure description and MEMOTE score with species names from the literature used for experimental data and modelling and their updated names (nomenclature update 12/22). Table S9. Mouse strains and their matched AGORA metabolic models. The selection was based on the highest ANI score between each mouse strain genome and the AGORA models’ corresponding genomes. Table S10. ANI score for alignment of mice species against AGORA model species (only top 5 ANI scores shown). ANI scores below 80% were not considered in this analysis. Table S11. ANI scores of mice species against AGORA2 model species. With green highlight the genomes included in AGORA2 that have the highest ANI score with the genome of interest, with orange the corresponding ones included in AGORA v1. Table S12. Comparison of the growth rates returned from MMT using pairs of mice AGORA and AGORA2 models. R1 stands for the growth of species A and R2 for the one of species B. Table S13. In silico media description (IVm and WD) where values represent the lower bounds of uptake fluxes. The lower bounds are set as the mmol/L concentration of every metabolite in the experimental medium assuming unit values for biomass and time.