Fig. 1. Global sensitivity analysis of synthetic co-cultures.

a, Cartoon depiction of the model (full description in Methods). G is the culture carbon source (for example, glucose), x_i is the essential metabolite produced/received by the co-culture system and y_i is the population of strain i. b, Initial simulations of the impact of metabolite production on host growth in yeast monocultures. Left: metabolite production rate over time. Right: metabolite production has a nonlinear relationship with maximal growth. Colours represent the strength of the ϕ₁ parameter, which governs the production of the exchanged metabolite x₁. c,d, Simulations of the two-member co-culture system at different strengths of metabolite exchange. c, Total population size at 72 h. d, Proportion of strain 1, y₁, in the culture. e–h, Global sensitivity analysis of the two-member co-culture. Model is described in Methods and results are fully discussed in Supplementary Note 2. Parameters are as follows: ϕ_i is the proportion of glucose flux going to production of metabolite i by strain y_i. $x_{0,i}$ is the initial concentration of metabolite i in the medium. $r_{0,i}$ is the initial starting population of strain i (note that $r_{0,1}+r_{0,2}=1$). $V_{\max ,G}^{y_i}$ is the maximum uptake rate of glucose G by strain y_i. $V_{\max ,j}^{y_i}$ is the maximum uptake rate of metabolite j by strain y_i. δ is the dummy parameter used for statistical tests in the local sensitivity analysis as described in Methods. Asterisk denotes either sensitivity or total sensitivity is significantly different (P < 0.01) from the dummy parameter as determined by a t-test using Bonferroni correction. Results are reported as mean ± s.d. for 100 resamplings. Shown are the sensitivities of the final OD₇₀₀ (e), the total batch culture time (f), the final population ratio (g) and the growth rate of each strain (h) to key parameters in the model.

Source data