Table 7.
Explanation of simulation data sets used to generate figures of aggregate results and comparative outcome for ABM-PHYS and ABM-ST.
| Parameter type | Analysis type | Number of simulations | Figure numbers |
|---|---|---|---|
| Variable parameter | Comparison of average phenotypes across | N = 300 | Supplementary Figures 3–5 |
| all outcome categories (clearance, containment, dissemination). | |||
| Comparison of statistical distribution of outcomes | N = 300 | Figure 6 | |
| for in vivo, ABM-PHYS, and ABM-ST. | |||
| Identification of statistically significant model | N = 300 | Supplementary Figure 6A | |
| parameters across all outcomes. | |||
| Identification of statistically significant model | N = 41 (of 150) | Supplementary Figure 6B | |
| parameters for containment and dissemination outcomes. | |||
| Fixed parameter | Outcome specific comparison of bacterial load phenotype. | N = 5 to 6 | Figure 5 |
| * Data sets where some | Comparative analysis of the hypoxic region of | N = 20(*4) | Figure 7 |
| are from variable | containment granulomas. | ||
| parameter runs | Comparison of the bacterial load in in vivo | N = 10,22(*1) | Figure 9 |
| vs. in silico containment granulomas. | |||
| Comparison of Mtb gene expression for containment | N = 20 (*10) | Supplementary Figure 2 | |
| and dissemination outcomes. |
Aggregate or average results are generated based on the N = 300 LHS simulation run set or from fixed parameter simulation sets (parameters used for fixed parameter studies are listed in Appendix 4 in Supplementary Material). Figure 6, Supplementary Figures 3–5, 6A are generated using the N = 300 simulation set for both the standard and integrated ABM models, where we categorized outcomes as described in Appendix 1 in Supplementary Material. Supplementary Figure 6B is produced using 41 of 150 simulation results (over the same parameter space as the N = 300 simulation set) in order to investigate which significant parameters result when we exclude the clearance outcomes. Figure 5, which compares the number of extracellular bacteria in the standard ABM to the integrated ABM, is generated from 5 to 6 fixed parameter simulations for each scenario. In the remaining figures we used fixed parameter simulation runs to generate a sufficient number of containment samples for comparative analysis and characterization of the containment response: Figure 7 results are based on 20 simulations representing 6 solid containment granulomas and 14 caseous containment granulomas (of the 20 simulations, 4 were from the N = 300 sample run set and 16 from fixed parameter simulation runs); Figure 9 results are based on 10 fixed parameter simulations for the standard ABM and 22 simulations for the integrated multiscale ABM (of the 22 simulations, 1 was from the N = 300 sample run and 21 from fixed parameter simulation runs); Supplementary Figure 2 results are based on 20 simulations (10 were from the N = 300 sample run set and 10 from fixed parameter simulation runs).