The development of new methods capable of more realistic modeling of microbial communities necessitates that their results be quantitatively comparable with experimental findings. In this research, a new integrated agent and constraint based modeling framework abbreviated ACBM has been proposed that integrates agent-based and constraint-based modeling approaches. ACBM models the cell population in three-dimensional space to predict spatial and temporal dynamics and metabolic interactions. When used to simulate the batch growth of C. beijerinckii and two-species communities of F. prausnitzii and B. adolescent., ACBM improved on predictions made by two previous models. Furthermore, when transcriptomic data were integrated with a metabolic model of E. coli to consider intracellular constraints in the metabolism, ACBM accurately predicted growth rate, half-rate constant, and concentration of biomass, glucose, and acidic products over time. The results also show that the framework was able to predict the metabolism changes in the early stationary compared to the log phase. Finally, ACBM was implemented to estimate starved cells under heterogeneous feeding and it was concluded that a percentage of cells are always subject to starvation in a bioreactor with high volume.

能够对微生物群落进行更现实的建模的新方法的开发，要求其结果在数量上可以与实验结果相媲美。在这项研究中，提出了一个新的基于代理和约束的集成建模框架，简称ACBM，该框架集成了基于代理和基于约束的建模方法。ACBM对三维空间中的细胞群进行建模，以预测时空动态和代谢相互作用。当用来模拟C. beijerinckii以及F. prausnitzii和B.青春期的两种物种的分批生长时，ACBM改进了先前两个模型所做的预测。此外，当转录组数据与代谢模型结合时考虑到代谢中的细胞内限制，大肠杆菌会准确预测生长速率，半速率常数以及生物量，葡萄糖和酸性产物随时间的浓度。结果还表明，与对数期相比，该框架能够预测静止早期的代谢变化。最后，采用ACBM估计异质饲喂下的饥饿细胞，并得出结论，在高容量的生物反应器中一定比例的细胞总是处于饥饿状态。