There is a growing awareness that the bacterial interactions should follow a highly nonlinear pattern in reality. However, it is challenging to tract the varying bacterial interactions using the pair-wise correlation analysis, which fails to explore their potential effects on the behavior of microbes. Here, we utilize the regularized S-map to capture the varying interspecific interactions from the time-series data of bacterial community under the exposure to nitrite. Our results show that the bacterial interactions are highly variable and asymmetric interactions dominate the interaction pattern in community. Furthermore, we propose a Jacobian coefficient-based statistical method to predict the harmony level of a bacterial community at each successive ecosystem state. The result shows that bacterial community exhibits higher harmony level in nitrite-treated samples than the control group. We show that the community harmony level is positively associated with the specific endogenous respiration rates and biofilm formation of the culture. In addition, the community tends to process lower diversity and structural stability under zero and high nitrite stresses. We demonstrate that harmony level, other than structural stability, is a useful index to unveil the underlying mechanism of bacterial performance. Overall, the regularized S-map can help us to understand the bacterial interactions in eco-systems more accurately than previous approaches.
Importance
It has long been acknowledged that bacterial interactions play important roles in the community structure and function. Revealing the interaction variability can allow an understanding of how bacteria response to perturbation and why bacterial community performance changes. Such information should improve our skills to engineer the bacterial community (e.g., wastewater treatment plant) and achieve better removal performance and lower energy consumption.