Bacteria almost always form communities, except in special cases. Catastrophic shifts in bacterial communities can cause severe problems, such as human diseases in the human gut, or damage to agricultural crops. Therefore, understanding community stability is needed for controlling bacterial community dynamics.
In microbial ecology, it has been suggested that communities may have multiple stable compositions. However, this hypothesis is mainly based on the fact that there is a bias in the observed patterns, which cannot be distinguished from cases where the bias exists merely in the observations themselves. In this study, we tracked the bacterial community dynamics in 5 eel aquaculture tanks over one year, collecting 1,376 observational data points and applied energy landscape analysis to investigate multiple stability in species-rich bacterial communities.
Energy landscape analysis can quantify the stability of the possible states of systems composed of many components. This method can be divided into two major steps. First, the expected values of occurrence probabilities for each possible state are calculated based on observational data. The expected values of occurrence probabilities can be interpreted as the stability, assuming that a more stable state changes less and has a higher observation probability when sampled randomly. Next, considering the transition network, the energy landscape, characterized by stable states, basins of attraction, and tipping points, can be estimated. This framework has been applied in various fields, such as physical chemistry and neuroscience.
As a result, multiple stable compositions were detected with varying stability levels within the bacterial communities. Furthermore, the stability indicator significantly correlated with the degree of change (dissimilarity) in community composition over 1 day. This demonstrated that low stability (high instability) effectively predicts changes in community composition.