Use of null models to compare the assembly of northeast Atlantic bacterial community in the presence of crude oil with either chemical dispersant or biosurfactant

Abstract

The compositions of marine microbial communities in response to crude oil in the presence of biosurfactant or synthetic dispersants have been extensively studied in the last decade. Assembly processes, however, in such communities are poorly understood. In this study, we used seven different but complementing null model approaches, such as elements of metacommunity structure, Raup-Crick beta-diversity, normalised stochasticity ratio, Tucker’s null model, quantitative process estimates, lottery assembly, and phylogenetic dispersion models, to quantify the relative importance of ecological process that drive the community assembly. We found that the presence of chemical dispersant in the oil-amended microcosms induced significant temporal changes in the assembly processes that were different from the oil-only or biogenic dispersant-amended microcosms. The assembly processes in all microcosms were neither purely deterministic nor stochastic, but increasingly deterministic in dispersant-amended microcosms. Furthermore, the relative importance of determinisms varied over time and was strongest during the middle phase of incubation. Tucker’s null model revealed that phylogenetically distinct taxa might have shaped the bacterial community assembly in the different microcosms towards more niche or neutral processes. Moreover, there was faster recruitment of phylogenetically distant species in the dispersant-amended community. Drift, homogenising selection and dispersal limitation were the dominant assembly processes in all microcosms, but variable selection was only important in dispersant-amended microcosms. In conclusion, our study highlights that the assembly processes in marine bacterial communities are not static but rather dynamic, and the chemical dispersant can cause significantly different patterns of community assembly compared to non-amended or biosurfactant-amended microcosms.ImportanceThe null model strategy is designed to intentionally exclude an ecological or evolutionary process of interest and create a beta diversity pattern that would be expected in the absence of this particular process – i.e. the community structure is random in respect to the process being tested. Recent advancements of bioinformatics and statistical tools have made it possible to apply theoretical macroecological concepts to microbial metagenomics in order to better understand and quantify the mechanisms and patterns controlling the complexity of microbial ecology. The conclusions from the null models can help predict the changes in microbial biodiversity and ecosystem services in oil polluted environments and therefore assist in making effective decisions with regards to what would be the best oil spill response option for similar environmental conditions.