Past, present, and future spatial distributions of deep-sea coral and sponge microbiomes revealed by predictive models

Abstract

Knowledge of the spatial distribution patterns of biodiversity in the ocean is key to evaluate and ensure ocean integrity and resilience. Especially for the deep ocean, where in situ monitoring requires sophisticated instruments and considerable [fi]nancial investments, modelling approaches are crucial to move from scattered data points to predictive continuous maps. Those modelling approaches are commonly run on the macrobial level, but spatio-temporal predictions of host-associated microbiomes are not being targeted. This is especially problematic as previous research has highlighted that host-associated microbes (microbiomes) may display distribution patterns that are not perfectly correlated with host animal biogeographies, but also with other factors such as prevailing environmental conditions. We here establish a new simulation approach and present predicted spatio-temporal distribution patterns of deep-sea sponge and coral microbiomes, making use of a combination of environmental data, host data and microbiome data to advance our understanding of deep-sea microbiomes. This approach allows predictions of microbiome spatio-temporal distribution patterns on scales that are currently not covered by classical sampling approaches at sea. This includes both predictions in space within regional oceanic provinces off eastern North America, and also in time, with predictions into the past and future, covering a time span of 214 years. In summary, our presented predictions allow (i) identi[fi]cation of microbial biodiversity hotspots in the past, present, and future, (ii) evaluation of microbial-macrobial connections at case-study sites through trait-based predictions, and (iii) identi[fi]cation of shifts in microbial community composition (key taxa) across environmental gradients and shifting environmental conditions.