Abstract
Bottom-up selection has an important role in microbial community assembly but is unable to account for all observed variance. Other processes like top-down selection (e.g., predation) may be partially responsible for the unexplained variance. However, top-down processes and their interaction with bottom-up selective pressures often remain unexplored. We utilised an in situ marine biofilm model system to test the effects of bottom-up (i.e., substrate properties) and top-down (i.e., large predator exclusion via 100 mu m mesh) selective pressures on community assembly over time (56 days). Prokaryotic and eukaryotic community compositions were monitored using 16 S and 18 S rRNA gene amplicon sequencing. Higher compositional variance was explained by growth substrate in early successional stages, but as biofilms mature, top-down predation becomes progressively more important. Wooden substrates promoted heterotrophic growth, whereas inert substrates' (i.e., plastic, glass, tile) lack of degradable material selected for autotrophs. Early wood communities contained more mixotrophs and heterotrophs (e.g., the total abundance of Proteobacteria and Euglenozoa was 34% and 41% greater within wood compared to inert substrates). Inert substrates instead showed twice the autotrophic abundance (e.g., cyanobacteria and ochrophyta made up 37% and 10% more of the total abundance within inert substrates than in wood). Late native (non-enclosed) communities were mostly dominated by autotrophs across all substrates, whereas high heterotrophic abundance characterised enclosed communities. Late communities were primarily under top-down control, where large predators successively pruned heterotrophs. Integrating a top-down control increased explainable variance by 7-52%, leading to increased understanding of the underlying ecological processes guiding multitrophic community assembly and successional dynamics.