Microbial community assembly in engineered biological systems is often simultaneously influenced by stochastic and deterministic processes, and the nexus of these two mechanisms remains to be further investigated. Here, three lab-scale activated sludge reactors were seeded with identical inoculum and operated in parallel under eight different sludge retention time (SRT) by sequentially reducing the SRT from 15 days to 1 day. Using 16S rRNA gene amplicon sequencing data, the microbial populations at the start-up (15-day SRT) and SRT-driven (≤10-day SRT) phases were observed to be noticeably different. Clustering results demonstrated ecological succession at the start-up phase with no consistent successional steps among the three reactors, suggesting that stochastic processes played an important role in the community assembly during primary succession. At the SRT-driven phase, the three reactors shared 31 core operational taxonomic units (OTUs). Putative primary acetate utilizers and secondary metabolizers were proposed based on K-means clustering, network and synchrony analysis. The shared core populations accounted for 65% of the total abundance, indicating that the microbial communities at the SRT-driven phase were shaped predominantly by deterministic processes. Sloan’s Neutral model and a null model analysis were performed to disentangle and quantify the relative influence of stochastic and deterministic processes on community assembly. The increased estimated migration rate in the neutral community model and the higher percentage of stochasticity in the null model implied that stochastic community assembly was intensified by strong deterministic factors. This was confirmed by the significantly different α- and β-diversity indices at SRTs shorter than 2 days and the observation that over half of the core OTUs were unshared or unsynchronized. Overall, this study provided quantitative insights into the nexus of stochastic and deterministic processes on microbial community assembly in a biological process.

A stand-alone down-flow hanging sponge (DHS) system with a two-stage configuration was operated for 700 days to treat synthetic soft drink wastewater at 3000 mg/L chemical oxygen demand (COD). Throughout the operation, >90% COD and total organic carbon (TOC) removal efficiency was obtained by the first stage, and a final effluent of COD <60 mg/L (TOC <20 mg/L) was consistently maintained with the second stage. Lower organic removal efficiency was observed to closely correlate with lower pH, higher volatile fatty acid (VFA) concentration, and higher suspended solid (SS) in the effluent. Occasionally, biomass sloughing was observed as a cause to unstable reactor performance in the first stage. The microbial community of the retained biomass on the sponges differed significantly based on spatial locations of sponges, sampling time points, and loading shocks. In general, Proteobacteria were found to be more abundant in the reactor at an organic removal efficiency >80% than that at <50%. Specifically, operational taxonomic units closely related to Tolumonas auensis and Rivicola pingtungensis were identified as important populations that were responsible for degrading the major substrate in the soft drink wastewater toward to the end of the reactor operation. In addition, high abundance of Bacteroidetes in the reactor was speculated to be responsible for the VFA accumulation in the effluent. This study demonstrated that stand-alone DHS reactor could be used in treating high-strength wastewater efficiently.