Qing Dai

Microbial habitats that receive repeated external input may not remain shaped by that input forever if local retention allows resident communities to build up over time. Here, we used a controlled bench-scale sink p-trap system to examine how community assembly unfolded during initial establishment in new, bleach-treated p-traps. Two p-traps received repeated handwashing-water input, while one received tap water as baseline. The treated p-traps, but not the control, showed clear successional change toward later resident-like states. Nested-model comparisons further showed that recent external input had its greatest influence early in succession, but the p-trap's own prior state remained the stronger predictor throughout. Final-day post-flush trajectories indicated short-term displacement from pre-flush positions, with later time points tending to move back toward late-stage resident centroids. Together, these results show that repeated inoculation does not necessarily keep communities under continued outside influence. Instead, retentive microbial habitats can shift over time from early sensitivity to external input toward persistence shaped more by local history.

Barcode crosstalk is a potential source of error in indexed multiplex sequencing that can mimic cross-contamination and generate false-positive signals, particularly in experiments with low-biomass samples and imbalanced designs. Here we identify and quantify barcode crosstalk in multiplexed Oxford Nanopore sequencing and introduce post-ligation pooling (PLP). PLP is a drop-in library-preparation modification that prevents barcode crosstalk, rather than simply mitigating its effects. These results highlight the potential for barcode crosstalk to have detrimental effects in multiplexed long-read sequencing and establish PLP as an immediately adoptable mitigation, especially where negative controls and low-abundance signals inform biological interpretation.