Environmental DNA sampling workflow matters: testing different approaches on marine microbial assemblages in three coastal areas of Greece

Abstract

Introduction Environmental DNA (eDNA) and high-throughput sequencing are now common tools for investigating marine microbial communities in coastal environments. Nevertheless, different eDNA sampling workflows are used across studies, while direct comparisons of their outputs under the same environmental and analytical framework remain limited. Methods We aimed to compare three eDNA sampling approaches, two active filtration workflows using membrane filtration (Filter) and Sterivex cartridges (Sterivex), and one passive Gauze-based workflow, applied simultaneously in three coastal gulfs of Greece, to examine whether sampling workflow influences the recovered bacterial and eukaryotic diversity and community structure via 16S and 18S rRNA gene metabarcoding. The workflows differed in sampling characteristics, including pre-screening, processed water volume and capture mechanism, with 200-5m pre-screening applied only to Filter and Sterivex. Results Filter and Sterivex produced concordant patterns of community structure and generally lower between-sample variability. They were nonetheless not fully equivalent: differences in diversity, taxonomic representation and network hub identity indicate that the two active workflows recover partially distinct community representations. Gauze produced a more variable and compositionally distinct signal, particularly for eukaryotic communities, where concordance with the active workflows was weak and occasional low-yield recovery occurred. Workflow-related differences were also observed in exploratory co-occurrence network topology and hub identity, indicating that workflow choice can influence downstream ecological interpretation. Discussion Each workflow appeared to be suitable for different research objectives, with passive sampling providing complementary information. Maintaining a consistent sampling workflow is therefore particularly important for monitoring programs and cross-study comparisons.