Abstract
ABSTRACT
Environmental DNA (eDNA) exists in three states: membrane‐bound, adsorbed and dissolved. These states differ in persistence and degradation, strongly influencing the interpretation of eDNA data. Despite this, they have rarely been separated and analyzed independently from environmental samples. We developed a state‐sorting workflow to isolate and analyze them, applying it to samples from 221 sites from 58 streams across eight lake watersheds with COI and ITS metabarcoding, targeting metazoans and plants, respectively, to reveal differences in biodiversity content and transport dynamics. Our results show that all three states contain both shared and unique taxonomic diversity of metazoan eDNA. However, the plant eDNA was only detected in the membrane‐bound state. For metazoans, membrane‐bound eDNA contained 87.8% of observed ASVs, far exceeding the adsorbed (37.2%) and dissolved (20.5%) states. Only membrane‐bound eDNA showed evidence of downstream transport, but its extent varied among watersheds due to local hydrology. While upstream eDNA was transported to stream–lake confluences, lake surface samples showed a turnover in community composition. Clarifying the fate of membrane‐bound eDNA within lakes will enhance catchment‐level detection from lake samples and understanding of lake hydrodynamics. Of the environmental parameters assessed, water temperature was most strongly aligned with changes in community composition between sites. Most previous studies have likely captured the majority of the eDNA diversity within their samples by inadvertently targeting the membrane‐bound state. This study demonstrates the utility of eDNA state‐sorting, but the methods require further refinement. Analyzing states independently improves the interpretation of eDNA data and elucidates the processes governing eDNA fate and transport.