Marine biomonitoring with eDNA: can metabarcoding of water samples cut it as a tool for surveying benthic communities?

In the marine realm, biomonitoring using eDNA of benthic communities
requires destructive direct sampling or the setting-up of settlement
structures. Comparatively much less effort is required to sample the water
column, which can be accessed remotely. In this study we assess the
feasibility of obtaining information from the eukaryotic benthic
communities by sampling the adjacent water layer. We studied two different
rocky-substrate benthic communities with a technique based on quadrat
sampling. We also took replicate water samples at distances from a few
centimetres to 20 m from the benthic habitat. Using as marker a fragment
of the Cytochrome c oxidase subunit I gene with universal primers, we
obtained a total of 3,543 molecular operational taxonomic units (MOTUs)
from the samples. The structure obtained in the two environments was
markedly different, with Metazoa, Archaeplastida Rhodophyta and
Stramenopiles being the most diverse group in benthic samples, and
HacrobiaAlveolata, Metazoa and Alveolata Rhizaria in the water. Only 265
MOTUs (7.5%) were shared between benthos and water samples, and of these
180 MOTUs (5.1%) were identified as benthic MOTUs that left their DNA in
the water. Most of them were found immediately adjacent to the benthos,
and their number decreased and The distribution of these benthic shared
MOTUs showed a decrease both in number of MOTUs and in number of reads as
we moved apart from the benthic habitat. It was concluded that water eDNA,
even in the close vicinity of the benthos, was a poor proxy for the
analysis of benthic structure, and that direct sampling methods are
required for monitoring these complex benthic communities via
metabarcoding.