Biodiversity detection
DNA metabarcoding has been proven to be a very effective and powerful
tool to study arthropod communities as a whole by overcoming the
taxonomic impediment, allowing detection of a large variety of taxa (see
Liu et al., 2020 and references therein). To the best of our knowledge,
this is the first time that such a wide biodiversity analysis has been
carried out at the species level in Mediterranean evergreen oak
woodlands (“dehesas” or “montados”). Previous studies on soil
arthropod communities from this habitat had been performed by grouping
at higher taxonomic levels: Class, Order and, in rarer cases, Family
(Andrés et al., 1999; Rota et al., 2015; Sadaka & Ponge, 2003). At a
lower taxonomic level, only checklists of species from certain families
have been reported (e.g. Zaballos, 1983), but in no case related to the
effects of livestock on arthropods. In the present study we could
analyse it at the MOTU level, despite not all of them could be assigned
to Linnean species.
Identification success differed among taxonomic levels depending on the
availability of reference sequences. Almost all arthropod MOTUs were
identified to Order level, and more than 80% to Family level. We could
thus identify most of the arthropod families coexisting in these Iberian
holm oak savannas. However, identification success at the species level
was lower: we could assign Linnean species names to only 40% of the
MOTUs. This highlights the importance of promoting classical taxonomy
research as the solid base to create comprehensive and reliable
reference databases of DNA barcodes, indispensable for the use of
molecular identification methods. Moreover, these databases should be
geographically unbiased, since in the Mediterranean region, where our
study took place, reference barcodes are scarcer than, for example, in
Central Europe, which decreases identification likelihood (Gaytán et
al., 2020).
The use of two different genetic markers increased biodiversity
detection success. The complementary use of 16S in addition to COI did
not contribute detection at the Order level, but it did at the Family
level. Although more families were exclusively detected by COI, others
were only retrieved by 16S. Moreover, these families may have had an
important role in the biodiversity analysis, as differences between
treatments in the richness at MOTU level were always more pronounced
using the 16S dataset. These results agree with a previous study which
concluded that the best markers to detect Hexapoda subphylum barcodes
are found in the mitochondrial ribosomal RNA (rRNA) genes, such as 12S
and 16S (Marquina et al., 2019b). However, our results suggest that
their conclusions may be applicable not just on Hexapoda, but also to
all Arthropoda phylum. Nevertheless, the lack of reference data compared
to COI ballasts the use of this marker if the main aim is species
identification.
For this work, we chose a non-destructive approach for the DNA
extraction from the bulk samples, as we were interested in keeping the
individual arthropods for further morphological analysis. A series of
recent studies have tested the suitability of these non-destructive DNA
extraction protocols, finding them as a promising alternative to sample
homogenization in terms of biodiversity detection (Batovska et al.,
2021; Iwaszkiewicz-Eggebrecht et
al., 2023; Marquina et al., 2022; Nielsen et al., 2019). However, there
may be limitations relevant for the analysis of our samples: for
instance, both Iwaszkiewicz-Eggebrecht et al. (2023) and Marquina et al.
(2022) found that ants (Hymenoptera: Formicidae) were detected in their
mock communities in much lower proportion and occasions than other taxa.
This could have affected our results, as we worked with ground-dwelling
arthropod communities, in which ants are prominent members. Nonetheless,
we did find a great diversity of MOTUs belonging to Formicidae (18 with
COI and 12 with 16S), which we consider acceptable.