Quantitative property of metabarcoding
Metabarcoding on the 18S rRNA region of DNA is a common strategy to
study zooplankton. Various sections of this region have been used as
barcodes (e.g. , V9 or V4 the standard markers for planktonic
eukaryotes since Tara Oceans studies). In this study, we choose to
amplify and sequence the V1V2 region, originally used for meiofaunal
zoobenthos (Fonseca et al., 2014) and specific to Metazoan (Lejzerowicz
et al., 2021). It appears to be one of the best markers to assess marine
community changes (Cordier et al., 2018). However, it is important to
consider that some taxa are not well characterized, e.g. ,Pleuromamma sp. or Chaetognatha . This may be due to a
secondary structure on their ribosomal region which might make their DNA
harder to amplify as the primer regions show high fidelity in the
available sequences in the NCBI. This could impact our results sincePleuromamma sp., which is highly abundant at BATS, is known to
rapidly react to the new primary production and to contribute
significantly to the ecosystem carbon flux (between 4% and 70%,
Steinberg et al., 2000). Chaetognaths are also common in this study
region. They are thought to be linked to the bloom community, since
their abundance usually follows copepod density maxima and decreases
with increasing stratification of the water column in the Sargasso Sea
(Ivory et al., 2019) and their morphology matches this cluster’s
characteristics. Therefore, due to chaetognath’s sequences low
amplification we might have missed potential interspecific interactions.
Generally, correlations are found between the number of reads per ASV
and the corresponding taxon biomass or abundance (Bucklin et al., 2019).
Significant relationship also exists between the proportion of input
material and the proportion of sequences obtained per species from
metabarcoding, however large uncertainties remain (Lamb et al., 2019).
Here, positive correlations were found for some coarse taxonomic groups.
A negative correlation was found for ostracods abundances versus
sequence reads proportions. It was not the case in Matthews et al.
(2021) in which they found a positive correlation between ostracods’
relative abundances and proportion of sequences for both 18S V4 and COI
primers. It might be due to a technical bias from the metabarcoding
process. On the contrary, some groups appeared to be well suited for a
quantitative metabarcoding use. Positive correlations were found for
abundance and biomass of Cyclopoida, Doliolida, Harpacticoida and
Larvacea. For biomass, similar trends were observed for Cyclopoida,
Doliolida and Larvacea (Matthews et al., 2021). The proportion of reads
of Chaetognatha and Cladocera well correlated with their abundance but
not their biomass. Chaetognatha are not well identified by 18S V1V2
(Matthews et al., 2021) which might explain the absence of correlation
between relative number of sequences and biomass. On the contrary,
relative number of sequences and biomass was positively correlated for
Cephalopoda despite their low abundance. For the remaining taxa, the
absence of significant correlations might be due to not-specific or
inaccurate conversion factors from biovolume to dry weight. Hence, more
studies such as the one by Maas et al. (2021) should be done before
assuming a quantitative nature.