Soil eukaryote community shift but not composition is consistently
recovered by different OTU inference methods applied to long read
metabarcoding data
Abstract
Long amplicon metabarcoding has opened the door for phylogenetic
analysis of the largely unknown communites of microeukaryotes in soil.
Here, we amplified and sequenced the ITS and LSU regions of the rDNA
operon (around 1500 bp) from grassland soils using PacBio SMRT
sequencing and evaluated the performance of three different methods for
generation of operational taxonomic units (OTUs). The field site at
Kungsängen Nature Reserve has drawn frequent visitors since Linnaeus’s
time, and its species rich vegetation includes the largest population of
Fritillaria meleagris in Sweden. To test the effect of different OTU
generation methods, we sampled soils across an abrupt moisture
transition that divides the meadow community into a Carex acuta
dominated plant community with low species richness in the wetter part,
which is visually distinct from the mesic-dry part that has a species
rich grass-dominated plant community including a high frequency of F.
meleagris. We used the moisture and plant community transition as a
framework to investigare how detected belowground microeukaryotic
community composition was influenced by OTU generation methods. Soil
communities in both moisture regimes were dominated by protists, a large
fraction of which were taxonomically assigned to Ciliophora (Alveolata)
while 30-40% of all reads were assigned to kingdom Fungi. Ecological
patterns were consistently recovered irrespective of OTU generation
method used. However, different methods strongly affect richness
estimates and the taxonomic and phylogenetic resolution of the
characterized community with implications for how well members of the
microeukaryotic communities can be recognized in the data.