DNA metabarcoding of mites from small soil samples: limited agreement with morphological identifications but improved results from long-read sequencing.

The characterization of soil mite (Acari) communities traditionally follows morphological identifications of specimens extracted from soil, which is a highly laborious and time-consuming process. Metabarcoding has become an increasingly utilized approach for species identification from environmental DNA (eDNA) samples, but whether the metabarcoding approaches align with the morphological identification data on soil mites has rarely been addressed. Here, we examine the congruence of soil mite communities between morphological and metabarcoding datasets. The morphological dataset was generated by extracting mite specimens from the soil samples, whereas molecular datasets represent two types of cytochrome c oxidase subunit I (COI) amplicons produced directly from soil eDNA (from 0.2 g and 2 g soil samples) and sequenced with Illumina (313 base pairs amplicons) and PacBio (658 base pairs amplicons) platforms. We found that specimen extraction from soil samples, followed by morphological identification, yielded the highest number of mite species. Despite significantly lower mite richness in the metabarcoding datasets, PacBio datasets provided more reliable community profiles that aligned strongly with the morphological data. This indicates that soil sample quantities generally used for microbial analyses are also informative in studying soil faunal communities. Furthermore, our results indicate that methodological choices (herein PacBio