Evaluating DNA extraction methods for eDNA metabarcoding of soil invertebrate diversity

Abstract

Metabarcoding of environmental DNA (eDNA) has been increasingly used in assessing soil biodiversity, primarily for microorganisms but also for invertebrates. Currently, conventional morphological identification (CMI) for detecting microarthropods and earthworms involves extracting them via heat treatment or hand-sorting from soil blocks, and subsequent morphological identification. To compare the soil fauna community composition assessment methods, we compared CMI, DNA metabarcoding of heat-extracted invertebrates (comDNA), and DNA extracted directly from soil (eDNA). For eDNA, two commercially available QIAGEN DNA extraction kits were further compared: DNeasy Powerlyzer PowerSoil kit (eDNA_PS), based on 0.25 g of soil, and DNeasy PowerMax soil kit (eDNA_PM), based on 10 g of soil. PowerMax captured higher richness, while PowerSoil captured diversity comparable to that of comDNA. In eDNA and comDNA samples, arthropods dominated the community composition, followed by annelids. Both eDNA and comDNA methods captured several overlapping species,; however, each method also detected unique ASVs. Interestingly, comDNA captured a higher abundance of several ASVs that were not detected in eDNA. Regardless of the methods used, the location of the soil sampled showed a significant effect on soil fauna community structure. Several species detected or shared in DNA-based methods were also shared with CMI, and a few collembolan species detected by eDNA were also correlated with the abundance data from CMI. Further, the community composition of collembolans varied between the comDNA and two eDNA (eDNA_PS, eDNA_PM) methods; however, more than one-third of the species were detected across all three methods. Our findings show the complementarity of eDNA and comDNA and support the integration of DNA-based methods in future soil fauna biodiversity assessment programs.