Comparison of DNA metabarcoding and morphological identification for stream macroinvertebrate biodiversity assessment and monitoring

Abstract

AbstractConventional morphology-based identification is commonly used for routine assessment of freshwater ecosystems. However, cost and time efficient techniques such as high-throughput sequencing (HTS) based approaches may resolve the constraints encountered in conducting morphology-based surveys. Here, we characterized stream macroinvertebrate species diversity and community composition via metabarcoding and morphological analysis from environmental samples collected from the Shigenobu River Basin in Ehime Prefecture, Japan. We compared diversity metrics and assessed both approaches’ ability to evaluate the relationship between macroinvertebrate community and environmental variables. In total, we morphologically identified 45 taxa (3 families, six subfamilies, 31 genera, and five species) from 8,276 collected individuals from ten study sites. We detected 44 species by metabarcoding, with 35 species collapsed into 11 groups matching the morphologically identified taxa. A significant positive correlation between logged depth (number of HTS reads) and abundance of morphological taxa was observed, which implied that quantitative data can be used for subsequent analyses. Relatively higher estimates of alpha diversity were calculated from the metabarcoding data in comparison to morphology-based data. However, beta diversity estimates between metabarcoding and morphology data based on both incidence and abundance-based matrices were correlated proving that community differences between sampling sites were preserved in the molecular data. Also, both models were significant, but metabarcoding data (93%) explained a relatively higher percentage of variation in the relationship between community composition and the environmental variables than morphological data (91%). Overall, we present both the feasibility and limitations of HTS-driven estimations of taxonomic richness, community composition, and diversity metrics, and that metabarcoding was proven comparable and more sensitive against morphology-based analysis for stream macroinvertebrate biodiversity assessment and environmental monitoring.