The accurate qualitative and quantitative assessment of biodiversity is essential in order to understand biodiversity and ecosystem function relationships, especially in the face of rapid biodiversity loss. However, the scale and intensity of contemporary biodiversity identification challenges are limited by traditional taxonomic approaches. As an alternative, Next Generation Sequencing applications (NGS) have been developed to rapidly detect diversity from bulk community samples, with particular focus on macroinvertebrates from freshwater ecosystems. Most commonly used approaches are PCR based (metabarcoding) that successfully detect species richness, but due to PCR biases, are estimated to be less accurate at determining species abundance. Recently, mito-metagenomic approaches, that incorporate shotgun sequencing of mitochondrial genomes from bulk samples, are predicted to provide more accurate abundance estimations. Using a carefully designed experiment, here we evaluate the performance of both metabarcoding vs. mito-metagenomics for qualitative and quantitative biodiversity assessment. Ten artificial communities comprising 13 aquatic macroinvertebrate species (1,500 specimens) were recreated using exact body measurements as a proxy for biomass calculation. Each bulk community was sequenced for three amplicons of the Cytochrome Oxidase Subunit I (COI) gene as well as shotgun sequenced for complete mitochondrial genomes. Individual reference genomes were also produced for each species.
Results/Conclusions
Metabarcoding of the COI region returned >1.4 million quality filtered for all three amplicons, while shotgun sequencing achieved ~2.5Gb data per community. Both methods performed well in the majority of cases, with few exceptions such as failing to detect some of the rarest and lowest in biomass species in some communities (<5%). Metabarcoding was less consistent as it also missed some species present in higher biomass content. Preliminary results on abundance estimation suggest that the mito-metagenomics approach predicted more accurately biomass and abundance in bulk samples. For metabarcoding, individual amplicons failed to match biomass relationships in several cases but combined results of all amplicons were significantly correlated with biomass counts. The results suggest that the mito-metagenomic approach could be more accurate for abundance estimation but comparable results could be obtained from metabarcoding if a combined multi-amplicon methodology is used. The use of mito-metagenomics approaches presents reliable estimations of detectable invertebrate biomass at the community level while reducing workload compared to metabarcoding. Collectively, understanding the quantitative nature of NGS approaches to biodiversity identification is an important step in addressing many questions across the ecological sphere with associated benefits for improving accuracy of ecological surveys for freshwater ecosystem monitoring.