Optimizing Environmental DNA Metabarcoding of Archived Suspended Particulate Matter for River Biodiversity Monitoring

Environmental DNA (eDNA) metabarcoding has become a prominent method in aquatic biodiversity research and monitoring. A widely underexplored sample type is suspended particulate matter (SPM), composed of small organic and inorganic particles that contain eDNA. SPM is typically collected with sedimentation boxes, for instance by the German Environmental Specimen Bank for chemical monitoring of rivers. Its standardized collection and storage make SPM a potentially ideal source for eDNA time series analysis, linking environmental pressures and responses. This study evaluated key steps in the laboratory workflow allowing for the implementation of SPM in biodiversity research and monitoring. After validating an in-house extraction protocol, we extracted eDNA from three 12-year-old SPM samples representing three river sites. To quantify replicate variation in species detection, each sample was processed with three subsamples, four extraction replicates per subsample, and four PCR replicates per extraction replicate (48 replicates per SPM sample). Additionally, mechanically ground samples were investigated to test for decreased inter-sample variation. Our analysis revealed extensive, and plausible taxa lists for fish and aquatic invertebrates with significant differences between the sites. Within sites, we found high stochasticity in species detection of all replicates with the highest similarity between PCR replicates. In agreement with this, a Bayesian occupancy model demonstrated that species detection probability decreases in the PCR, while replicates earlier in the workflow were more effective in increasing species richness detection compared to PCR replicates or further increasing sequencing depth. Generally, 4-8 replicates were sufficient to recover 50% of all species, which contributed over 90% of the reads. Based on our results, we recommend at least eight replicates for biodiversity assessments using SPM. These should include at least one subsample or extraction replicate to maximize species richness. Our study supports the future use of SPM samples in biodiversity monitoring and provides a framework for assessing other eDNA matrices for robust species detection.