5- Conclusion
Here, we provide a thorough examination of the influence of PCR replication and read sampling depth on two common measures of diversity in environmental DNA metabarcoding research: α diversity and β diversity. We find that metabarcoding PCR robustly estimates measures of diversity that rely on high abundance taxa, and that low numbers of PCR replicates are sufficient to distinguish sites from each other. However, we also find that low abundance taxa tend to occur in only one or a few replicate PCRs, and often require deeper sequencing of PCR amplicon pools to be counted, in particular when read cutoff thresholds are high. Importantly, the challenges of counting rare taxa are so great, in particular in environmentally diverse sites, that even large numbers of PCR replicates and relatively deep sequencing are insufficient to guarantee that they will be recorded. Experiments aiming to catalogue diversity should consider using replicate PCRs, replicate extracts, and even replicate field sampling. While this problem necessarily limits the utility of environmental DNA as a mechanisms to catalogue the full diversity of taxa present at a given site, it also shows the potential power of eDNA to recover even the rarest of taxa, although the authenticity of exceptional taxa may need to be explored using a different enrichment approach, such as hybridization capture or targeted amplification.