2. Re-analysis of the time-series changes in eDNA particle size distribution
Our meta-analyses provided new insights into the relationship between eDNA persistence and its state. We then re-analysed the dataset from a previous tank experiment (Jo et al., 2019b) to estimate mitochondrial and nuclear eDNA decay rates at multiple size fractions and water temperature levels. The results of the re-analysis appeared to be generally consistent with those of the meta-analyses; as indicated by the meta-analyses, eDNA persistence depended on the interactions between its size fraction, type of the target gene, and water temperature (Table 3; Figure 4). In particular, a significant interaction between filter pore size and temperature indicated that inflow of the degraded, larger-sized eDNA into smaller size fractions could buffer the effect of temperature on eDNA degradation in these smaller size fractions, as described in previous sections. The dependence of eDNA degradation on water temperature would likely be smaller when targeting smaller-sized eDNA or using a smaller pore size filter.
Some recent studies attempted to estimate species biomass and abundance by integrating quantitative eDNA analysis and hydrodynamic modelling, allowing the consideration of eDNA dynamics, such as its production, transport, and degradation (Carraro et al., 2018; Tillotson et al., 2018; Fukaya et al., 2020). For a more accurate estimation, environmental parameters affecting these eDNA dynamics may be included in the statistical modelling framework. The effect of temperature on eDNA degradation can be minimized during statistical modelling by considering eDNA particles at smaller fraction sizes, which will allow simplification of the modelling procedure while retaining its accuracy and reliability. However, considering the apparent suppression of eDNA degradation in smaller size fractions, owing to the inflow of the degraded larger-sized eDNA, it is possible that such smaller-sized eDNA yield ‘older and less fresh’ biological signals than the larger-sized eDNA. Such non-fresh eDNA signals can result in false-positives during eDNA detection (Yamamoto et al, 2016; Jo et al., 2017), in which case the use of eDNA particles in the smaller size fractions would be disadvantageous for eDNA-based biomass or abundance estimation. The applicability of smaller-sized eDNA for such estimations can be verified by comparing the correlation between eDNA quantification and species biomass and abundance, and the availability of longer eDNA fragments among the filter pore sizes or eDNA particle sizes, for which meta-analyses such as the present study may be suitable.