From anarchy to clarity, data pre-processing and statistical choices influence quantitative environmental DNA (eDNA) analyses

Abstract

Environmental DNA (eDNA) analyses hold great potential for increasing species detection sensitivities and estimating species abundances. The rapidly growing user base, continuous method development and optimisation have led to diverse approaches for capturing and analysing eDNA. While significant efforts have been made to standardise field and laboratory protocols, a notable gap remains in understanding the consequences of data pre-processing and statistical choices on the final results obtained, particularly in quantitative eDNA analyses. These insights are crucial for developing best-practice guidelines that can harmonise analytical workflows. To address this gap, we conducted an extensive literature review focusing on quantitative species-specific eDNA studies. We assessed the diversity of data pre-processing and statistical choices made to evaluate the correlation between eDNA concentrations and species' abundance or biomass, and collected the raw datasets when available. We then applied commonly used data analysis strategies to the datasets to formulate general recommendations for improving the reliability and reproducibility of quantitative eDNA analyses. Our results indicate that, within the available literature, statistical methods are not always clearly described and raw data are rarely made publicly available. Furthermore, the choice of data pre-processing strategies and statistical tests used to assess quantitative correlations can significantly influence the likelihood of detecting positive correlations and the effect sizes. Overall, we recommend the following: (i) increase transparency in method descriptions and data availability; (ii) assess correlations using mixed-effect models that can account for data characteristics; (iii) avoid pre-processing quantitative eDNA data, especially when combined with sub-optimal statistical tests. Implementing these guidelines should enhance the accessibility and transparency of quantitative eDNA data and ultimately their use for managers and policy makers.