Reliability and reproducibility in targeted eDNA detection: the implications of species distributions and genetic diversity

Abstract

Environmental DNA (eDNA) sampling is an established tool for biodiversity studies. The design of assays to detect species’ DNA in samples underpins this tool. Although much progress has been made to develop validation frameworks for assay performance, an in silico validation framework is sorely missing. A conceptual framework was developed to incorporate the elements of assay design and likely outcomes of assay performance through a step-by-step decision tree evaluating in silico sensitivity and specificity. Using an in silico PCR tool, the sensitivity and specificity of published assays were evaluated to determine the potential ability of an assay to amplify and detect known target and non-target species sequences and place them within the conceptual framework. Results showed that most published assays tested exhibited the potential for Type I and/or Type II errors, with very few assays showing unique species specificity. Results also demonstrate that in silico validation can predict potential problematic non-target species co-amplification. The developed conceptual framework demonstrates the ability to provide a robust and reproducible in silico validation step highlighting key components of the design stage that need to be published to improve transparency, repeatability, reproducibility, and confidence in eDNA sampling.