RESUMEN
Taxon-specific quantitative PCR (qPCR) assays are commonly used for environmental DNA sampling-based inference of animal presence. These assays require thorough validation to ensure that amplification truly indicates detection of the target taxon, but a thorough validation is difficult when there are potentially many non-target taxa, some of which may have incomplete taxonomies. Here, we use a previously published, quantitative model of cross-amplification risk to describe a framework for assessing qPCR assay specificity when there is missing information and it is not possible to assess assay specificity for each individual non-target confamilial. In this framework, we predict assay specificity against unsampled taxa (non-target taxa without sequence data available) using the sequence information that is available for other confamilials. We demonstrate this framework using four case study assays for: (1) An endemic, freshwater arthropod (meltwater stonefly; Lednia tumana), (2) a globally distributed, marine ascidian (Didemnum perlucidum), (3) a continentally distributed freshwater crustacean (virile crayfish; Faxonius virilis, deanae and nais species complex) and (4) a globally distributed freshwater teleost (common carp; Cyprinus carpio and its close relative C. rubrofuscus). We tested the robustness of our approach to missing information by simulating application of our framework for all possible subsamples of 20-all non-target taxa. Our results suggest that the modelling framework results in estimates which are largely concordant with observed levels of cross-amplification risk using all available sequence data, even when there are high levels of data missingness. We explore potential limitations and extensions of this approach for assessing assay specificity and provide users with an R Markdown template for generating reproducible reports to support their own assay validation efforts.