Reducing Artifacts and Uncertainty in Environmental DNA Monitoring

Genetic analysis of environmental DNA (eDNA) provides site occupancy inferences for aquatic macrofauna that are often easier to obtain than direct observations of organisms. Research on the ‘ecology of eDNA’ from aquatic macrofauna - i.e., the origin, state, transport, and fate - is needed to describe the spatiotemporal context for eDNA-based occupancy inferences, and to guide eDNA sampling design. One often overlooked origin is contamination. Contamination sources, primarily PCR product and tissue, are abundant in most laboratories with interest and capacity to analyze eDNA. Unfortunately, standard precautions (i.e., autoclaving) are inadequate for destroying DNA contamination. Even worse, standard negative controls (i.e., water blanks) are inadequate for observing contamination occurrence and abundance. Having learned some of these old lessons the hard way, I present several case studies involving eDNA data that I ultimately concluded were best explained by contamination. These conclusions are supported by research in other fields that make similar inferences about organisms, populations, and communities based solely on indirect genetic evidence from environmental samples: microbiology, forensics, paleogenetics, fecal source tracking, and ecotoxicology. I suggest a framework for accurately estimating contamination occurrence/abundance and demonstrate its application in the context of a mechanistic modeling approach to understanding the ecology of eDNA.