Factors affecting DNA barcoding classification accuracy for piscine prey: An experimental assessment with invasive lionfish

Abstract

The use of DNA barcoding in fish diet studies is becoming more widespread, but the effect of prey digestion on barcoding accuracy has been poorly studied. We conducted a series of controlled feeding experiments, with red lionfish (Pterois volitans) as predators and Gulf killifish (Fundulus grandis) as prey, designed to test the effect of digestion time (12–42 h), prey item (sample) preservation (inside versus outside stomach), and sample preparation (washed versus unwashed) on DNA barcoding classification accuracy. Competitive interactions occurred between lionfish and killifish DNA during PCR amplification when universal COI primers were utilized; thus, primers were designed that matched the same priming site in both killifish and lionfish DNA. The proportion of amplified killifish DNA relative to lionfish DNA per sample was estimated using multiplex quantitative PCR (qPCR) and shown to decrease with longer digestion times, especially when samples were retained within lionfish stomachs for preservation. The effects of digestion and preservation method on amplifiable DNA resulted in significant effects on barcoding accuracy, resulting in erroneous identification of the predator (i.e., self-DNA) for nearly 25% of samples. Removing prey from predator stomachs prior to preservation significantly increased the proportion of amplified killifish DNA from samples, and enhanced barcoding accuracy. Overall, study results have important implications for the probability of false negatives for prey items, as well as false positives for self-DNA, when utilizing barcoding to characterize piscine diet. Results also highlight the need to conduct more method validation research given the increasing prevalence of DNA barcoding as a means to identify specific prey items in fish diet studies.