Beyond Species Richness: A Statistically Rigorous Comparison of eDNA and Conventional Fish Sampling Methods.

The comparative performance of different fish sampling methods remains unclear due to a lack of statistically rigorous comparisons. This issue has been exacerbated in recent years due to environmental DNA (eDNA) metabarcoding becoming increasingly popular. However, limited research has compared multiple conventional methods simultaneously to eDNA metabarcoding. To address this, 12S metabarcoding, COI metabarcoding, electrofishing, hoop netting, seine netting, minnow trapping, and gill netting samples were taken at 10 sites across four waterbodies from July eighth to July 13th, 2023. Typically, during sampling method comparisons, diversity or species richness is used as the metric. However, relying solely on these metrics often leads to misinterpretation. In order to rigorously statistically compare the methods, four metrics were used: accuracy, consistency, diversity, and congruency. A novel integrative statistical approach for conducting methodological comparisons was conducted by utilizing the four metrics. By conducting a thorough statistical analysis, a variety of novel insights are highlighted. Specifically, hoop netting was accurate when compared to seine netting and gill netting, COI was accurate when compared to seine netting, seine netting was consistent with COI and hoop netting, eDNA (COI + 12S) found higher levels of species richness, and COI was congruent with four conventional methods. When the specific results are synthesized, three main findings were discovered: (1) COI outperformed 12S on three out of the four metrics, (2) seine netting is largely redundant in the presence of COI, and (3) eDNA captures a more complete picture of biodiversity and is best compared to a combination of conventional methods; however, in order to accomplish this, multiple markers need to be utilized (12S + COI). However, the findings also raise concerns with respect to the validity and accuracy of 12S results, suggest that a combination of methods need to be utilized in the majority of situations, and question the overall reliability of eDNA metabarcoding. Ultimately, we recommend that researchers conduct rigorous statistical analyses, moving beyond superficial comparisons, to develop numerical evidence to support and strengthen their conclusions.