Data from: Assessing vertebrate biodiversity in a kelp forest ecosystem using environmental DNA

Preserving biodiversity is a global challenge requiring data on species’ distribution and abundance over large geographic and temporal scales. However, traditional methods to survey mobile species’ distribution and abundance in marine environments are often inefficient, environmentally destructive, or resource-intensive. Metabarcoding of environmental DNA (eDNA) offers a new means to assess biodiversity and on much larger scales, but adoption of this approach for surveying whole animal communities in large, dynamic aquatic systems has been slowed by significant unknowns surrounding error rates of detection and relevant spatial resolution of eDNA surveys. Here, we report the results of a 2.5 km eDNA transect surveying the vertebrate fauna present along a gradation of diverse marine habitats associated with a kelp forest ecosystem. Using PCR primers that target the mitochondrial 12S rRNA gene of marine fishes and mammals, we generated eDNA sequence data and compared it to simultaneous visual dive surveys. We find spatial concordance between individual species’ eDNA and visual survey trends, and that eDNA is able to distinguish vertebrate community assemblages from habitats separated by as little as ~60 m. eDNA reliably detected vertebrates with low false-negative error rates (1/12 taxa) when compared to the surveys and revealed cryptic species known to occupy the habitats, but overlooked by visual methods. This study also presents an explicit accounting of false negatives and positives in metabarcoding data, which illustrate the influence of gene marker selection, replication, contamination, biases impacting eDNA count data, and ecology of target species on eDNA detection rates in an open ecosystem.保护生物多样性是一项全球性挑战，亟需获取大地理尺度与时间尺度下的物种分布与丰度数据。然而，传统海洋环境中移动物种分布与丰度调查方法往往存在效率低下、破坏生态环境或资源消耗过大的问题。环境DNA（environmental DNA, eDNA）元条形码（metabarcoding）技术为生物多样性评估提供了全新途径，且可适配更大尺度的调查工作，但在大型动态水生系统中开展全动物群落调查时，该方法的推广应用却因检测误差率与eDNA调查相关空间分辨率的诸多未知因素而受阻。本研究报道了一段2.5千米的eDNA样带调查结果，该样带用于调查与海藻林生态系统相关的多样海洋生境梯度中分布的脊椎动物群落。本研究使用靶向海洋鱼类与哺乳类线粒体12S核糖体RNA基因的聚合酶链式反应（PCR）引物，获取了eDNA序列数据，并将其与同步开展的水下视觉调查结果进行对比。研究发现，单个物种的eDNA检测结果与视觉调查的空间分布趋势具有一致性，且eDNA能够区分间距仅约60米的不同生境中的脊椎动物群落组成。与视觉调查相比，eDNA检测脊椎动物的假阴性错误率较低（1/12类群），检测结果可靠；同时还发现了已知栖息于该生境但被视觉调查忽略的隐秘物种。本研究还对元条形码数据中的假阴性与假阳性结果进行了明确量化分析，阐明了基因标记选择、重复实验、污染、影响eDNA计数数据的偏差，以及目标物种生态学特性对开放生态系统中eDNA检测率的影响。