Prasetyo Andhika P, Cusa Marine, Murray Joanna M, Agung Firdaus, Muttaqin Efin, Mariani Stefano, McDevitt Allan D
School of Science, Engineering and Environment, University of Salford, Salford, UK.
Centre Fisheries Research, Ministry for Marine Affairs and Fisheries, Jakarta, Indonesia.
iScience. 2023 Jun 7;26(7):107065. doi: 10.1016/j.isci.2023.107065. eCollection 2023 Jul 21.
Trade restrictions for endangered elasmobranch species exist to disincentivise their exploitation and curb their declines. However, trade monitoring is challenging due to product variety and the complexity of import/export routes. We investigate the use of a portable, universal, DNA-based tool which would greatly facilitate monitoring. We collected shark and ray samples across the Island of Java, Indonesia, and selected 28 commonly encountered species (including 22 CITES-listed species) to test a recently developed real-time PCR single-assay originally developed for screening bony fish. In the absence of a bespoke elasmobranch identification online platform in the original FASTFISH-ID model, we employed a deep learning algorithm to recognize species based on DNA melt-curve signatures. By combining visual and machine-learning assignment methods, we distinguished 25/28 species, 20 of which were CITES-listed. With further refinement, this method can improve monitoring of the elasmobranch trade worldwide, without a lab or species-specific assays.
对濒危板鳃亚纲物种实施贸易限制，旨在抑制对它们的开发利用并遏制其数量减少。然而，由于产品种类繁多以及进出口路线复杂，贸易监测颇具挑战性。我们研究了一种便携式、通用的基于DNA的工具的使用情况，该工具将极大地便利监测工作。我们在印度尼西亚爪哇岛采集了鲨鱼和鳐鱼样本，并挑选了28种常见物种（包括22种列入《濒危野生动植物种国际贸易公约》的物种），以测试一种最近开发的实时聚合酶链反应单检测法，该检测法最初是为筛选硬骨鱼而开发的。在原始的FASTFISH-ID模型中缺乏定制的板鳃亚纲物种识别在线平台的情况下，我们采用了一种深度学习算法，根据DNA熔解曲线特征来识别物种。通过结合视觉和机器学习分类方法，我们区分出了28种中的25种，其中20种是列入《濒危野生动植物种国际贸易公约》的物种。经过进一步完善，这种方法可以在无需实验室或特定物种检测法的情况下，改进全球范围内对板鳃亚纲贸易的监测。
