Tumbling Dice Ltd, Newcastle upon Tyne, UK.
Pest Manag Sci. 2018 Mar;74(3):705-714. doi: 10.1002/ps.4766. Epub 2017 Nov 29.
This paper describes the progress that we have made in assessing the feasibility of 'fingerprinting' using imaged SDS-PAGE gels of haemolymph proteins, to identify soft-bodied wood-boring insect larvae such as the Asian longhorn beetle, Anoplophora glabripennis (Motscholsky, 1853) (Coleoptera: Cerambycidae). Because of stringent import restrictions and difficulty in obtaining licences to work with these organisms, we opted to work with four species of scarab beetle, Mecynorhina polyphemus (Fabricius, 1781), Pachnoda sinuata (Fabricius, 1775), Eucidella shiratica (Csiki, 1909) and Eucidella shultzeorum (Kolbe, 1906) which have near identical larval morphologies.
We show that this technology when combined with an advanced pattern matching system (Digital Automated Identification SYstem - DAISY) can classify soft-bodied insect larvae that are almost identical morphologically to species at a level of accuracy is in excess of 98%. The study also indicates that the technology copes well with noisy data and small training sets.
The experience gained in undertaking this study gives us confidence that we will be able to develop a field deployable system in the medium term. We believe that as a high-throughput identification tool, this technology is superior to competitor technologies (e.g. fingerprinting of imaged DNA gels) in terms of speed, cost and ease of use; and therefore, is suitable for low-cost deployment in the field. © 2017 Society of Chemical Industry.
本文描述了我们在评估使用 SDS-PAGE 凝胶图像进行“指纹识别”的可行性方面所取得的进展,以识别亚洲长角天牛(Anoplophora glabripennis(Motscholsky,1853))(鞘翅目:天牛科)等软体蛀木昆虫幼虫。由于严格的进口限制和获得这些生物工作许可证的困难,我们选择了四种金龟子,包括多纹丽金龟(Mecynorhina polyphemus(Fabricius,1781))、沟胫天牛(Pachnoda sinuata(Fabricius,1775))、台湾弧丽金龟(Eucidella shiratica(Csiki,1909))和台湾弧丽金龟(Eucidella shultzeorum(Kolbe,1906)),它们的幼虫形态非常相似。
我们表明,这项技术与先进的模式匹配系统(Digital Automated Identification SYstem - DAISY)结合使用,可以对形态几乎相同的软体昆虫幼虫进行分类,其准确性超过 98%。该研究还表明,该技术能够很好地处理嘈杂数据和小训练集。
开展这项研究所获得的经验使我们有信心能够在中期内开发出一种可现场部署的系统。我们相信,作为一种高通量识别工具,与竞争对手技术(例如成像 DNA 凝胶的指纹识别)相比,该技术在速度、成本和易用性方面具有优势;因此,非常适合在现场进行低成本部署。© 2017 化学工业学会。