AI结合multiplex-HRM进行海南岛法医昆虫种属鉴定的基础研究

法医昆虫学是目前已知最有希望解决晚期死亡时间推断的技术，其中以形态学鉴别为金标准的法医昆虫种属鉴定是其核心环节及应用瓶颈，分子种属鉴定是其重要补充手段。本研究通过建立海南岛法医昆虫形态摄像大数据库，利用基于深度学习的人工智能技术(Artificial Intelligence ,AI)，探索实现法医昆虫形态学种属鉴定的难题，解决目前法医昆虫学研究中昆虫形态学种属鉴别人才不足的困境。同时，探索将简单、快速、准确、低成本的高分辨率溶解曲线分析技术（high resolution melting，HRM）及DNA barcoding技术应用于法医昆虫学研究，在此基础上探索建立一次实验即可实现多位点检测的multiplex HRM-Barcoding技术体系的构建，解决形态学难以区分的疑难法医昆虫的高分辨率种属鉴定，为将推动法医昆虫学在海南岛内进入实践应用做出重要基础性探索和积累。

Forensic entomology is the most promising technology to determine the time of late death, among which morphological identification is the core link and application bottleneck, and molecular species identification is an important supplement. In this study, by establishing a large database of forensic insect morphology images in Hainan Island, and using the artificial intelligence technology based on deep learning, the problem of forensic insect morphology species identification is explored, and the problem of insufficient personnel in forensic entomological species identification is solved. At the same time, the high resolution melting (HRM) and DNA barcoding techniques, which are simple, fast, accurate and low-cost, are applied to the research of forensic entomology. On this basis, the multiplex that can be used to detect multiple DNA loci can be established by one experiment. The construction of HRM barcoding technology system may solve the high-resolution species identification of insects which are difficult to distinguish morphologically, and make important basic exploration and accumulation for promoting the practical application of forensic entomology in Hainan Island.