Mallette Jennifer R, Casale John F, Jordan James, Morello David R, Beyer Paul M
U.S. Drug Enforcement Administration, Special Testing and Research Laboratory, Dulles, VA 20166 USA.
National Geospatial Intelligence Agency, Springfield, VA 20150 USA.
Sci Rep. 2016 Mar 23;6:23520. doi: 10.1038/srep23520.
Previously, geo-sourcing to five major coca growing regions within South America was accomplished. However, the expansion of coca cultivation throughout South America made sub-regional origin determinations increasingly difficult. The former methodology was recently enhanced with additional stable isotope analyses ((2)H and (18)O) to fully characterize cocaine due to the varying environmental conditions in which the coca was grown. An improved data analysis method was implemented with the combination of machine learning and multivariate statistical analysis methods to provide further partitioning between growing regions. Here, we show how the combination of trace cocaine alkaloids, stable isotopes, and multivariate statistical analyses can be used to classify illicit cocaine as originating from one of 19 growing regions within South America. The data obtained through this approach can be used to describe current coca cultivation and production trends, highlight trafficking routes, as well as identify new coca growing regions.
此前，已实现了对南美洲五个主要古柯种植区的地理溯源。然而，古柯种植在南美洲各地的扩张使得确定次区域来源变得越来越困难。由于古柯生长的环境条件各不相同，最近通过额外的稳定同位素分析（氘和氧-18）改进了之前的方法，以全面表征可卡因。结合机器学习和多元统计分析方法实施了一种改进的数据分析方法，以在种植区之间进行进一步划分。在此，我们展示了如何将痕量可卡因生物碱、稳定同位素和多元统计分析相结合，用于将非法可卡因归类为源自南美洲19个种植区中的某一个。通过这种方法获得的数据可用于描述当前古柯种植和生产趋势、突出贩运路线以及识别新的古柯种植区。
