【摘要】 目的利用GC-MS检测死后家猪脑组织的代谢产物,推断死亡时间。方法成年家猪大脑,置于25℃、75%RH的人工气候箱中,于0 h-84 h内,每间隔6 h取材,GC-MS检测各时间点脑组织代谢物变化。结果PCA显示:平台期和窗口期的时间组彼此分开。建立PLS模型,通过VIP> 1且Kruskal-Wallis检验（P <0.05）筛选出18种重要的代谢物,线性回归模型和参数检验均有统计学意义。多元回归方程为:Y_PMI=6.610+16.29X_十八烷酸+14.56X_{5-氨基缬草酸}+5.517X_丙氨酸（R²=0.909、SE=6.323）或Y_PMI=15.78+9.690 X_{5-氨基缬草酸}+86.45X_亮氨酸-82.35X_甘氨酸（R²=0.952、SE=4.271）。结论 GC-MS检测出家猪死后脑组织的多种产物与PMI存在显著相关性,证实了其理论和技术推断PMI的可行性。综合多指标多元逐步回归分析和PLS-DA等多元模式分析方法建立PMI推断模型,可提高推断模型的准确性和精确度。更多还原

【Abstract】 Objective To apply GC-MS to detect metabolites in brain of swines and to study the estimation of postmortem interval（PMI）. Methods The adult swine brains were placed in a 25℃ and 75% humidity artificial climate box, the brains were then collected every 6 h of the interval from 0 to 84 h and analyzed with GC-MS. Results PCA model showed that the groups of the platform periods and window periods are separated from each other. The PLS model was established, and 18 important metabolites were screened through Kruskal-Wallis test with VIP > 1 and P < 0.05., and the linear regression model and parameter test were statistically significant. The multiple regression equation is Y_PMI= 6.610+ 16.29 X_Sa+ 14.56 X_{5-Aminovaleric acid}+ 5.517 XAla（R² = 0.909、SE = 6.323） or Y_PMI= 15.78 + 9.690 X_{5-Aminovaleric acid} + 86.45 X_Leu-82.35 X_Gly（R² = 0.952、SE = 4.271）. Conclusion There is a significant correlation between products detected by GC-MS and PMI, confirming the feasibility of the GC-MS theory and technology in inferring the PMI. Integrated multi-indicator multiple stepwise regression analysis and multi-modal analysis methods such as PLS-DA to establish a PMI inference model, could improve the accuracy and veracity of the PMI inference.更多还原