用骨骼肌mRNA变化推断损伤时间探索  

An exploratory study on wound age estimation of skeletal muscle injury under the influence of postmortem interval

作　　者：刘元新 张帅 梁芯瑞 郗延如 王英元[1] 孙俊红[1] 曹洁[1] 靳茜茜[1] 杜秋香[1] Liu Yuanxin;Zhang Shuai;Liang Xinrui;Xi Yanni;Wang Yingyuan;Sun Junhong;Cao Jie;Jin Qianqian;Du Qiuxiang(School of Forensic Medicine,Shanxi Medical University,Jinzhong 030600,China)

机构地区：[1]山西医科大学法医学院,山西晋中030600

出　　处：《中国法医学杂志》2022年第3期261-266,271,共7页Chinese Journal of Forensic Medicine

基　　金：国家自然科学基金面上项目(81971795);山西省优秀青年基金项目(201901D211351)。

摘　　要：目的研究大鼠骨骼肌损伤后mRNA表达水平的变化趋势及其是否受死亡时间影响,并筛选代表性指标建立损伤时间推断的数学模型,为死亡时间影响下的损伤时间推断提供新的方法和思路。方法将168只SD大鼠随机分为对照组和损伤组(4 h、8 h、12 h、16 h、20 h、24 h,n=24只),各组进一步分为三个死亡时间亚组(死后0 h、12 h、24 h,n=8只),建立大鼠骨骼肌挫伤及死后模型;使用Real-time PCR检测Slfn3、Rcc1l、Lrrc41、Trit1、Impact、Gpnmb 6个基因的mRNA表达情况;使用SPSS 25.0相关分析(Pearson和Spearman法)评价mRNA指标表达水平与损伤时间和死亡时间的相关关系;建立损伤时间Fisher判别模型并用留一法(Leave-One-Out,LOO)评估模型准确性。结果不同mRNA受死亡时间影响程度不同;经相关分析,剔除与损伤时间无关的Trit1、Lrrc41,损伤时间Fisher判别模型预测准确率有所提升;重新分组后,模型判别准确率为91.7%,可有效区分对照组、损伤后4 h~8 h、12 h~16 h、20 h~24 h组。结论应用相关分析筛选与损伤时间相关指标,建立损伤时间的Fisher判别模型,可以有效弱化死亡时间对mRNA的影响,从而在不区分死亡时间的情况下推断损伤时间;结果显示模型性能良好,验证了本研究筛选指标所用方法的有效性,为死亡时间影响下的损伤时间推断提供新的方法和思路。Objective Based on the different temporal changes of mRNA expression levels in rat skeletal muscle after contusion and death,the representative indicators were selected,and the mathematical models of wound age estimation was established,providing new methods and ideas for the screening and application of indicators in forensic research.Methods A total of 168 SD rats were randomly divided into control group and contusion group(4 h,8 h,12 h,16 h,20 h,24 h,n=24).All group were further divided into three postmortem interval(PMI)subgroups(0 h,12 h,24 h postmortem,n=8).The skeletal muscle contusion and postmortem rat model was established.The expression levels of 6mRNAs(Slfn3,Rcc1l,Lrrc41,Trit1,Impact,Gpnmb)in rats’skeletal muscle were detected by Real-time PCR.SPSS 25.0correlation analysis(Pearson and Spearman)were used to evaluate the correlation between target mRNA expression level and wound age or PMI.Fisher discriminant analysis(FDA)models for wound age estimation were established,and LeaveOne-Out(LOO)was used for internal cross-validation to evaluate the accuracy of the models.Results Different mRNAs was affected by PMI to different degrees.The prediction accuracy of the Fisher discriminant model for inference of wound age was improved when Trit1,Lrrc41 indicators unrelated to wound age were eliminated.Further combining wound age,the accuracy of model was improved to 91.7%,which could effectively distinguish the control group,4 h~8 h,12 h~16 h,20 h~24 h.Conclusion The Fisher discriminant model of wound age was established by using correlation analysis to screen the indicators related to the wound age,which can effectively weaken the in?uence of PMI on mRNA,so that the wound age could be inferred without distinguishing the PMI.The results showed that the model had good performance,which verified the effectiveness of the method used to screen the indicators in this study,and provided a new research idea for the screening and application of mRNA indicators for inference of wound age under the in?uence of PMI.

关 键 词：法医病理学 损伤时间推断 相关分析 骨骼肌挫伤 mRNA指标筛选 

分 类 号：D919.4[医药卫生—法医学]