Application Effectiveness Evaluation And Module Development In The SepDNA Software For DNA Mixture Analysis

Objective: The most common DNA mixture appears in rape cases, which usually consists of the sperm(male) and vaginal fluid(female). Currently a variety of methods have been used to separate the male and female components from the DNA mixture. However, in some gang rapes sometimes appears the mixed sperm of two or more males, or in some murder scenes appears the mixed blood or mixed epithelial cells. How to identify different individuals from the DNA mixture is an obstacle in DNA mixture analysis.Since 1998, the appearance of softwares for the DNA mixture separation has improved the DNA mixture analysis in biostatistics and provided a basis for the case investigation. However, there have no software for the separation and analysis of DNA mixture developed indenpendently in our country. Thus,this kind of development is of realistic significance.Previously, we has developed a software called sep DNA for the analysis of DNA mixture in our laboratory, which contains two STR detection modules- IDfiler(specific to the Identifiler kit) and Yfiler(specific to the Yfiler kit),and two mathematical models- Bayes and Iter. It can be used to separate two-male and three-male mixed DNA in theory. Through a joint Bayes-Iter analysis and comprehensive consideration, we can infer the source of DNA mixture and achieve personal identification. Therefore, our research focused on the sep DNA software in the evaluation of application and the module addition in order to expand the application range.Methods:1 Part of Application Effectiveness EvaluationCollect DNA data from Beijing Municipal Public Security Bureau and Shanxi Provincial Public Security Department which have been solved andinvolved the mixed DNA profiles(containing the mixed STR data and the corresponding known individual STR data). After analyzing in the sep DNA software, we compare the analysis results with the corresponding known individual STR data and then get the separation accuracy of the software.2 Part of Module DevelopmentExtract DNA from blood samples of fifty unrelated male individuals from Hebei Province Blood Center using the Pure Link? Genomic DNA Kit(Life Technologies in America). After building the two-male mixed models and three-male mixed models in different mixed ratios, we amplify these DNA models by PCR and conduct electrophoresis by 3130 genetic analyzer with three commercial kits called Amp F?STR?Identifiler? Plus Kit, Amp F?STR?Minifiler TM Kit and Goldeneye? DNA Identification System 20 A Kit. The parameters assessed include APH/APA, Mx, Hb and Ci. After analyzing these model data in the sep DNA software, we compare the analysis result with the known individual DNA datas by using some statistical analysis methods(eg.Kruskal-Wallis H test, Student-Newman-Keuls test, Mc Nemar test, Pearson Chi-square test and Dunnett's test), and then get the separation accuracy of three commercial kits.Results:1 Part of Application Effectiveness EvaluationBoth the overall separation accuracy and the locus-separation accuracies of Iter model were lower than Bayes model. Through joint Bayes-Iter analysis and comprehensive consideration, the separation accuracies have increased,especially after considering both the Best and Better Genotypes. We picked out the cases of which the average peak height is higher than 1250 rfu and then found that the separation accuracies were higher than all the cases.2 Part of Module DevelopmentWith the decrease of the average peak height, D value increased in the Identifiler Plus kit, while 20 A and Minifiler kits were not the case. Average peak height and average peak area showed a linear relationship in all modules of three different commercial kits. The differences of heterozygote balanceratio in different loci and different mixed ratios were almost of no statistical significance. Overall, Ci of Identifiler Plus and Minifiler Kits were better than20 A kit, and Ci in loci with blue and green fluorescence were higher than red.After a joint Best-Better analysis, the separation accuracy of Bayes model in two-male mixed models of Identifiler Plus kit was 93.5%. Overall, the separation accuracy of Identifiler Plus kit is higher than 20 A kit, which is higher than Minifiler kit. However, the separation accuracy of Minifiler kitis higher than Identifiler Plus kitat CSF1 PO andis higher than 20 A kitat D2S1338.The separation accuracy of 20 A kit in three-male mixed modelsis higher than others.Conclusions:1 We can use the sep DNA software to separate the mixed DNA in actual cases. Without regard to drop-out, the separation accuracy of the sep DNA software in actual cases can reach more than 70%. If the average peak height in an actual case is higher than 1250 rfu, after a joint Bayes-Iter analysis and a joint Best-Better analysis, the separation accuracy can reach 90%.2 The two-male and three-male mixed models are used to add three modules(Identifiler Plus, Minifiler and20A) to sep DNA software in order to expand the range of software application, so that the sep DNA software can be more widely used in the mixed DNA separation.