Separation of DNA mixtures

Mixtures of DNA often appear in relation to crime cases. In our modern societies the a propor analysis of such cases are of great importance.  The aim of this PhD project is to investigate this problem from a statistical point of view in order to develop a better methodology to model the DNA mixture problem. The project is a continuation of my MSc project where I applied a missing data approach based on a multivariate normal distribution. The models developed in the MSc thesis were limited in application to real crime cases. The often arising problems of stutters, drop-ins and -outs, contamination and degradation of the DNA and pull-up effects are important issues of the future work. The current models are formed to fit STR DNA data, but newer standards and techniques such as SNP DNA will also be taken into consideration. Multiple contributor mixtures with more than two donors occur frequently in real crime cases (e.g. gang rapes). Extensions to cope with more than two donors with some unknown is therefore also central in terms of future work.