Statistical models for forensic DNA data

The need for thorough and accurate model frame works for evaluating forensic evidence continue to grow with the increased used of technical data to solve crimes. This project focus on extending the existing methodology in forensic genetics enabling the inclusion of more of the available data in the evaluation of the evidence. A part of this includes programming software applications for

useful to forensic geneticists.

Of particular interest is the development of software for DNA mixture analysis, which are cases where more than one individual has contributed to the biological material. The biological material may be exposed to e.g. bacteria and sunlight, which can cause breakages in DNA strand. Hence, the analysis is complicated further since the probability of partially observed DNA pro les increases. Models for the entire signal (electropherogram, EPG) obtained in the biotechnological analysis are appealing since issues of thresholding the EPG are avoided.