Estimation of parameters in DNA mixture analysis

Abstract

In [7], a Bayesian network for analysis of mixed traces of DNA was presented using gamma distributions for modelling peak sizes in the electropherogram. It was demonstrated that the analysis was sensitive to the choice of a variance factor and hence this should be adapted to any new trace analysed. In this paper, we discuss how the variance parameter can be estimated by maximum likelihood to achieve this. The unknown proportions of DNA from each contributor can similarly be estimated by maximum likelihood jointly with the variance parameter. Furthermore, we discuss how to incorporate prior knowledge about the parameters in a Bayesian analysis. The proposed estimation methods are illustrated through a few examples of applications for calculating evidential value in casework and for mixture deconvolution.

Keywords:

• Bayesian network
• forensic identification
• Markov chain Monte Carlo methods
• DNA mixture deconvolution

Acknowledgements

We are grateful to Kjell Konis for advice and flexibility in adapting RHugin to serve the purpose of this analysis and to Robert Cowell and Julia Mortera for helpful comments on a previous version of the manuscript.