Use of fully continuous models for biostatistical evaluation of forensic DNA analytical findings Experiences of the project group "Biostatistical DNA calculations"

The biostatistical evaluation of DNA profiles supports the courts in the assessment of the evidential value of a DNA stain. In practice, an increasing number of stains characterized by minimal amounts of DNA and possible allelic drop-in and drop-out as well as complex mixtures are analyzed. Often such complex DNA profiles cannot, or only to a limited extent, be evaluated biostatistically with a traditional binary approach of interpretation.The development of fully continuous models (FCM) renders a large number of these previously incalculable DNA profiles accessible for biostatistical evaluation. Almost all available information of the DNA profile is included into this calculation. While these probabilistic genotyping methods are widely used internationally, only few experiences have become known in German-speaking countries.In order to understand the functionality, possibilities and limitations of FCM calculations, mixed DNA stains of known composition were comparatively analyzed with four currently available software products. The evaluation examined central aspects of FCM, such as concordance of calculation results, influence of drop-in and drop-out events on the calculated fully continuous likelihood ratios (LRfc) values, as well as the derivation of database-suitable DNA profiles derived from probability-based predictions (deconvolution).The experiences gained in this study, together with other already published data, provide a basis for recommendations on the use of fully continuous probabilistic genotyping software in the biostatistical evaluation of DNA profiles.