Accounting for inter-laboratory DNA recovery data variability when performing evaluations given activities

Duncan Taylor; Luke Volgin; Peter Gill; Bas Kokshoorn

doi:10.1016/j.fsigen.2025.103283

Accounting for inter-laboratory DNA recovery data variability when performing evaluations given activities

Forensic Sci Int Genet. 2025 Apr 8:78:103283. doi: 10.1016/j.fsigen.2025.103283. Online ahead of print.

Authors

Duncan Taylor¹, Luke Volgin², Peter Gill³, Bas Kokshoorn⁴

Affiliations

¹ Forensic Science SA, GPO Box 2790, Adelaide, SA 5001, Australia; School of Biological Sciences, Flinders University, GPO Box 2100, Adelaide, SA 5001, Australia. Electronic address: Duncan.Taylor@sa.gov.au.
² Forensic Science SA, GPO Box 2790, Adelaide, SA 5001, Australia.
³ Forensic Genetics Research Group, Department of Forensic Sciences, Oslo University Hospital, Oslo, Norway; Department of Clinical Medicine, University of Oslo, Oslo, Norway.
⁴ Forensic Trace Dynamics, Faculty of Technology, Amsterdam University of Applied Sciences, Amsterdam (AUAS), the Netherlands; Netherlands Forensic Institute, The Hague, the Netherlands.

PMID: 40239301
DOI: 10.1016/j.fsigen.2025.103283

Abstract

A large, recently published, inter-laboratory study by the ReAct group has shown that there is considerable variability in DNA recovery that exists between forensic laboratories. The presence of this inter-laboratory variability presents issues when one laboratory wishes to carry out an evaluation and needs to use the data produced by another laboratory. One option proposed by the ReAct group is for laboratories to carry out a calibration exercise so that appropriate adjustments between laboratories can be made. This will address some issues, but leave others unanswered, such as how to make use of the decades of transfer and persistence data that has already been published. In this work we present a method to utilise data produced in other laboratories (whether it provides DNA amounts or a probability of transfer) that takes into account inter-laboratory variability within an evaluation. This will allow evaluations to continue, without calibration data, and ensures that the strength of findings is appropriately represented. In this paper we discuss complicating factors with the various ways in which previous data has been reported, and their limitations in supporting probability assignments when carrying out an evaluation. We show that a combination of producing calibration information for new data (as suggested by the ReAct group) and development of strategies where calibration data is not available will provide the best way forward in the field of evaluations given activities.

Keywords: Activity level evaluation; Activity level propositions; Bayesian Network; DNA recovery; DNA transfer; ReAct project.