Forensic Science International: Reports (Nov 2021)

Developmental validation of a software implementation of a flexible framework for the assignment of likelihood ratios for forensic investigations

  • Hannah Kelly,
  • Zane Kerr,
  • Kevin Cheng,
  • Maarten Kruijver,
  • Jo-Anne Bright

Journal volume & issue
Vol. 4
p. 100231

Abstract

Read online

The developmental validation of the software DBLR™ (‘database likelihood ratios’) is described. DBLR™ implements a flexible framework for the assignment of likelihood ratios for forensic DNA profiles. This framework allows the user to assign a likelihood ratio for virtually any conceivable scenario using any combination of single-source or mixed DNA profiles. Propositions are defined by constructing pedigrees and then linking DNA typing results to pedigree individuals. Complex kinship scenarios including incestuous relationships can be evaluated. Additional functionality within DBLR™ allows the user to explore the discriminatory power of any DNA profile (Explore Deconvolution), search a profile against a database of known profiles using flexible search strategies (Search Database), investigate whether two mixtures share one or more common donors (Mixture to Mixture), and search a database of known profiles for an assumed common donor between two or more profiles (Common Donor). Developmental validation was carried out in accordance to the FBI Quality Assurance Standards for Forensic DNA Testing Laboratories (2020) and included functional and reliability testing as well as accuracy, precision, sensitivity, and specificity studies. The results of developmental validation demonstrate that each of the functions within DBLR™ is performing as expected.

Keywords