An examination of aspects of the probabilistic genotyping tool: Forensic Statistical Tool. Issue 3 (3rd October 2019)

• Record Type: Journal Article
• Authors: Gasston, Julia; Kruijver, Maarten; Curran, James M.; Bright, Jo‐Anne; Pugh, Simone; Buckleton, John S.
• Is Part Of: WIREs. Volume 2:Issue 3(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2019-10-03
• Subjects: code review -- Forensic DNA -- FST -- probabilistic genotyping; Forensic sciences -- Periodicals; Forensic sciences; Electronic journals; Periodicals; 363.25
• DOI: 10.1002/wfs2.1362
• Languages: English
• ISSNs: 2573-9468

Abstract

Abstract: The probabilistic genotyping software Forensic Statistical Tool implements a semi‐continuous model for DNA interpretation. This software omits any locus where the sum of the allele probabilities equals or exceeds .97. There has been criticism that this function is neither signaled by the software nor disclosed in publications. We investigate the effect of this function by creating a near clone of the model and applying it to five‐ and six‐allele loci for three‐person mixtures created in the expected population proportions. On average, the dropping of a locus is conservative for six‐peak loci and nonconservative for five‐peak loci. For persons of interest (POIs) with rare alleles, the dropping is usually conservative. For POIs with common alleles, the dropping of the locus is often nonconservative. This article is categorized under: Forensic Biology > Interpretation of Biological Evidence Forensic Biology > Forensic DNA Technologies Jurisprudence and Regulatory Oversight > Expert Evidence and Narrative Abstract : The probabilistic genotyping software Forensic Statistical Tool locus omission feature is demonstrated to be conservative for six‐peak loci and nonconservative for five‐peak loci within a three‐person mixture.