Statistical analysis tools of mixture DNA samples: When the same software provides different results

• Tipo: Artigo em Revista Científica Internacional
• Data: 2022

Autores

Costa C. (Autor) Outra; Figueiredo C. (Autor) Outra; Amorim, A (Autor) FCUP; Prieto L. (Autor) Outra; Costa S. (Autor) Outra; Ferreira P.M. (Autor) Outra; Pinto N. (Autor) Outra

Revista

Título: Forensic Science International: Genetics Supplement Series

Vol. 8

Páginas: 37-39

ISSN: 1875-1768

Editora: Elsevier

Indexação

Scopus - 0 Citações

Outras Informações

ID Authenticus: P-00X-9XQ

DOI: 10.1016/j.fsigss.2022.09.014

Abstract (EN): The high complexity of the genetic analysis of crime scene samples is mainly related to the unknown number of contributors, low DNA quantity and quality, and associated stochastic effects. The difficulty and subjectivity of interpreting casework samples was the motto for the development of software to mitigate these conditions and allow the quantification of the genetic evidence. Currently, there are several tools for statistical analysis of mixture samples based on either qualitative or quantitative models. The first considers the electropherograms¿ qualitative information, while the latter also considers the associated quantitative information. This work's main goal was to evaluate the effect that parameters¿ settings variation may have on the LR computation, specifically the drop-in frequency parameter. For that, a qualitative ¿ LRmix Studio ¿ and two quantitative software ¿ STRmix¿ and EuroForMix ¿ were considered and an intra-software analysis was performed, using as input real casework samples. The drop-in frequency variation showed an impact, leading to differences higher than four units (log10 scale) for some pairs of samples. In addition, for some cases, no comparisons were performed either because the tool computed a null LR value or displayed an error message. Thus, this work reinforces the importance of proper parameters¿ modeling and estimation in forensic casework evaluation.

Idioma: Inglês

Tipo (Avaliação Docente): Científica