Scholarlink

Background and objective <p>In the molecular genetic context, deconvolution describes the derivation of individual DNA profiles from a&#xa0;DNA mixture. Whereas mixture deconvolution is conventionally carried out manually by an expert, there are various probabilistic software applications to automate the process and two of these applications, EuroForMix (EFM) and EFMrep, are used in numerous forensic laboratories. They are particularly helpful when trace material is degraded or PCR artifacts are present and multiple replicates, if applicable from different DNA extracts, are combined for a better assessment. The aim of this study was to evaluate software-assisted deconvolution (EFM and EFMrep) with respect to correctly derived genotypes or alleles per profile in comparison to manual deconvolution based on a&#xa0;combination of at least two different DNA extracts per case.</p> Material and methods <p>To this end, 16&#xa0;cases from former routine work were selected as examples and re-evaluated with EFM and EFMrep. In all cases, the manually derived profiles were incomplete due to the complexity of the underlying mixtures but still generated a&#xa0;hit in the German DNA Analysis Database (DAD). Likelihood calculations were performed for each match and the corresponding mixtures before the respective profiles were accepted as ground truth and the reference for deconvolution.</p> Results and conclusion <p>The results show that both forms of software are a&#xa0;useful addition to manual deconvolution, although they cannot replace it. Thus, when run in parallel they can objectively support the expert when there is uncertainty about an allele or genotype. In several cases it was shown that alleles were not correctly derived despite taking the threshold for the probability of results suggested in the current recommendations into account.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

DNA mixture deconvolution using fully continuous models EuroForMix and EFMrep

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年1月20日

查看全文
  • Roberta Schiller,
  • Jan Fleckhaus,
  • Dagmar von Máriássy,
  • Verena Brune,
  • Birgit Bayer,
  • Katja Anslinger

摘要

Background and objective

In the molecular genetic context, deconvolution describes the derivation of individual DNA profiles from a DNA mixture. Whereas mixture deconvolution is conventionally carried out manually by an expert, there are various probabilistic software applications to automate the process and two of these applications, EuroForMix (EFM) and EFMrep, are used in numerous forensic laboratories. They are particularly helpful when trace material is degraded or PCR artifacts are present and multiple replicates, if applicable from different DNA extracts, are combined for a better assessment. The aim of this study was to evaluate software-assisted deconvolution (EFM and EFMrep) with respect to correctly derived genotypes or alleles per profile in comparison to manual deconvolution based on a combination of at least two different DNA extracts per case.

Material and methods

To this end, 16 cases from former routine work were selected as examples and re-evaluated with EFM and EFMrep. In all cases, the manually derived profiles were incomplete due to the complexity of the underlying mixtures but still generated a hit in the German DNA Analysis Database (DAD). Likelihood calculations were performed for each match and the corresponding mixtures before the respective profiles were accepted as ground truth and the reference for deconvolution.

Results and conclusion

The results show that both forms of software are a useful addition to manual deconvolution, although they cannot replace it. Thus, when run in parallel they can objectively support the expert when there is uncertainty about an allele or genotype. In several cases it was shown that alleles were not correctly derived despite taking the threshold for the probability of results suggested in the current recommendations into account.