A forensic genetics laboratory can be described as carrying two broad scientific responsibilities [1]: To produce genetic data able to maximally discriminate forensically relevant hypotheses; and to report the value of them. The branch of forensic science dedicated to improving the quantity of genetic information has supported advances in mega-plex panels that simultaneously target more than 20 forensically relevant markers, the emergence of NGS in forensics, and the development of novel collection devices that recover more biological material from a substrate. With these practical advances also came improvements to the way in which data were interpreted and included the adoption of Bayesian reasoning by forensic scientists, articulation of a hierarchy of propositions, and the implementation of probabilistic genotyping. 

With the interpretive framework being mostly constructed, attention is being paid to efforts seeking to appraise the consistency of evaluations within and across forensic science service providers (FSSPs), as was done in [2, 3]. The findings show that, in the main, mixture interpretation is subject to sometimes impactful effects originating from service provider’s policy decisions on matters pertaining to NoC and suitability [3] or the laboratory treatments and settings used to generate the data [4]. 
 

What makes forensic DNA data diverge can, arguably, be traced back to the laboratory processes used to generate it. Here, the DNA molecules carrying allele information at targeted genetic locations are extracted while cells from, potentially, a high number of donors are still mixed. What generally follows is the fractionation of the extract into at least two portions: one that is stored and one that is amplified. Being that the number of copies of a given allele within a volume fraction varies, the extract fraction carried to PCR conveys an unknown and variable number of DNA molecules of given allele type. The DNA molecules that are amplified and tagged with fluorophores are then detected. The result is a fluorescence signature that is a superposition of allele and artifact signal from numerous donors whose signal may not be fully represented and can be further obfuscated by noise. Application of signal thresholds can serve to exacerbate these divergences. 

If generating as much useful information is as valued as making the best interpretive use of that data, the question then becomes: Is it possible to uncover what laboratory treatments give maximal amounts of relevant information for a given technology? If so, can the level of useful information across laboratories be similar despite differences in platforms and assays? In this workshop we attend to these questions. 
 

The workshop is structured as follows: To begin, we review counting techniques, relevant definitions, and known distributions like the binomial distribution. Then we apply the concepts to predict the number of amplified DNA molecules of a given type. Next, we convert these numbers to a fluorescence, finally producing a distribution of peak heights for different extract fractions carrying an unknown number of target DNA molecules. We explore these distributions to examine if it is possible to uncover data generating procedures from which we receive maximal levels of genetic information across the broadest number of donors for a single amplification, regardless of platform or assay. Lastly, the group will explore the implications of the findings and discuss their impacts in light of the recent report in Forensic DNA Interpretation and Human Factors [5].

References

[1] W. Goodwin, A. Linacre, and S. Hadi, An Introduction to Forensic Genetics, 2nd ed. Wiley, 2010.

[2] L. M. Brinkac, N. Richetelli, J. M. Davoren, R. A. Bever, and R. A. Hicklin, "DNAmix 2021: Laboratory policies, procedures, and casework scenarios summary and dataset,"

Data in Brief, vol. 48, p. 109150, 2023/06/01/ 2023, doi: https://doi.org/10.1016/j.dib.2023.109150.

[3] R. A. Hicklin, N. Richetelli, B. L. Emerick, R. A. Bever, and J. M. Davoren, "Variation in assessments of suitability and number of contributors for DNA mixtures," Forensic Science International: Genetics, vol. 65, 2023, doi: 10.1016/j.fsigen.2023.102892.

[4] K. C. Peters, H. Swaminathan, J. Sheehan, K. R. Duffy, D. S. Lun, and C. M. Grgicak, "Production of high-fidelity electropherograms results in improved and consistent DNA interpretation: Standardizing the forensic validation process," Forensic Science International: Genetics, vol. 31, pp. 160-170, 2017, doi: 10.1016/j.fsigen.2017.09.005.

[5] Expert Working Group in Forensic DNA Interpretation, "Forensic DNA Interpretation and Human Factors: Improving Practice Through a Systems Approach," in "Human Factors in Forensic Sciences Expert Working Group Series," National Institute of Standards and Technology, 2024.