A Mixed DNA Profile Controversy: Corrigendum

Marie Semaan¹, Sarah Abbas^1,2 and Issam Mansour^1*
*Correspondence: Issam Mansour, Department of Laboratory Science and Technology, American University of Science and Technology, Beirut, Lebanon, Tel: +9611218716, Email:

Keywords

Forensic DNA • Mixed traces • Mixture software • Inbreeding • Likelihood ratio

Introduction

Traces involving DNA mixtures are frequently encountered in forensic casework [1]. Such traces originate from two or more contributors [2]. Interpretation of mixed traces is facilitated when 1. The amount of amplified DNA from all contributors is sufficient and above the analytical threshold, 2. The ratio of DNA contributed by each source is reflected on the peak heights and consequently the possible genotypes of major and minor contributors may be determined [3], 3. There is no degradation and allele drop out, 4. No artifacts and allele drop in and 5. The contributors are unrelated and have few shared alleles. But since the situation of forensic cases is rarely ideal and the likelihood in which it is not possible to distinguish the alleles of the different contributors does exist, several models have been suggested for the interpretation of mixed DNA profiles [4].

The present report describes a challenging case where eight POIs could not be excluded from a mixed DNA trace and highlights the fact that semicontinuous expert DNA mixture software might not be sufficient in populations with increased relatedness even if inbreeding coefficient is considered, when major pieces of information such as peak heights and interpretation strategies (such as conditioning) that allow better specificity and sensitivity in mixture cases are not accounted to.

Case Note

The present case is a mixed DNA trace. Profiling was performed with 23 STR systems by combining two multiplex STR kits: PowerPlex^®16 HS, the PowerPlex^® ESI 17 (Promega Corporation; Madison, WI, USA) (Table 1).

Systems D3S1358, D5S818, TH01, D21S11, TPOX, D7S820, D2S1338, Penta D and D1S1656 (highlighted in blue) and systems D8S1179, SE33 and D12S391 (highlighted in yellow) show that at least two individuals contributed to this trace.

Only two alleles per locus appear in systems D19S433, vWA, D13S317, FGA, D16S539, D18S51, CSF1PO, Penta E, D22S1045 and D2S441 (highlighted in green), which is potentially due to homozygosity and allele sharing [5] characteristics that are frequent in the Lebanese population due increased inbreeding in the community under study [6-8]. Allele drop out could be a further reason to whether only two or more profiles contributed to the mixture obtained from the trace.

Eight POIs are to be considered, if all other non-DNA evidence relevant to the allegation failed to provide an alibi [9]. Before assigning the number of contributors (NOC) to the mixture and formulating any proposition for interpretation, all eight DNA profiles showed a complete match with the mixed DNA trace (Table 2).

Statistical interpretation was performed based on the semi-continuous model using the DNA mixture analysis expert software LRmix studio, under the following proposition LR_iu/uu.

where,

H_iu: DNA comes from Pi and an unknown unrelated person U.

H_uu: DNA comes from two unknown people U, U unrelated to P_i or each other.

The allele frequencies of the Lebanese population were considered, and the analysis was repeated under the following situations: with and without the inbreeding coefficient Ɵ that reflects the rate of inbreeding in the Lebanese population 0.01.

If the NOC is limited to two contributors only and drop in is not accounted to, the obtained LR for POI 1 was 0 due to system D12S391. The obtained Likelihood ratios for the other POIs ranged between 10¹⁸ and 10²² (Table 2). The statistical results confirmed the inclusion of seven POIs and none of the POIs could be excluded by any of the 23 tested loci.

The number of systems was increased to 28 by adding the systems of the PCR CS7 kit (Promega Corporation; Madison, WI, USA) to the previously used PowerPlex^®16 HS, PowerPlex^® ESI 17 and PowerPlex^® kits.

By increasing the number of tested loci, three POIs were still not excluded (Table 3a). The statistical results supported the proposition that the POIs had contributed to the mixture rather than an unknown person for three POIs and none of the individuals could be excluded by any of the 28 tested loci (Table 3b).

Efforts in the recent decade have been made to assess the degree of uncertainty in the analysis of STR profiles, in particular the mixed DNA profiles. In undertaking such a study, it was important to know the true contributors of the DNA mixture before matching it against the database.

517 samples were collected randomly from Lebanese villages of different religious backgrounds. Profiling was performed with three different profile sizes: 15, 23 and 28 STR systems using three multiplex STR kits: PowerPlex^® 16 HS, the PowerPlex^® ESI 17 and the PowerPlex^® CS7 (Promega Corporation; Madison, WI, USA). The obtained profiles were used to simulate DNA mixtures of two contributors. Each of the 517 profiles was then probed against the electronically simulated two-contributor mixtures (the same was done with each of the 15, 23 and 28 profile sizes). Mixture analysis was performed, and Likelihood Ratios were generated whenever false inclusions were detected.

In the present case, profiles of individuals #3 and #6 were the true contributors of the mixture.

DNA mixture analysis software (based on peak designation only) yielded LR values of 2 x 10¹⁹, 6.4 x 10²², 1.4 x 10¹⁸, 2 x 10²¹ and 2.9 x 10²² for POIs 2, 4, 5, 7 and 8, respectively with 23 locus profiles that supported the propositions that they contributed to the mixture rather than an unrelated person. In some cases, the LRs were larger than the LR values of the true contributors. Since we do have background information on these POIs we intended to consider the below hypothesis:

H_iu: DNA comes from P_i and an unknown person U.

H_uu: DNA comes from an unknown individual U and R who is a relative to P_i.

And the LRs were as follows:

LR2u/ur=4.9 x 10⁶, LR4u/ur=1.2 x 10⁶, LR5u/ur=5.0 x 10⁵, LR7u/ur=5.3 x 10⁶, LR8u/ur=6.6 x 10⁶

The only situation where ground truth was supported by the LR values is when we considered conditioning and used continuous model PG software.

Discussion and Conclusion

While increasing the number of STRs successfully helped excluding individuals 4, 5, 7 and 8, it strongly supported the false inclusion of individual 2 with a higher LR value, thus results supporting the proposition that was not ground truth (however, in this case both propositions were false, as the ground truth involved a close relative).

Interpreting DNA mixtures is among the most challenging forensic casework and cases like the one presented may be encountered in inbred populations. Equally important to using the relevant allele frequency dataset [10,11] and taking into account the inbreeding coefficient, is the choice of the DNA mixture analysis software and of the propositions. As apparent in this work, DNA mixture analysis software that uses semi-continuous models (i.e. do not account to peak heights) can falsely support the inclusion of noncontributors.

Presenting the DNA evidence without statistics renders the DNA evidence inadmissible. However, these findings challenge the admissibility of semicontinuous DNA mixture statistics and raise the attentiveness to forensic DNA mixture inclusive conclusions when dealing with communities with high-level of inbreeding. To reduce the rate of false inclusions, information based on peak heights and the use of conditioning profiles [12], whenever possible, are better alternatives [13].

References

1. Torres, Yolanda, Inmaculada Flores, Victoria Prieto and Manuel López-Soto, et al. "DNA mixtures in forensic casework: a 4-year retrospective study." Forensic Sci Int 134 (2003): 180-186.

Google Scholar, Crossref, Indexed at

2. Bieber, Frederick R., John S. Buckleton, Bruce Budowle and John M. Butler, et al. "Evaluation of forensic DNA mixture evidence: protocol for evaluation, interpretation, and statistical calculations using the combined probability of inclusion." BMC Genetics 17 (2016): 1-15.

Google Scholar, Crossref, Indexed at

3. Gill, Peter, Charles H. Brenner, John S. Buckleton and Angel Carracedo, et al. "DNA commission of the International Society of Forensic Genetics: recommendations on the interpretation of mixtures." Forensic Sci Int 160 (2006): 90-101.

Google Scholar, Crossref, Indexed at

4. Ladd, Carll, Henry C. Lee, Nicholas Yang and Frederick R. Bieber. "Interpretation of complex forensic DNA mixtures." Croatian Med J 42 (2001): 244-246.

Google Scholar, Indexed at

5. Thompson, William C., Laurence Mueller and Dan Krane. "Forensic DNA statistics: Still controversial in some cases." The Champion 36 (2012): 12-23.

Google Scholar, Indexed at

6. Al-Azem, M., A. El Andari, and I. Mansour. "Estimation of allele and haplotype frequencies for 23 YSTR markers in the Lebanese population." Forensic Res Criminol Int J 5 (2017): 00150.

Google Scholar, Crossref, Indexed at

7. Abbas, S., L. Mourad and I. Mansour. "Evaluation of Sibling-Ship Analysis in Secluded Lebanese Villages with Increased Mating Patterns." J Forensic Investigation 6 (2018): 4.

Google Scholar, Indexed at

8. El Andari, Ansar and Issam Mansour. "Is he or is he not the father?: Pros, Cons effect of following standardized interpretations." Forensic Sci Int: Genetics Supplement Series 4 (2013): e186-e187.

Google Scholar, Crossref

9. Nichols, Richard A. "Interpreting DNA Evidence: Statistical Genetics for Forensic Scientists." Heredity 82 (1999): 585-586.

Google Scholar, Crossref, Indexed at

10. El Andari, Ansar, Lama Mourad and Issam Mansour. "Genetic variations and population data on five supplementary STR markers in Lebanon." Ann Hum Genet 83 (2019): 82-85.

Google Scholar, Crossref, Indexed at

11. El Andari, Ansar, Hiba Othman, Franco Taroni and Issam Mansour. "Population genetic data for 23 STR markers from Lebanon." Forensic Sci Int: Genet 7 (2013): e108-e113.

Google Scholar, Crossref, Indexed at

12. Hicks, Tacha, ane Kerr, Simone Pugh and Jo-Anne Bright, et al. "Comparing multiple POI to DNA mixtures." Forensic Sci Int: Genet 52 (2021): 102481.

Google Scholar, Crossref, Indexed at

13. Kalafut, Tim, Simone Pugh, Peter Gill and Sarah Abbas, et al. "A mixed DNA profile controversy revisited." J Forensic Sci 67 (2022): 128-135.

Google Scholar, Crossref, Indexed at