TrueAllele solves uninterpretable DNA in mother and daughter double homicide

Back to Presentations

Getting more from less: low-level DNA mixtures on cartridges

Kari R. Danser, MS, Jennifer M. Bracamontes, MS, Megan M. Foley, MS, Matthew M. Legler, BS, and Mark W. Perlin, PhD, MD, PhD, "Getting more from less: low-level DNA mixtures on cartridges", Mid-Atlantic Association of Forensic Scientists Annual Meeting, Pittsburgh, PA, 10-May-2024.


PowerPoint presentation of the conference talk.

PDFDownload Handout
PPTDownload PowerPoint


How much identification information can be recovered from firearm cartridges? Our study examined DNA data from different casing materials and collection methods. On the same STR data, we compared TrueAllele® computer interpretation with simple allele counting. TrueAllele measured more information and found previously unidentified contributors.


Cartridge casings are the empty shells left behind after a gun is fired. They are typically made of different metals and have different calibers. Nearly 200,000 cartridge cases are recovered annually at United States crime scenes. 1 Crimes where cartridges can be collected include homicide, aggravated assault, robbery, and gang-related crime. It has been shown that fired cartridges degrade any remaining DNA and contain significantly less DNA 2 . Items with less DNA make it harder for crime labs to interpret the data, or for crime scene investigators to collect DNA.

The cartridge study was designed to determine how much single source identification information can be recovered from firearm cartridges. The design included finding the best collection method and the most informative cartridge type. Our study examined DNA data from seven different casing materials for a total of 910 cartridge samples. The cartridge samples were then collected using five different DNA collection methods. Once the DNA was collected, both manual and “probabilistic” genotyping interpretation, TrueAllele ® Casework, were used to analyze the cartridge data.

Once the single source cartridge samples were created, a George Washington University forensic lab manually interpreted the DNA data. They used a peak-height threshold and an allele counting method. The laboratory counted how many alleles matched the reference that were expected to be present in the sample. The laboratory found the reference sample was present in 205 of the cartridge samples.

A TrueAllele-trained analyst set up single source DNA interpretation requests for the computer system to solve. Upon review of the data, the cartridge data was seen to contain low-level mixtures. There were 202 low-level samples. 431 mixtures were present, comprising 47% of the samples.

Additional contributors were present. Since these mixtures were an unexpected finding, a TrueAllele analyst conducted further computer processing. They considered multiple contributors, assuming the number of contributors observed in the data. These mixtures contained 2-5 contributors.

TrueAllele found more DNA dentification information than human review. The computer could examine the lower-level data and mixtures; it found a previously unidentified contributor. This unknown profile was informative, with a Kullback-Leibler (KL, or expected genotype match statistic) of 30.36 ban. As measured by likelihood ratio (LR), the unknown contributor was found in 138 of the samples, across all seven cartridge types. Manual interpretation did not assess this unknown person, since that method focused solely on the reference’s allele pair, limiting data interpretation.

Based on the results of the study, the known reference was found in 351 cartridge samples using TrueAllele, compared to 251 samples using manual interpretation. The unknown person was found in 138 samples using TrueAllele, while manual interpretation didn’t find the unknown person.

Based on KL genotype information values, the most informative collection methods were wet:wet and wet:dry. The least informative collection methods were scraping and soak and sonicate. Based on KL and LR, the most informative cartridge types were aluminum and steel. The least informative cartridge type was 45 fired.

In conclusion, TrueAllele computer interpretation was better than manual review at extracting DNA information from cartridge casings. This improvement was especially notable with cartridges containing low-level and mixture DNA.


  1. “Shelling out Evidence: NIST Ballistic Standard Helps Tie Guns to Criminals.” NIST, 23 Jan. 2023, ballistic-standard-helps-tie-guns-criminals.

  2. Prasad, Elisha, et al. “Touch DNA recovery from unfired and fired cartridges: Comparison of swabbing, tape lifting and soaking.” Forensic Science International, vol. 330, Jan. 2022, p. 111101,