STR profiles obtained from trace samples are often incomplete and subject to uncertainty due to stochastic effects, including PCR artifacts, inhibitors and locus-specific degradation. To improve the reliability of observed alleles, consensus profiles are commonly generated based on replicate PCR amplifications. These profiles can be compared with available reference profiles and support inference of unknown contributor genotypes. Especially for low-template traces and complex mixtures, manual comparison of alleles across multiple amplifications is time-consuming and error-prone.

To enable automated and reproducible consensus profile generation, we present SmarTRace, a novel and user-friendly Shiny application implemented in R. SmarTRace processes standard electropherogram data across different autosomal and Y-chromosomal STR kits. Consensus profiles are generated using a deterministic approach based on user-defined criteria. Additionally, a probabilistic approach models peak heights and stutter peaks using gamma distributions and incorporates PCR counts to estimate locus-specific allele probabilities. SmarTRace compares the generated consensus profiles with reference profiles applying user-defined parameters, resulting in multilingual reports to support forensic casework. Finally, SmarTRace infers major contributors of mixtures and quantifies these with statistical support values. All observed alleles and peak heights, inferred consensus profiles, comparison results and inferred major contributors are exported as customizable csv files. SmarTRace was validated using electropherogram data from two to four PCR replicates of trace samples ranging from touch DNA to blood stains, generated with standard STR kits PowerPlex® ESX and ESI 17 Fast Systems (both Promega) and NGM Detect™ PCR Amplification Kit (ThermoFisher) and analyzed using GeneMapper™ ID-X 1.7.3 (ThermoFisher). For comparisons between trace and reference samples, SmarTRace consistently outperformed manual analysis in accuracy and speed. For cases flagged by SmarTRace, manual inspection remains recommended to avoid wrongful exclusion of contributors due to unlabeled peaks at stutter positions or below analytical thresholds.

Major contributors were extracted from low complexity mixtures with varying contributor ratios from 1:1/1:1:1 to 50:1/50:50:1 and with total DNA inputs from 12.5 to 1000 pg. While major contributors were successfully identified in most mixtures, challenging conditions - including increased number of contributors, low DNA input and balanced mixture ratios - resulted in errors that were reliably flagged by SmarTRace.

To conclude, the open-source tool SmarTRace supports forensic geneticists by reducing manual error, facilitating major component identification and increasing objectivity and efficiency in STR profile comparison, while maintaining expert review for critical cases.

Authors

• Zehra Köksal (Department of Forensic Genetics, Institute of Forensic Medicine, University of Basel, Switzerland)
• Hannah Seifried (Department of Forensic Genetics, Institute of Forensic Medicine, University of Basel, Switzerland)
• Sarah Kron (Department of Forensic Genetics, Institute of Forensic Medicine, University of Basel, Switzerland)
• Simon Egger (Department of Forensic Genetics, Institute of Forensic Medicine, University of Basel, Switzerland)
• Janine Schulte (Department of Forensic Genetics, Institute of Forensic Medicine, University of Basel, Switzerland)
• Iris Schulz (Department of Forensic Genetics, Institute of Forensic Medicine, University of Basel, Switzerland)