The forensic DNA field seems perpetually faced with the challenge of labor intensive and time-consuming casework alongside the criminal justice community’s demand for more timely analysis of evidence from crime scenes. Over time, forensic analysis methods for DNA amplification and STR profiling have greatly increased in sensitivity, allowing for samples containing degraded or trace amounts of DNA, such as touch DNA, to be analyzed. Touch DNA samples are those that result from the transferal of minute amounts of biological material from the body when contact is made with a surface [1, 2]. These touch DNA samples often contain low levels of template DNA (less than 100pg available for STR amplification), which becomes problematic during PCR as some target regions may be preferentially amplified over others [3]. This can result in STR allelic drop in, drop out, or peak imbalances [4] in the resulting profiles, making their interpretation more difficult. This process is further complicated because these touch DNA samples are inherently prone to mixtures – i.e. samples that include DNA from more than one contributor – since they originate from surfaces that may have been touched by numerous individuals [5].

Unfortunately, in the current forensic DNA workflow, the number of contributors in a sample is not revealed until the final step of analysis – a process that can easily take weeks from start (initial sample evaluation and serology) to finish (case report issued). When a mixture is present, along with low amounts of DNA, resulting data often includes one or more of the contributors’ allele peaks falling below the analytical threshold often leading to “inconclusive” reporting. Additionally, because this information is not available until endpoint analysis, it is not possible to make earlier analytical adjustments to the protocols or workflow that may serve to increase the likelihood of generating a profile with a distinguishable minor contributor. While reamplification of a low, mixed DNA sample may be possible, it is time consuming. Additionally, with low template or touch samples, the samples are more often consumed during initial testing leaving little or no remaining DNA for a second analysis. Due to these potential complications, many laboratories elect not to routinely accept, process, or interpret touch DNA samples [6]. A screening assay that could detect the contributor nature of a sample and potentially provide early exclusionary information earlier in the forensic DNA workflow would give analysts more options for processing low-level touch DNA samples, which may, in turn make laboratories to be less reluctant to process them. Knowledge of the presence of a mixture (versus a single source sample) could redirect the DNA workflow in an effort to improve the number of passing profiles obtained during the first round of testing, subsequently reducing retest rates.

The ability to distinguish between single source and mixed biological samples may be most easily introduced at the real time PCR (qPCR) stage. Real time PCR instruments combine the functions of a thermal cycler and fluorometer; they are routinely used by forensic DNA laboratories to detect the amount of amplifiable human DNA in a sample, using commercialized qPCR kits. Regardless of the exact model used, their multi-channel fluorescent detection, coupled with amplicon melt capabilities, allows for an opportunity to integrate additional assays into existing quantification kits [7,8,9,10] – thereby providing more information about evidentiary samples, earlier in the workflow. The Applied Biosystems™ 7500 (ABI 7500, Thermo Fisher, Waltham, MA) is the most frequently used qPCR instrument in the forensic community historically due to its broad acceptance of producing reliable and accurate data [8]. It is a five-color platform that allows for customizable melt curve protocols to be performed and analyzed with its melt curve analysis (MCA) software [11, 12]. The Rotor-Gene® Q (QIAGEN, Hilden, Germany) is another instrument that offers quantification, amplification, and high-resolution melt (HRM) curve analysis, and it is equipped with seven dye channels; one being the specially tuned extended green HRM channel. This, combined with its comprehensive HRM software, allows for a relatively simple melt curve analysis [13, 14]. The QuantStudio™ 5 system and 6 Flex system (Thermo Fisher) are qPCR platforms that have a decoupled six color and coupled five color filter set, respectively, that allows for the addition of custom dyes and post-PCR HRM [15]. Further, these QuantStudio™ models offer on-board melt curve analysis with its custom HRM software module [15, 16]. The QuantStudio™ models are newer than the Rotor-Gene® Q and are supported for use with several common forensic DNA quantification kits [17,18,19,20,21].

The availability of melt curve (or dissociation) analysis with most modern qPCR instruments affords a functionality that can be easily exploited to gain more information about a sample at this early stage. The melt curve function detects and measures the change in fluorescence of the amplified products over time as the temperature is slowly increased. Thus, as the temperature increases, the DNA dissociates (melts), corresponding to differences in DNA composition such as length and sequence, suggesting it may be possible to use these curves to differentiate between various alleles or genotypes at a given genetic locus and even distinguish mixtures [22,23,24,25,26,27]. For example, Kuehnert et al. determined that the intercalating dye EvaGreen® (Biotium, Freemont, CA) detected in the green channel at 510 nm, could be used to detect and distinguish single source STR amplicons from mixed contributor amplicons using the HRM channel of the Rotor-Gene® Q [23].

In fact, several previous studies have utilized STR loci D5S818 and D18S51 for development of HRM-based mixture detection assays [23, 26, 28, 29]. D5S818 has small amplicon sizes (115–178 bp) and a small range of repeats (6–18) whereas, the STR locus D18S51 has larger amplicons (262–342 bp) and a larger range of repeats (7–27) [26, 30]. The difference in amplicon length prevents overlap of the resulting D5S818 and D18S51 melt curves, allowing these loci to be easily duplexed (amplified and melted simultaneously in a single reaction). In conjunction with machine learning algorithms, such as linear discriminant analysis (LDA) and Support Vector Machines (SVM), this approach can be a very flexible technique for making sample contributor predictions. However, for practical application, an assay of this type would be best if incorporated into the existing qPCR (quantification) step, rather than as an additional, stand-alone process.

The Investigator Quantiplex® quantification kit (Quantiplex®, QIAGEN, Hilden, Germany) is a commercially available qPCR kit that quantifies human genomic DNA in a sample. Quantiplex® has only one target – a human target in the FAM dye channel - and an internal PCR control in the VIC dye channel. The goal of the work described herein was to establish foundational knowledge for the direct integration of a melt curve assay into a qPCR kit, the Quantiplex® quantification kit, as a way to provide additional data that could be used to assign evidentiary DNA samples to either a single source genotype or identify it as a mixture (containing DNA from multiple contributors). With the success of this work and the fundamental principles for an integrated qPCR-HRM assay outlined, this work has been used as a premise for more complex and informative integrated qPCR-HRM assays, specifically the integrated Quantifiler Trio™-HRM assay.