Forensic Science International: Genetics

Transcription

1 Forensic Science International: Genetics 28 (2017) Contents lists available at ScienceDirect Forensic Science International: Genetics journa l homepage: lsevier.com/locate/fsig Research paper Validation of a rapid DNA process with the RapidHIT 1 ID system using GlobalFiler 1 Express chemistry, a platform optimized for decentralized testing environments Susana Salceda, Arnaldo Barican, Jacklyn Buscaino, Bruce Goldman, Jim Klevenberg, Melissa Kuhn, Dennis Lehto, Frank Lin, Phong Nguyen, Charles Park, Francesca Pearson, Rick Pittaro, Sayali Salodkar, Robert Schueren, Corey Smith, Charles Troup, Dean Tsou, Mattias Vangbo, Justus Wunderle, David King* IntegenX Inc, 5720 Stoneridge Drive, Suite 300, Pleasanton, CA, 94588, USA A R T I C L E I N F O Article history: Received 29 September 2016 Received in revised form 11 December 2016 Accepted 9 January 2017 Available online 12 January 2017 Keywords: SWGDAM developmental validation Rapid DNA RapidHIT 1 RapidHIT 1 ID RapidLINK TM GlobalFiler 1 Express STR Presumed single source sample A B S T R A C T The RapidHIT 1 ID is a fully automated sample-to-answer system for short tandem repeat (STR)-based human identification. The RapidHIT ID has been optimized for use in decentralized environments and processes presumed single source DNA samples, generating Combined DNA Index System (CODIS)- compatible DNA profiles in less than 90 min. The system is easy to use, requiring less than one minute of hands-on time. Profiles are reviewed using centralized linking software, RapidLINK TM (IntegenX, Pleasanton, CA), a software tool designed to collate DNA profiles from single or multiple RapidHIT ID systems at different geographic locations. The RapidHIT ID has been designed to employ GlobalFiler 1 Express and AmpFLSTR 1 NGMSElect TM, Thermo Fisher Scientific (Waltham, MA) STR chemistries. The Developmental Validation studies were performed using GlobalFiler 1 Express with single source reference samples according to Scientific Working Group for DNA Analysis Methods guidelines. These results show that multiple RapidHIT ID systems networked with RapidLINK software form a highly reliable system for wide-scale deployment in locations such as police booking stations and border crossings enabling real-time testing of arrestees, potential human trafficking victims, and other instances where rapid turnaround is essential The Author(s). Published by Elsevier Ireland Ltd. This is an open access article under the CC BY-NC- ND license ( 1. Introduction Use of short tandem repeats (STR) has become the pillar of the worldwide forensic DNA community [1 3]. Currently, processing and analyzing samples require highly trained forensic analysts and technicians operating multiple instruments in a centralized laboratory setting. However, this system may not be able to respond expeditiously in some situations and such delays may lead to premature release of suspects linked to previous crimes. Decentralized environments (e.g. booking stations, border crossings and small-scale forensic labs) could benefit from faster timeto-results for such criminal suspects. A previous generation of Rapid DNA technology was developed to address this issue, and has enabled integrated sample-in to CODIS-ready profile-out * Corresponding author. address: davidk@integenx.com (D. King). workflow. The RapidHIT System (IntegenX, Pleasanton, CA) has been developmentally [4 6] and internally validated [7 14], and has been used to upload more than 1000 STR profiles into national DNA databases such as CODIS and the UK National DNA Database. A next-generation Rapid DNA instrument, RapidHIT 1 ID (IntegenX, Pleasanton, CA), has been developed specifically for the needs of decentralized environments, Fig. 1. The RapidHIT ID system is a fully automated, sample-to-codis file system for STRbased human identification. The RapidHIT ID processes presumed single source samples and generates CODIS-compatible STR profiles in less than 90 min with less than one minute of handson time. The system uses a low-cost, single-use cartridge for each sample tested, (Fig. 2a). Cartridges utilize GlobalFiler 1 Express (Thermo Fisher Scientific, Waltham, MA) or AmpFLSTR 1 NGMSElect TM kit (Thermo Fisher Scientific, Waltham, MA) multiplexes. Kit-dependent control cartridges are loaded with allelic ladders / 2017 The Author(s). Published by Elsevier Ireland Ltd. This is an open access article under the CC BY-NC-ND license (

2 22 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Materials and methods 2.1. DNA samples Fig. 1. RapidHIT ID System. Bulk reagents and the capillary electrophoresis module are housed in a 150-use consumable ( printer toner -like) primary cartridge (Fig. 2b). The sample and primary cartridges have a minimum of 6 months stability at temperatures up to 25 C. Once an STR profile is generated, the RapidHIT ID transfers the data to a central computer operating RapidLINK software (IntegenX) for processing, including manual profile review (Fig. 3). RapidHIT ID has a built-in fingerprint reader and camera for authenticating access and automatic audit tracking. RapidLINK also provides predictive analytics for consumable inventory management across multiple geographically-distributed RapidHIT ID systems. The results of testing the RapidHIT ID system with GlobalFiler 1 Express are reported herein for sensitivity, specificity, interfering substances, size precision, concordance and overall performance. Experiments were designed following Federal Bureau of Investigation (FBI) Quality Assurance Standards [15] and Scientific Working Group for DNA Analysis Methods (SWGDAM) guideline [16]. To establish correlation between the RapidHIT ID and conventional bench systems, two sets of buccal swab samples were collected from consenting donors using cotton-tipped swabs (Puritan Medical Products Company LLC, Guilford, ME). Each donor was instructed to swipe the inside of the cheek 5 times for RapidHIT ID samples and 10 times for the conventional bench processing and return the swab to the original paper package labeled with the date of collection and anonymized identification number. The swabs were stored at room temperature prior to processing on RapidHIT ID or used to generate reference STR profiles by traditional methods. Reference STR profiles were generated using the GlobalFiler 1 Express (Thermo Fisher Scientific) buccal swab protocol [17] described in Section To simplify concordance comparisons, a database of reference profiles was established. Positive control swabs were used for several studies, were prepared using human male fibroblast cell line HTB-157 (ATCC, Manassas, VA), designated 1000 M, and deposited on cottontipped swabs. Cells were stored at 80 C in 90% FBS plus 10% DMSO (Aragen Bioscience, Morgan Hill, CA). To prepare for use, the cells were thawed, washed and re-suspended twice in PBS buffer and the resulting solution was quantified for DNA concentration using a Scepter Handheld Automated Cell Counter (Millipore, Billerica, MA). A working cellular concentration of cells/ml was prepared for dilution. Thirty microliter aliquots of the appropriate dilution of cells were added to swabs, which were then air-dried overnight at room temperature. The 1000 M cell line is the same as component F in the National Institute of Standards and Technology (NIST, Gaithersburg, MD) DNA Profiling Standard SRM 2391c and the certified profile from NIST was used as the reference for concordance. DNA samples for species specificity studies comprised bovine, chicken, horse, porcine and rabbit obtained from BioChain (Newark, CA). Purified genomic DNA from several humanassociated microorganisms in the oral cavity also were tested (ATCC, Manassas, VA) DNA extraction On the RapidHIT ID, extraction of DNA from samples is performed in the swab chamber of the sample cartridge, Fig. 2(a). The sample is incubated at 75 C in 500 ml Prep-n-Go Fig. 2. RapidHIT ID consumables. (a) sample cartridge, (b) primary cartridge.

3 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 3. RapidLINK TM software. Enables data review and control of geographically dispersed RapidHIT ID systems. Buffer (Thermo Fisher Scientific). The Prep-n-Go buffer is dispensed into the sample cartridge from the primary cartridge, Fig. 2(b). A boundary study was performed on the incubation time to verify the performance of DNA extraction. The incubation times studied were 8 min, 10 min (standard condition) and 12 min. Three swabs of 1000 M with 100,000 cell-load were used for each tested condition Polymerase chain reaction (PCR) amplification, sample electrophoresis and data analysis Manual bench processing method Reference profiles for the buccal and 1000 M swabs were prepared by processing the swabs with Prep-N-Go followed by PCR on the GeneAmp PCR System 9700 (Thermo Fisher Scientific) thermal cycler using the GlobalFiler 1 Express PCR Amplification kit and separated on the 3130 l Genetic Analyzer (Thermo Fisher Scientific). In summary, cell lysates were generated by adding 300 ml of Prep-N-Go buffer to 1.5 ml Eppendorf tubes, inserting a sample swab and incubating at 70 C in a heating block for 15 min. PCR amplification of the lysate was prepared by combining 6 ml of GlobalFiler 1 Express primer set, 6 ml of master mix and 3 ml of cell lysate to give a total reaction volume of 15 ml, per the manufacturer s protocol [17]. Thermal cycling was performed on the GeneAmp PCR System 9700 (Thermo Fisher Scientific) with a 96-well gold-plated silver block. The following thermal cycling parameters were used in the 9700 max mode: enzyme activation at 95 C for 1 min, followed by 27 cycles of denaturation at 94 C for 3 s and annealing/extension at 60 C for 30 s. The final extension step was performed at 60 C for 8 min, followed by a final hold at 4 C until PCR products were used. After thermal cycling was complete, samples were prepared for capillary electrophoresis according to the manufacturer s protocol [17] using a 16 capillary 3130 l Genetic Analyzer. The electrophoresis results were analyzed using GeneMarker HID genotyping software (SoftGenetics) using the default analysis settings for GlobalFiler 1 Express chemistry RapidHIT ID system The RapidHIT ID sample cartridge has GlobalFiler 1 Express primer set and master mix preloaded in separate reagent vials (Fig. 2a). The sizing standard, based on MapMarker Dy base pair (bp) size standard (BioVentures, Inc., Murfreesboro, TN) with additional oligonucleotide fragments to enhance sizing accuracy, is also pre-loaded in a reagent vial. Lysis buffer and electrophoretic reagents for 150 sample cartridge runs are stored in the primary cartridge. Following DNA extraction, the lysate is moved within the sample cartridge from the swab chamber to the PCR chamber. The PCR chamber contains a 1.2 mm diameter paper disc and traps the lysate. Primer and master mixes are then simultaneously pumped into the PCR chamber. The volume of the PCR chamber is approximately 12 ml and the ratio of primer to master mix volume follows the manufacturer s recommendation [17]. Thermal cycling occurs within the PCR chamber. The thermal cycling parameters executed are enzyme activation at 95 C for 1 min, followed by 28 cycles of denaturation at 94 C for 5 s and annealing/extension at 60 C for 40 s. The final extension step is performed at 60 C for 8 min. Upon completion of thermal cycling, the size standard is pumped through the PCR chamber advancing the amplified product into a mix chamber where both size standard and amplified sample combine. The resulting mixture is then moved through a heat denaturation zone (95 C)

4 24 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) and into the primary cartridge (Fig. 2b), where it is injected into the electrophoresis capillary. The DNA fragments are separated and optically detected. The RapidHIT ID processes the raw data and generates an STR profile. Electrophoresis gel is replaced in the capillary prior to every sample or control run. Control cartridges containing GlobalFiler 1 Express allelic ladder are processed by the RapidHIT ID with the same protocol as a sample cartridge omitting the thermal cycling step. The RapidHIT ID signal processing pipeline trims primer peaks, performs multicomponent analysis for spectral overlap, baselines, filters noise, uniformly re-scales the fluorescence intensity of all data, and generates an electropherogram trace file in fsa file format. The fsa files are automatically imported into GeneMarker 1 HID AutoLite (SoftGenetics LLC, State College, PA) for allele calling. GeneMarker HID AutoLite employs a library of allelic ladders used to size and identify alleles from electropherograms. An allelic ladder is automatically selected by GeneMarker HID AutoLite for analysis of each sample. The ladder library for the RapidHIT ID comprises ladders from multiple RapidHIT ID systems as well as an allelic ladder derived from a control cartridge that was last recorded on that system. The analytical (AT) and stochastic thresholds (ST) are set on a per locus basis. The methodology for setting AT is derived from a receiver-operator curve approach following the procedure outlined by Rakay et al. [18]. For the data analyzed in this paper, the value of ST for a given locus is determined by dividing the locus AT by the smaller of 0.5 or the lowest peak height ratio observed at that locus listed in Table 3 (Min%). Final electropherograms and STR profiles were reviewed using RapidLINK software networked to the multiple RapidHIT ID systems used in this study. Throughout this study, STR profiles are referred to as complete if all 24 of the GlobalFiler 1 Express loci have alleles automatically identified by the signal processing pipeline and concordant if all the identified alleles are an exact match to those alleles identified using the reference method described in Section Thermal cycling parameters PCR cycling parameters were tested around the standard set of conditions selected for the GlobalFiler 1 Express assay on the RapidHIT ID system. Three 1000 M control swabs containing 100,000 cells were used to test each of the thermal cycling parameters. The following parameters were tested (bold font indicates the standard operating conditions): activation temperature: 93 C, 95 C, and 97 C; denaturation temperature: 92 C, 94 C, and 96 C; annealing/extension temperature: 58 C, 60 C, and 62 C; final extension time: 6 min, 8 min, and 10 min; cycle number: 27, 28, and 29 cycles (using the times described in 2.3.2) Mock inhibition study GlobalFiler 1 Express has been developed and optimized (Thermo Fisher Scientific) to allow direct amplification from buccal swabs treated with Prep-n-Go Buffer without the need for sample purification. Samples collected from the oral cavity may contain common inhibitors of PCR, such as caffeine and nicotine. The robustness of this method was challenged by spiking these two PCR inhibitors onto 1000 M control swabs containing 100,000 cells. Three replicates for each inhibitor concentration were tested. The inhibitors were prepared as follows: (1) brewed black coffee at volumes of 2 ml, 10 ml, 50 ml and 100 ml was added directly onto the 1000 M control swabs; (2) tobacco: 2.5 g of chewing tobacco were mixed with 25 ml of water, ground in a mortar and allowed to soak for four hours. The tobacco slurry was stored overnight at room temperature, and then 2 ml, 10 ml, 50 ml and 100 ml were added directly to the 1000 M control swabs. Table 1 Effect of tobacco on control swabs (n = 3 per condition) with 100, M cells on the percent of alleles detected. Tobacco Concentration % Total Alleles Detected none 100.0% 2 ml 100.0% 10 ml 100.0% 50 ml 98.4% 100 ml 85.7% 2.6. Species specificity Specificity was tested using DNA from five non-primate sources as well as pooled DNA from six microorganisms. For the nonprimate DNA samples (bovine, chicken, horse, porcine and rabbit), a mixture was made comprising 8 ng of non-primate DNA together with PCR master and primer mixes consistent with the reagent volume used in the sample cartridge. Sample cartridges were then filled with this mixture following standard manufacturing processes. These modified sample cartridges were then run as blank samples on a RapidHIT ID. For the microorganisms (Fusobacterium nucleatum, Lactobacillus casei, Streptococcus mitis, Streptococcus mutans, Streptococcus salivarius and Enterococcus faecalis), the same protocol was followed replacing non-primate DNA with pooled bacterial DNA consisting of 10 5 copies of each microorganism. All samples were run in triplicate on a RapidHIT ID. Table 2 Reproducible peaks present in species DNA (n = 3 per species DNA). In bold, peaks that were previously reported [19]. Average Peak Height (RFU) Dye Marker Allele Size (bp) Chicken 173 TAZ SE Horse 111 VIC NA OL FAM D3S1358 OB Porcine 937 TAZ SE Table 3 Heterozygote peak height ratios (PHR), standard deviation (SD), and minimum values for 22 GlobalFiler 1 Express markers calculated from samples run in the concordance study. Marker Number of allele pairs Average PHR (%) SD PHR (%) Min (%) D3S % 9.1% 52.8% vwa % 10.4% 59.9% D16S % 9.8% 57.6% CSF1PO % 11.2% 43.3% TPOX % 7.9% 67.6% AMEL % 12.3% 44.3% D8S % 9.9% 62.3% D21S % 7.3% 69.8% D18S % 8.2% 67.8% D2S % 5.7% 78.0% D19S % 12.3% 24.0% TH % 9.3% 50.5% FGA % 6.2% 76.8% D22S % 11.7% 48.3% D5S % 7.5% 73.5% D13S % 7.4% 62.2% D7S % 7.5% 69.2% SE % 13.9% 45.7% D10S % 7.2% 71.9% D1S % 10.6% 46.6% D12S % 9.3% 64.0% D2S % 14.7% 31.0%

5 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Sensitivity Sensitivity was tested by running the following sets of cottontipped swabs: (1) two-fold serial dilutions of 1000 M cells from 200,000 cells down to 3125 cells (1.2 mg ng DNA based on 6 pg DNA/diploid cell). Cells were prepared and added to cottontipped swabs in triplicate per dilution; (2) mock collection samples: cotton-tipped swabs with varying amounts of DNA collected in triplicate from a female and a male subject using differing swab collection techniques: 1 touch, 1 swipe, 2 swipes, 5 swipes, 10 swipes and 20 swipes inside the cheek Concordance and carryover study A total of 54 donor buccal swabs (51 unique individuals and 3 repeats) were processed on two RapidHIT ID systems, and genotype concordance was compared with the reference database described in Section To test the RapidHIT ID systems for sample-to-sample carryover a total of 30 blank samples (i.e., a sample cartridge without sample) were processed. The blank samples were processed in a quasi-checkerboard sequence with the concordance samples: a blank sample was processed after every two to three donor swabs. Carryover or contamination is indicated by the presence of electropherogram peaks above the AT. To further evaluate the risk of carryover with very high-load DNA samples, an alternating sequence of high-load DNA samples and blanks were run on two RapidHIT ID systems. Each sample comprised 500, M cells pipetted onto a cotton-tipped swab. To further increase the DNA carryover load, the number of PCR cycles for the 1000 M swabs was increased from the standard operating condition of 28 to 30. Quantification of the DNA in lysates from 1000 M swabs with 500,000 cells and swabs from a donor (with 5 swipes from the inside the cheek) was performed by incubating the swabs with 400 ml of lysis buffer (DNA IQ System, Promega, Madison, WI) for 30 min at 75 C. Ethanol was added to the lysate to reach 80% by volume, centrifuged at 4 C for 1 h at 25,000 rpm. The supernatant was aspirated, pellet washed with 80% ethanol followed by 4 C centrifugation for 30 min at 25,000 rpm. The supernatant was aspirated, and the DNA pellet was re-suspended in 50 ml of TE buffer. The DNA concentration was measured using Quantifiler Human DNA Quantification Kit (Thermo Fisher Scientific) on the 7500 Real-Time PCR System (Thermo Fisher Scientific) Swab retrieval and re-extraction To demonstrate that samples run on the RapidHIT ID can subsequently be retrieved and re-run on a conventional system, twenty buccal swab samples from the concordance study were randomly selected. The swabs were removed from the sample cartridge and reprocessed on the bench following the process described in Section DNA was precipitated from the lysate and quantified as described in Section 2.8. STR profiles generated by the RapidHIT ID were compared with the profiles from the same sample manually processed, as well as with their profile in the reference database Repeatability and reproducibility Five 1000 M control swabs containing 100,000 cells each were used to assess repeatability (using one RapidHIT ID system) and reproducibility (using three RapidHIT ID systems). Concordance was checked against the NIST certified genotype. Heterozygote peak height balance, average peak height and intracolor balance were calculated Electrophoresis sizing accuracy, stutter calculations and precision The sample profiles from the concordance study (Section 2.8) were used to measure sizing accuracy and to calculate stutter. The AT was lowered to 1 RFU and the stutter filters were set to 1% in the GeneMarker panel file to detect all stutter peaks and their corresponding peak heights. The proportion of stutter product relative to the main allele (i.e., percent stutter) was measured by dividing the height of the stutter peak by the height of the associated allele peak. RapidHIT ID precision was calculated by running 10 Global- Filer 1 Express control cartridges in each of three systems for a total of 30 allelic ladder samples. Precision was determined by calculating the standard deviation of the size for each allele in the combined data set Data analysis The balance between allelic peaks within heterozygous loci (also known as peak height ratio (PHR), intralocus balance or heterozygous balance) was calculated on a per locus basis by Fig. 4. Effect of lysis incubation time on average peak height and heterozygote peak height ratio.

6 26 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 5. PCR boundary study results using control swabs (n = 3 per condition; average SD). Asterisk indicates standard condition. dividing the lower allele peak height of a heterozygous sample by the higher allele peak height with the result expressed as a percentage. Overall average peak height for a sample was determined by averaging heterozygous peaks, dividing the homozygous peaks in half and then calculating the overall average. Profiles were examined for balance within each dye (also known as intracolor balance or ICB) by averaging heterozygous peaks and dividing the homozygous peaks by two. Once normalized for diploidy, the lowest normalized peak height was divided by the highest within a given dye, and the result expressed as a percentage. 3. Results and discussion 3.1. DNA extraction Modification of the incubation time of the buccal cotton-tipped swab with the lysis buffer did not affect the performance on the RapidHIT ID system. Complete concordance was obtained at all incubation times. Decreasing the lysis time had no noticeable impact on the average peak height and heterozygote peak height ratios as shown in Fig. 4. Fig. 6. Effect of tobacco on control swabs (n = 3 per condition) with 100, M cells (average SD).

7 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 7. Sensitivity study with 1000 M cells applied to swabs (n = 3 per condition; average SD) Thermal cycling parameters Boundary studies of the optimized thermal cycling parameters were conducted to verify the performance conditions selected for the GlobalFiler 1 Express assay on the RapidHIT ID system. A two degree Celsius increase or decrease of temperature for enzyme activation, denaturation and annealing/extension steps did not exhibit any impact on the average peak height or the heterozygote peak height ratios as shown in Fig. 5. In addition, complete concordance was obtained at each temperature for every sample. Reducing the final extension time by 2 min did not affect the STR profile and there was no evidence of incomplete adenylation. The optimum number of PCR cycles (i.e., 28) showed a higher average peak height than the one cycle lower and similar to the one cycle higher. In all three test cases, the heterozygote peak height ratios were similar. Overall, the results presented here indicate that the selected PCR conditions are highly robust Mock inhibition study A mock inhibition study was performed by adding coffee or tobacco to cotton-tipped swabs containing 100, M cells. Complete and concordant profiles were obtained when coffee was present at all four different concentrations. Average peak heights and average heterozygote peak height ratios were similar to the control containing no coffee (data not shown). However, increasing the amount of tobacco decreases the average peak height with no observable effect on average heterozygote peak height ratio (Fig. 6). Complete and concordant profiles were obtained with up to 10 ml of tobacco slurry. At 50 ml and 100 ml of tobacco slurry, 98.4% and 85.7% of the total number of the expected alleles, respectively, were detected (Table 1). Table 4 Stutter range (for N 4), average and standard deviation (SD) (for N + 4, N 4 and N 8) on the RapidHIT ID system from samples run in the concordance study. Locus N 4 Stutter Range(%) N 4 Stutter Av(%) N 4 SD(%) N + 4 Av(%) N + 4 SD(%) N 8 Av (%) N 8 SD (%) D3S vwa D16S CSF1PO TPOX D8S D21S D18S DYS D2S D19S TH FGA D22S D5S D13S D7S SE D10S D1S D12S D2S

8 28 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Species specificity A few reproducible cross-reactivity peaks were observed in all three replicates for chicken, horse and porcine as summarized in Table 2. The peaks indicated in bold font for horse and porcine samples have previously been reported in a validation study of GlobalFiler 1 Express [19]. The presence of the additional reproducible peaks was confirmed by processing samples using Fig. 8. Electropherograms from the sensitivity study for the male donor. Note the vertical axis scale is 0 24,000 RFU for (d) (f) and 0 7,000 RFU from 2 swipes to 1 touch for (a) (c).

9 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 8. (Continued)

10 30 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 9. Accuracy. Size difference (in base pairs) between an allele and its corresponding allele in the allelic ladder used to size the sample (53 samples, 2138 alleles). The color of each data point indicates the corresponding dye in the GlobalFiler 1 Express assay; FAM (blue), VIC (green), NED (yellow), TAZ (red) and SID (purple). the bench method (Section 2.3.1) in a separate laboratory using a different operator and equipment Sensitivity Complete and concordant profiles were obtained for 1000 M samples from 12,500 to 200,000 cells/swab. With 6250 cells, 51.6% of the alleles were detected. As expected, the average height for the peaks above AT decreased with the lower input of cells (Fig. 7). On careful inspection, no artifacts (spectral or PCR-based) were observed above the analytical threshold on the high DNA-load samples (200k cells and 20 swipes). Stutter was well within the ranges described in Table 4. All replicates of the male and female donor samples produced complete and concordant profiles for all sample collection types, including one touch to the cheek, indicating that there is a sufficient amount of cells and DNA being recovered to yield full DNA profiles. Peak height decreased with lower input of donor cells as shown for the male samples in Fig Concordance and carryover contamination To study concordance, a set of donor buccal swabs (51 different individuals and 3 repeats) was processed on two RapidHIT ID systems. One profile exhibited low resolution size standard peaks and was excluded from analysis. Genotype concordance was checked for all 53 remaining profiles against reference profiles generated from the GlobalFiler 1 Express runs on the PCR System 9700/3130 l Genetic Analyzer instruments as described in All samples produced complete and concordant STR profiles for all GlobalFiler 1 Express loci. The average heterozygote peak height ratio was calculated using all 53 donor samples and ranged from 77.9% to 90.5% as shown in Table 3. Overall, only 8 of the 915 (or the equivalent to 0.9% of the total heterozygote pairs) heterozygote peak height ratios in the study were below 50%. To calculate the sizing accuracy, the difference in bp was calculated between the size of each sample allele and its corresponding allele in the allelic ladder used to size that sample. The calculation was performed on the concordance data set of 53 samples. All alleles tested were well within 0.5 bp from the corresponding allele in the allelic ladder (Fig. 9). These results show the RapidHIT ID demonstrates appropriate accuracy for sizing alleles that differ in length by a single bp. Stutter peak height percentages at one repeat unit shorter than the true allele (N 4) were calculated by dividing the peak height of the stutter peak by its associated true allele peak height. For the concordance sample set, the stutter peak height averages, ranges (minimum and maximum) and standard deviation (SD) were calculated and are shown in Table 4. The same process was used for N 8 and N + 4 stutter peaks. Zero values in Table 4 indicate insignificant stutter incidence rate (<1%).The N 4 values are comparable to those included in the GlobalFiler 1 Express (Thermo Fisher Scientific, Waltham, MA) User Guide Rev C [17]. The default stutter values used to evaluate the concordance data set were the maxima observed in Table 4. Single source reference DNA samples might be affected by sample-to-sample carryover when a sample containing a large amount of DNA is processed before a sample containing a small amount of DNA. To evaluate sample-to-sample carryover risk two methods were used. In the first method, blank samples were interspersed with the concordance samples in a quasi-checkerboard sequence. No carryover peaks were observed in any of the 30 blank runs. Two blanks exhibited primer flare (Fig. 10(a) and (b)) and one blank exhibited a dye blob (Fig. 10(c)) characterized by a broad peak above the analytical threshold simultaneously occurring in >3 colors.

11 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 10. Electropherograms for blank runs showing (a) and (b) primer flare peaks, (c) dye blobs, (d) anomaly peak in D2S441.For all graphs, the vertical axis is intensity in RFU and the horizontal axis is fragment length from 76 to 200 bp. The second method used a checkerboard sequence of blank and very high-load DNA samples. High-load samples were prepared using cotton-tipped swabs loaded with 500, M cells. The amount of DNA recovered from the lysate of the swabs was in average 1.5 mg, approximately 7 times more DNA than the amount from a medium to high shedder donor swab. To further increase the DNA load through the system, the high-load samples were processed with two additional PCR cycles. Ten high-load and ten blank samples were processed in an alternating sequence on two RapidHIT ID instruments. Blank samples were processed with the optimized number of PCR cycles. The electropherograms of the blank samples were examined for peaks above the AT. In agreement with the previous study, no carryover peaks were observed in the blank runs in the second method described above. A very small peak (54 RFU) was observed at the short fragment end of D2S441 on one of the twenty blank runs. This peak was located at 90.4 bp, between alleles 10 and 11. No variant allele has been reported between alleles 10 and 11 of D2S441 for GlobalFiler1 Express kit [20]. This information, in conjunction with the fact that this is the sole peak present in the electropherogram, suggests that this anomaly is most likely to be a primer flare peak (Fig. 10(c)). These two studies combined indicate that the RapidHIT ID is highly robust against sample-to-sample carryover Swab retrieval and re-extraction For certain use cases, it is useful to reprocess a sample that has already been processed by the RapidHIT ID system [7]. Twenty samples were reprocessed following the method described in Section 2.9. STR profiles were complete and concordant for all samples processed using the RapidHIT ID and the subsequent bench method. For the bench method, DNA yields from the lysate ranged from 26 ng to 291 ng Repeatability and reproducibility Five 1000 M control swabs, loaded with 100,000 cells, were processed on three RapidHIT ID systems for a total of 15 control swabs. All samples produced complete and concordant STR profiles. The three systems produced comparable average peak heights and average heterozygote peak height ratios (Fig. 11). Size difference between sample alleles and the corresponding alleles in

12 32 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 11. Average peak height and average heterozygote peak height ratio for each of three RapidHIT ID systems and the combined result (n = 5 per system; n = 15 combined; average SD). Fig. 12. Size difference between allele and corresponding allele in allelic ladder for fifteen 1000 M samples (average SD). the allelic ladder was calculated for all three systems following the method described in Section 3.6. Allele sizing differences were all well within 0.5 bp (Fig. 12), demonstrating RapidHIT ID systems can robustly identify alleles separated by 1 bp. Finally, the intracolor balance was calculated for the fifteen 1000 M swabs. The results demonstrate intracolor balance equivalent to other sample-in STR profile-out systems (Fig. 13) [4] Precision Slight variability in allele migration leads to slight differences in the size of any given allele across multiple samples. GlobalFiler 1 Express is designed to identify many alleles which are separated by 1 bp. To prevent an allele from being identified as its neighbor the allele size must be within 0.5 bp of its corresponding ladder allele size. To evaluate a system s performance against this requirement, it is necessary to study the size of a large number of alleles across the entire GlobalFiler 1 Express size range. Size precision is the standard deviation of the size of a given allele. To significantly reduce the risk of an allele being identified as a neighbor separated by 1 bp, three times the size precision for any allele must be less than 0.5 bp stated differently, the size precision should be less than a standard deviation of bp. In this study, precision was determined by processing ten control cartridges (allelic ladder) in each of three RapidHIT ID systems. The size of each allele was calculated and the standard deviation of each allele size as a function of allele size is shown in Fig. 14. All alleles from each control cartridge sample were included. All data points in Fig. 14 have size standard deviation below bp. The slow upward trend of the sizing precision

13 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) Fig. 13. Intracolor balance (ICB) for fifteen 1000 M samples using 3 instruments (average SD). Fig. 14. Standard deviation of allele size. Data represents 10,290 total alleles from 30 allelic ladder samples on three RapidHIT ID systems. reflects a slight increase in migration variation for long fragment length alleles. Overall, the RapidHIT ID shows highly reliable size precision. 4. Discussion The RapidHIT ID system and RapidLINK software are designed to meet the requirements for decentralized Rapid DNA analysis in a wide-scale hub-and-spoke deployment model. The RapidHIT ID requires less than a minute of hands-on time from the operator and generates a full GlobalFiler 1 Express profile in 90 min. The results presented here show highly robust performance of the RapidHIT ID system over a wide range of operating conditions as well as 100% genotype concordance with standard bench-processing methods with the samples tested herein. PCR and extraction conditions were optimized on the RapidHIT ID system to generate full profiles from buccal swabs. Boundary studies for the different PCR steps showed the robustness of the optimized conditions. Complete profiles were obtained with as little as 12,500 cells as well as one-cheek-touch swabs. A large number of donor swabs were processed within a week of collection and demonstrated exceptional sizing accuracy and full concordance. In addition, 20 of those swabs were removed from the RapidHIT ID system and reprocessed with a bench method, which resulted in complete and concordant reference profiles. Negligible interference was demonstrated from other species DNA or common buccal microorganisms. Standard inhibitors were shown to have negligible effect.

14 34 S. Salceda et al. / Forensic Science International: Genetics 28 (2017) No sample-to-sample carryover was detected, including very high-load DNA samples amplified using two incremental PCR cycles. Highly acceptable size precision was demonstrated. The success rate for the concordance data set was 98%. Overall, for all data sets, excluding inhibition, the success rate was 94%. The developmental validation studies described here used the National DNA Index System (NDIS)-approved GlobalFiler 1 Express kit without alteration of the product. Studies of GlobalFiler 1 Express have previously been performed [17,19], including developmental validation studies as required by SWGDAM guidelines [16] for characterization of loci, population studies, and PCR components. The studies from the manufacturer, in conjunction with the studies presented here, can be used for decentralized facilities and laboratories to support internal validation studies. Conflict of interest statement None. Role of funding All work was supported by IntegenX. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgements The authors acknowledge the indispensable work in early phases of research and development of Kaiwan Chear, David Eberhart, Helen Franklin, Stevan Jovanovich, William Nielsen, Jeff Perry, Robert Porporato, Teresa Snyder-Leiby (SoftGenetics LLC), Nancy Stainton and Steve Williams. References [1] A.J. Jeffreys, V. Wilson, S.L. Thein, Hypervariable minisatellite regions in human DNA, Nature 314 (1985) [2] A. Edwards, A. Civitello, H.A. Hammond, C.T. Caskey, DNA typing and genetic mapping with trimeric and tetrameric tandem repeats, Am. J. Hum. Genet. 49 (1991) [3] C.P. Kimpton, P. Gill, A. Walton, A. Urquhart, E.S. Millican, M. Adams, Automated DNA profiling employing multiplex amplification of short tandem repeat loci, PCR Methods Appl. 3 (1993) [4] L.K. Hennessy, N. Mehendale, K. Chear, S. Jovanovich, S. Williams, C. Park, S. Gangano, Developmental validation of the GlobalFiler 1 Express kit, a 24- marker STR assay, on the RapidHIT 1 System, Forensic Sci. Int. Genet. 13 (2014) [5] S. Jovanovich, G. Bogdan, R. Belcinski, J. Buscaino, D. Burgi, E.L. Butts, K. Chear, B. Ciopyk, D. Eberhart, O. El-Sissi, H. Franklin, S. Gangano, J. Gass, D. Harris, L. Hennessy, A. Kindwall, D. King, J. Klevenberg, Y. Li, N. Mehendale, R. McIntosh, B. Nielsen, C. Park, F. Pearson, R. Schueren, N. Stainton, C. Troup, P.M. Vallone, M. Vangbo, T. Woudenberg, D. Wyrick, S. Williams, Developmental validation of a fully integrated sample-to-profile rapid human identification system for processing single-source reference buccal samples, Forensic Sci. Int. Genet. 16 (2015) [6] L.K. Hennessy, H. Franklin, Y. Li, J. Buscaino, K. Chear, J. Gass, N. Mehendale, S. Williams, S. Jovanovich, D. Harris, K. Elliott, W. Nielsen, Developmental validation studies on the RapidHIT TM human DNA identification system, Forensic Sci. Int. Genet. Suppl. Ser. 4 (2013) e7 e8. [7] B.L. LaRue, A. Moore, J.L. King, P.L. Marshall, B. Budowle, An evaluation of the RapidHIT 1 system for reliably genotyping reference samples, Forensic Sci. Int. Genet. 13 (2014) [8] M. Holland, F. Wendt, Evaluation of the RapidHIT TM 200, an automated human identification system for STR analysis of single source samples, Forensic Sci. Int. Genet. 14 (2015) [9] E.L. Romsos, P.M. Vallone, Rapid PCR of STR markers: applications to human identification, Forensic Sci. Int. Genet 18 (2015) [10] Z. Thong, Y.H. Phua, E.S. Loo, S.K. Goh, J. Ang, W.F. Looi, C.K. Syn, Evaluation of the RapidHIT TM 200 system: a comparative study of its performance with Maxwell 1 DNA IQ TM /Identifiler 1 Plus/ABI 3500L workflow, Forensic Sci. Int. Genet. 19 (2015) [11] S. Verheij, L. Clarisse, M. van den Berge, T. Sijen, RapidHIT TM 200, a promising system for rapid DNA analysis, Forensic Sci. Int. Genet. Suppl. Ser. 4 (2013) e254 e255. [12] H.S. Mogensen, R. Frank-Hansen, B.T. Simonsen, N. Morling, Performance of the RapidHIT TM 200, Forensic Sci. Int. Genet. Suppl. Ser. 4 (2013) e286 e287. [13] S. Gangano, K. Elliott, K. Anoruo, J. Gass, J. Buscaino, S. Jovanovich, D. Harris, DNA investigative lead development from blood and saliva samples in less than two hours using the RapidHIT TM human DNA identification system, Forensic Sci. Int. Genet. Suppl. Ser. 4 (2013) e43 e44. [14] M. Date-Chong, W.R. Hudlow, M.R. Buoncristiani, Evaluation of the RapidHIT TM 200 and RapidHIT GlobalFiler 1 Express kit for fully automated STR genotyping, Forensic Sci. Int. Genet. 23 (2016) 1 8. [15] Federal Bureau of Investigation, Quality Assurance Standards (QAS) for Forensic DNA Testing Laboratories (effective September 1, 2011), Quality Assurance Standards for Convicted Offender DNA Databasing Laboratories (effective September 1, 2011) and Addendum to the Quality Assurance Standards for DNA Databasing Laboratories performing Rapid DNA Analysis and Modified Rapid DNA Analysis Using a Rapid DNA Instrument (effective December 1, 2014). Available at. biometric-analysis/codis. [16] Scientific Working Group on DNA Analysis Methods (SWGDAM). Validation Guidelines for DNA Analysis Methods. (Approved December 2012) Available at [17] GlobalFiler 1 Express PCR Amplification Kit User Guide, Rev C. Pub.No Thermo Fisher Scientific. [18] C.A. Rakay, J. Bregu, C.M. Grgicak, Maximizing allele detection: effects of analytical threshold and DNA levels on rates of allele and locus drop-out, Forensic Sci. Int. Genet. 6 (2012) [19] D.Y. Wang, S. Gopinath, R.E. Lagacé, W. Norona, L.K. Hennessy, M.L. Short, J.J. Mulero, Developmental validation of the GlobalFiler 1 Express PCR amplification kit: a 6-dye multiplex assay for the direct amplification of reference samples, Forensic Sci. Int. Genet. 19 (2015) [20] Short Tandem Repeat DNA Internet DataBase (STRBase). NIST Standard Reference Database SRD (Accessed 14 September 2016).