DNA recovery after sequential processing of latent fingerprints on black polyethylene plastic

Latent fingerprints on plastic substrates can be visualized by using sequential treatments to enhance the contrast between the fingerprint residues and underlying substrate; however, the extent to which these processes affect subsequent DNA analysis is mostly unknown. Latent fingerprints deposited on black plastic by one donor were visualized with single‐process fingerprint powders (i.e., white powder, bichromatic powder, or bichromatic magnetic powder) or sequential treatments (i.e., laser → reflected ultraviolet imaging system (RUVIS) → CA fuming → RUVIS → Rhodamine 6G, Ardrox, and MBD (RAM) or CA fuming → RAM/laser → bichromatic magnetic powder). Samples were examined after the addition of each treatment. DNA was collected using cotton swabs, extracted, quantified, and amplified. DNA yields, peak heights, number of alleles obtained, and percentage of DNA profiles eligible for CODIS upload were examined. Latent fingerprints processed with the laser and up to three sequential treatments generated DNA profiles with significantly higher peaks heights than those of the untreated samples. Fingerprints processed with the laser and up to two sequential treatments generated DNA profiles with significantly more alleles. All methods beginning with laser enhancement generated more CODIS‐eligible profiles. Additional research is needed to determine the extent to which initial laser enhancement impacts the success of downstream DNA profiling results. Although DNA profile development is not guaranteed due to the variable quantities of DNA contained within latent fingerprints, the selection of an appropriate latent fingerprint visualization method could maximize both fingerprint detection and the generation of CODIS‐eligible DNA profiles.


| INTRODUC TI ON
Fingerprint visualization and forensic DNA testing may be performed when examining evidence during crime scenes investigations and in forensic laboratories.When the use of both methods is not possible for the same evidentiary item, examiners have to determine which is likely to offer the best chance for identification.Latent fingerprint residues, which contain sebaceous oils, eccrine sweat, and contaminants from the environment, can be visualized on nonporous substrates with single-process methods such as alternative light source (ALS) enhancement [1,2], cyanoacrylate (CA) fuming [3], and fingerprint powders [4][5][6], and quantifiable amounts of DNA can be recovered from latent fingerprints treated with these methods, although the profile quality may be negatively affected [7][8][9][10].Further processing can be performed to enhance contrast and improve the visualization of the latent fingerprints [5]; however, little is known about the extent to which multi-step sequential processes affect DNA recovery.
Identifying these effects can aid fingerprint examiners in determining when to stop processing evidence items that will undergo downstream DNA processing.For example, CA fuming alone has not been found to have an adverse effect on DNA typing of latent fingerprints on plastic [7], whereas Kumar et al. observed that the use of CA fuming followed by Rhodamine 6G dye to visualize latent fingerprints on plastic foil resulted in a negative effect on DNA typing [8].
Sequential processing of latent fingerprints on nonporous substrates typically involves visual examination, sequentially followed by examination with a laser or ALS, CA fuming, staining with a fluorescent dye, laser or ALS examination again, and fingerprint powder application [4,5,11,12].During the fuming process, the CA vaporizes as it is heated and polymerizes as a white residue on latent fingerprint residue [13,14].As a result, latent fingerprints become fixed on the substrate and are readily detectable if on a color-contrasting substrate [14].Following CA fuming, latent fingerprints can be treated with powder or fluorescent dye stain that adheres to the white CA residue [15] and then examined with a laser/ALS and/or a reflected ultraviolet imaging system (RUVIS) [6,10,11,13].A dye stain mixture of Rhodamine 6G, Ardrox, and 7-(p-methoxybenzylam ino)-4-nitrobenz-2-oxa-1,3-diazole (MBD) (RAM) has been used to enhance latent fingerprints detected with CA fuming [4,16,17] and is effective on many colors of nonporous surfaces [11].RAM dye stain is visualized with a laser or ALS (415-532 nm excitation range) to further enhance the latent fingerprints [4,11,17].If a latent fingerprint has yet to be located, the surface can be dusted with a fingerprint powder [11], which is selected based on the type and color of the surface being processed [4], or processed with vacuum metal deposition [18].
The aim of this pilot study was to evaluate the effects of single and sequential latent fingerprint visualization processes on downstream DNA analysis.Three fingerprint powders used for visualization of latent fingerprints on dark, nonporous surfaces were chosen for single-process methods in this study: white powder [5], bichromatic powder [4], and bichromatic magnetic powder [4].Two sequential processes for visualizing latent fingerprints on plastic were evaluated: (1) enhancement with a 532 nm laser → enhancement with RUVIS → CA fuming → enhancement with RUVIS → RAM dye stain [4,11,12,17] and (2) CA fuming → RAM dye stain (enhanced with a 532 nm laser) → bichromatic magnetic powder [4,11,12].

| Fingerprint preparation
A total of 91 fingerprints were prepared on black 22 μm polyethylene plastic bags (Brighton 56 gal professional super heavy strength black trash bags) by one donor over the course of 8 days.Polyethylene bags were selected due to the prevalence of polyethylene in common household items and the frequency with which plastic bags are received as evidence exhibits [19].The plastic bags were cut into 1.5″ × 2.0″ segments and decontaminated by short-wave UV irradiation in a Spectronics Spectrolinker™ XL-1500 with a 254 nm source.
Substrates were placed on a desktop scale, and the index, middle, and ring fingers from both hands were used to deposit fingerprints by pressing down until 2 kg was reached after a total substrate contact time of approximately 1 s, at which point the finger was removed from the substrate.The fingerprint deposition approach was adapted from Hefetz et al. [20] and Jasuja et al. [21].Fingerprint preparation instructions were provided to the donor in order to limit intra-donor variation; however, to mimic real-world handled evidence, fingertips were not charged with additional sebaceous oils, handwashing was kept to a minimum, and at least 2 h elapsed after handwashing and between fingerprint depositions.

| Fingerprint processing
Samples were treated with single-process fingerprint powders or sequential latent fingerprint visualization methods.Five replicates • Latent prints on polyethylene were visualized with powder, laser, RUVIS, CA fuming, and RAM stain.
• More CODIS-eligible profiles were obtained after visualization starting with laser enhancement.
• CODIS eligibility was negatively impacted with the addition of each treatment after laser + RUVIS.
were processed with either white powder, bichromatic powder, or bichromatic magnetic powder.Two sequential processes were examined: (1) laser → RUVIS → CA fuming → RUVIS → RAM; and (2) CA fuming → RAM/laser → bichromatic magnetic powder.Within each process, eight replicates were examined after the addition of each treatment to determine its effect on DNA recovery.Additionally, 12 untreated fingerprints were included as controls.

| Laser
During laser enhancement, latent fingerprints were exposed to 532 nm irradiation for approximately 10 s using a 4 W TracER™ laser (Coherent, Inc.).

| Cyanoacrylate fuming
CA fuming was conducted using 1.0 g Cyanobloom (Foster + Freeman Ltd.) in a model MVC3000 fingerprint fuming cabinet (Foster + Freeman Ltd.).The cabinet was utilized with an autocycle feature, which involved three stages: initial purge and humidification for 15 min, fuming for 18 min, and final purge for 20 min.

| Fingerprint powders
White powder (The Safariland Group) and bichromatic powder (The Safariland Group) were applied using disposable synthetic fiber fingerprint brushes (Arrowhead Forensics).Bichromatic magnetic powder (The Safariland Group) was applied using a magnetic fingerprint wand (Lynn Peavey Company).To avoid cross-contamination, all powders were dispensed into singleuse portions for each sample.A new disposable brush was used for each sample, and the magnetic wand was sterilized between samples.

| Reflected ultraviolet imaging system
RUVIS exposure was simulated using a 12 V DC Spectroline model EF-140/12H shortwave ultraviolet lamp (Clarion Safety Systems).
Latent fingerprints were exposed to UV irradiation for approximately 30 s.

| RAM dye stain
RAM dye stain was prepared following the Federal Bureau of Investigation's guidelines for developing latent fingerprints (Table 1) [11].Samples were briefly rinsed with the reagent and air dried for approximately 30-60 min.

| Sample collection
DNA was collected directly from the plastic bag segments with 100% cotton Bode SecurSwabs (Bode Technology) by wetting one side of the swab head with 1-2 drops of sterile DNA grade water from a 3 mL AddiPak brand water vial, swabbing the area where the fingerprint was deposited and/or visualized, turning the swab to the dry side of the swab head, and swabbing the area again.To determine where in the sequential process the DNA may be affected, DNA was collected after each additional visualization method.

| DNA analysis
Samples were extracted using the EZ1® DNA Investigator Kit (Qiagen).Lysis was performed by incubating the whole swab head in 95 μL Buffer G2, 95 μL ddH 2 O, and 20 μL Proteinase K at 56°C with shaking at 900 rpm for 1 h.To collect the lysates, samples were transferred to spin baskets and centrifuged at 20,000 g for 2 min.The spin basket containing the swab head was discarded, and 1 μL carrier RNA (1 μg/μL) was added to the collected lysate.The samples were purified on an EZ1 Advanced Instrument (Qiagen) using the trace proto- Reagent blanks, substrate negative controls, extraction positives, positive amplification controls, and negative amplification controls were also processed.

| Data analysis
The effects of the single process and sequential fingerprint visualization methods were evaluated by examining DNA yield, STR profile peak height values, number of alleles obtained, and percentage of profiles eligible for Combined DNA Index System (CODIS) upload (percentage of profiles containing alleles from at least eight loci with a match rarity of at least one in 10 million [22]).For each sample, the obtained profile was compared to the known genotype of the donor.The maximum number of alleles that could be obtained was 46.Expected alleles that did not meet the AT were assigned peak heights of 0 RFU.Homozygous genotypes were expected at D13S317 and D19S433 and were counted as two alleles when the peak heights were above the ST, in which case peak heights were divided in half and applied to each allele.
JMP® Software (SAS Institute Inc.) was used to perform statistical analysis of the data.The Shapiro-Wilk test for normality was performed to determine whether the data were normally distributed.
Shapiro-Wilk showed that data distribution departed significantly from normality (number of alleles, p < 0.001; DNA yield, p < 0.001; peak height, p = 0.010).Based on the data distribution and the small sample sizes, rank-based nonparametric tests were used for further comparisons.For DNA yield, number of alleles obtained, and peak height, comparisons to the untreated controls samples were made using Steel With Control, a nonparametric analog to Dunnett's test.
The Wilcoxon test (i.e., Mann-Whitney test) was used to compare peak heights for lower (<200 bp) and higher (>200 bp) molecular weight loci.For all statistical analyses, differences were considered statistically significant at p < 0.05.

| Latent fingerprints visualized with single-process fingerprint powders
DNA yields, peak heights, number of alleles obtained, and percentage of profiles eligible for CODIS upload obtained from the untreated latent fingerprints and the latent fingerprints treated with either white powder, bichromatic powder, or bichromatic magnetic powder were compared (Figure 1).Medians and p-values resulting from the Steel With Control comparisons can be found in Table 2.The number of alleles obtained for samples subjected to each visualization method can be found in Table 3.The DNA yields of the untreated samples ranged from 0.009-1.240ng and were significantly higher than the yields for samples visualized with white powder (p = 0.011), bichromatic powder (p = 0.028), and bichromatic magnetic powder (p = 0.028).The untreated samples also resulted in significantly higher peak heights when compared to those treated with white powder (p < 0.001), bichromatic powder (p = 0.001), and bichromatic magnetic powder (p < 0.001).Peak heights for the lower molecular weight loci (<200 bp) were significantly greater than the higher molecular weight loci (>200 bp) for all samples: untreated (p = 0.001), white powder (p = 0.024), bichromatic powder (p < 0.001), bichromatic magnetic powder (p = 0.038).These significant differences can be attributed to the alleles that dropped out, which were assigned peak heights of 0 RFU.The white powder and bichromatic magnetic powder samples did not produce any peak heights above AT for the higher molecular weight loci, although allelic dropout was observed at all loci across all samples.No significant differences were observed between the number of alleles obtained for the untreated and treated samples.Three untreated samples, two white powder samples, and two bichromatic magnetic powder samples produced no alleles.Only one untreated sample generated a profile eligible for CODIS upload.

| Latent prints treated with sequence 1: Laser → RUVIS → CA fuming → RUVIS → RAM
Latent fingerprints treated sequentially with the laser, RUVIS, CA fuming, RUVIS again, and RAM were evaluated using the aforementioned metrics (Figure 1).Medians and p-values resulting from the Steel With Control comparisons can be found in Table 2.The number of alleles obtained for samples subjected to each visualization method can be found in Table 3.No significant differences were observed between the DNA yields obtained from the untreated and sequentially treated samples.Significantly higher peak heights were obtained from laser, laser + RUVIS, laser + RUVIS + CA fuming, and laser + RUVIS + CA fuming + RUVIS samples when compared to the untreated samples (p < 0.001).The number of alleles obtained following laser, laser + RUVIS, and laser + RUVIS + CA treatments were significantly higher than those obtained by the untreated samples (p = 0.037, p = 0.013, p = 0.019, respectively).One laser + RUVIS + CA fuming + RUVIS + RAM sample produced no alleles.
Peak heights for the lower molecular weight loci were significantly greater than the higher molecular weight loci for all samples and visualization methods (p < 0.001).Stochastic effects and allelic dropout

| Latent prints treated with sequence 2: CA Fuming → RAM/laser → bichromatic magnetic powder
The effects of sequential treatments involving CA fuming, RAM/ laser, and bichromatic magnetic powder were also evaluated (Figure 1).Medians and p-values resulting from the Steel With Control comparisons can be found in Table 2.The number of alleles obtained for samples subjected to each visualization method can be found in

| DISCUSS ION
Latent fingerprints treated with single-process and sequential latent fingerprint visualization methods produced DNA profiles with varying degrees of success.The DNA results obtained from most of the treated fingerprints, specifically those treated with white powder, TA B L E 2 Medians and p-values obtained from comparisons to the untreated samples.The small sample sizes examined in this work limit the definitive conclusions that can be made from the data.Small data sets can be prone to bias, with high variability affecting accuracy.Some erable DNA [30], the success of alternative swabs types (e.g., nylon flocked) is determined by the type of sample being collected and the substrate on which it is deposited [31].An evaluation of alternative methods for collecting DNA from visualized fingerprints, such as cutting samples directly from the substrate [32], may improve DNA recovery and result in better quality STR profiles.Furthermore, DNA extraction can result in the loss of ≥72% of the initial template amount [33,34].Alternative DNA collection and processing methods may improve the overall quality of the STR profiles.In particular, direct PCR, a method in which a cutting or swab is added directly to an amplification reaction without prior extraction or quantification, has been identified as an effective method for improving DNA profiles from low-yield samples [35].Although research has shown that various single-process fingerprint powders (i.e., white hadonite, silver aluminium, HiFi volcano silk black, and black magnetic) do not interfere with direct PCR [36], the effects of sequential fingerprint visualization methods on direct PCR remain mostly unknown.
Despite using one donor to remove inter-donor variations related to individuals' differing propensities to shed epithelial cells, intra-donor variability was observed when examining the untreated control fingerprints.DNA was recovered from all untreated fingerprints (0.009-1.240 ng); however, only one untreated sample yielded a CODIS-eligible STR profile, and the number of alleles per profile varied widely (0-46 alleles).Previous studies have shown that DNA recovery from fingerprint samples deposited by one donor can be inconsistent due to intra-donor variation caused by a number of factors [20,[37][38][39][40][41].During sample preparation, factors that contribute to intra-donor variation were mitigated as much as possible for samples that were intended to mimic real-world handled evidence; however, the amount of DNA contained within a fingerprint cannot be fully controlled or standardized.

| CON CLUS IONS
Practitioners may have to determine whether both latent fingerprint visualization and forensic DNA testing will be performed on an col and eluting into 50 μL TE buffer.Next, 2 μL of each DNA extract were quantified in 11 μL reaction volumes using the Quantifiler® Trio DNA Quantification Kit (Thermo Fisher) on an ABI™ 7500 Real-Time PCR System with HID Real-Time PCR Software (Thermo Fisher).All samples were concentrated to 10 μL using Microcon® DNA Fast Flow centrifugal filter units (EMD Millipore, Billerica, MA) treated with 1 μL of carrier RNA (1 μg/μL) in TE buffer.All samples were amplified for 28 cycles on a GeneAmp™ PCR System 9700 (Thermo Fisher) using a GlobalFiler® PCR Amplification Kit (Thermo Fisher) in 12.5 μL reaction volumes with the maximum DNA input volume (7.5 μL).Extraction and amplification positive controls were TA B L E 1 Preparation of the RAM dye stain working solution.Rhodamine 6G (Sirchie) 1000 mL of methanol (Peroxide-Free/Sequencing, Fisher Scientific) MBD stock solution 1 g of 4-(4-Methoxybenzylamino)-7nitrobenzofurazan (MBD) (Sigma) 1000 mL of acetone (ACS Grade, Fisher Scientific) RAM working solution 3 mL of Rhodamine 6G stock solution 2 mL of Ardrox P133D (Armor Forensics) 7 mL of MBD stock solution 20 mL of methanol (Peroxide-Free/Sequencing) 10 mL of isopropanol (Optima Grade, Fisher Scientific) 8 mL of acetonitrile (HPLC grade, Fisher Scientific) 950 mL of petroleum ether (ACS grade, Fisher Scientific) amplified with a target template DNA quantity of 1 ng.Capillary electrophoresis was performed on a 3500xL Genetic Analyzer (Thermo Fisher) using 1 μL amplification product, POP-4™ polymer (Thermo Fisher), and 3500/3500xL Data Collection Software, v3.1.The injection time was 24 s with an injection voltage of 1.2 kV.Results were analyzed using ABI GeneMapper® ID-X v1.4 software (Thermo Fisher) with an analytical threshold (AT) of 125 relative fluorescent units (RFU) and a stochastic threshold (ST) of 600 RFU.
are commonly observed at higher molecular weight loci in low-yield DNA samples.Eighty-eight percent of profiles resulting from laser and laser + RUVIS samples met the threshold for CODIS eligibility, and the percentage of profiles eligible for CODIS decreased with the addition of each treatment.F I G U R E 1 (A) Distribution of the DNA yields obtained from each sample following each visualization method.(B) Distribution of the peak heights obtained for each allele at the lower (<200 bp, blue dot) and higher (>200 bp, green dot) molecular weight loci, with an AT of 125 RFU.(C) Distribution of the number of alleles obtained from each sample following each visualization method.(D) The percentage of profiles eligible for CODIS upload (≥8 loci with a match rarity of at least 1 in 10 million) for each process.
of the issues associated with small sample sizes were mitigated by selecting rank-based nonparametric tests for data comparisons due to their ability to handle non-normal data, small sample sizes, outliers, and extreme data values.However, expanding the scope of future work to include larger sample sizes and more donors will strengthen future findings.Notably, the untreated fingerprints did not produce any CODISeligible DNA profiles.In their previous work, Bathrick et al. found that 42% of untreated fingerprints deposited on copy paper generated CODIS-eligible profiles[27].Untreated fingerprints for the plastic and paper substrate studies were prepared by the same donor within the same time frame, and DNA collection and processing were performed with identical methods.Polyethylene plastic's propensity to bind DNA may have contributed to the loss of extracellular DNA[28,29] that would have been available if deposited on a different substrate, resulting in poor DNA typing results.The lowquality profiles obtained from the untreated fingerprints may also be partially due to inefficient DNA collection with the wet cotton swabbing method.While cotton swabs can retain up to 50% of the recov-

Table 3
. DNA yields for CA fuming and CA fuming + RAM/laser + bichromatic magnetic powder were significantly lower than the untreated controls (p = 0.014, p = 0.003).Peak heights for CA fuming, CA fuming + RAM/laser, and CA fuming + RAM/laser + bichromatic magnetic powder were also significantly lower than

Processing type Visualization method Sample size (n) DNA yield (ng) Peak height (RFU) # of alleles Median p Median p Median p
[23]er of alleles produced by the samples after each visualization method.RAM/laser, CA fuming + RAM/laser + bichromatic magnetic powder, and laser + RUVIS + CA fuming + RUVIS + RAM, were significantly lower than or did not differ significantly from the untreated samples.DNA being swept out of the target area by the brushing action or accumulating on the brush may have contributed to the lower DNA yields and peak heights obtained for samples treated with fingerprint powders[23].