The comparison of excision, swabbing, tape-lifting, and scraping as techniques for DNA evidence recovery: A preliminary study

Abstract

In the Final Report of the Commission of Inquiry into Forensic DNA Testing in Queensland, Recommendation 28 recommended that alternative procedures to the scraping method should be investigated for the recovery of biological material from crime scene exhibits for downstream DNA testing. This study forms the first phase of a larger project comparing swabbing, tape-lifting, scraping and excision of biological fluids from a fabric substrate to determine the most effective recovery procedure. Calico cotton swatches were seeded with two concentrations (neat and 1:1000) of blood, semen and saliva, samples commonly encountered in forensic casework. Each sample type was collected using the four evidence recovery methods, extracted, quantified, and the 1/1000 dilution samples were short tandem repeat (STR) profiled to assess the practical limits of each method. Excision had the greatest DNA yield with a large effect size and best STR profile quality compared to all other recovery methods for blood (p < 0.0001, ŋ² > 0.95), semen (p < 0.0001, ŋ² = 0.96) and saliva (p < 0.0001, ŋ² = 0.82). There were no statistical differences found between the MWE Sterile Tubed Dryswab™ Rayon swab 1028, the Puritan® 6” Standard Rayon Swab w/Polystyrene Handle – 806-PR, HazPak tape-lift or scraping for neat blood and semen. Tape-lifting had a statistically higher efficiency than scraping (p = 0.0012), the Puritan (p = 0.0046) and the MWE swabs (p = 0.0145). Results suggest that excision is the most effective biological material recovery technique for blood, semen, and saliva from 100 % white cotton.

1. Introduction

The successful collection of biological material from exhibits encountered in crime scenes or submitted to forensic laboratories is critical to generating interpretable DNA profiles to aid investigations. This process is often complicated by the nature of the surface or substrate upon which the material resides. Physiochemical properties such as porosity, texture, absorbency, and the interaction between substrate and material significantly influence the effectiveness of evidence recovery [1,2].

A range of DNA sampling techniques are employed in forensic laboratories to maximise recovery, with selection often guided by substrate type. For instance, smooth, non-porous surfaces such as glass and metal typically allow for more efficient sample collection using swabs. In contrast, porous or fibrous substrates—such as clothing, upholstery, or paper—tend to retain biological material in ways that reduce recovery efficiency [3]. The type of swab material, wetting agent, sampling motion, duration, and applied pressure all influence the quantity of DNA recovered [3].

Tape-lifting is particularly useful for recovering epithelial cells from porous surfaces and has been shown to outperform swabbing when used on fabrics such as denim, cotton, and polyester [4,5]. Vacuum-based collection methods have also emerged as robust alternatives, with dry vacuum devices for paper [6] and wet vacuum devices demonstrating improved recovery of trace DNA from porous materials like brick [7], carpet and fabrics [8,9].

Excision, which involves physically cutting out a portion of the substrate containing the suspected stain for direct extraction, has been shown to yield higher DNA quantities from blood-stained fabrics [10]. However, unlike the above techniques, excision is inherently destructive and less suitable in cases where preserving the integrity of the exhibit is essential [10,11].

The scraping method is conventionally employed to collect biological traces, particularly from textured surfaces, however, its utility is increasingly scrutinised due to contamination risks (cross-transfer between items or to the examiner), operator variability, and occupational safety concerns [12]. Apart from being labour intensive, the scraping procedure is also difficult to standardise across casework. Recognising these limitations, the Final Report of the Commission of Inquiry (CoI) into Forensic DNA Testing in Queensland [12] recommended a re-evaluation and potential phasing out of the scraping method. Recommendation 28 (Rec 28) specifically directed Forensic Science Queensland (FSQ) to investigate alternative DNA recovery techniques that may be more effective and safer than scraping [12].

This study addresses CoI Rec 28 b y systematically evaluating the efficiency of three commonly used DNA evidence recovery methods—swabbing, tape-lifting, and excision—compared to scraping, for recovering biological material from fabric substrates. Common biological fluids were spiked onto cotton fabric swatches at varying concentrations and collected using all four techniques. The study aimed to determine the most effective method for FSQ operational consideration based on key metrics, including DNA yield, degradation or inhibition, and the quality, strength, and allele count of the resulting DNA profiles.

2. Materials and methods

2.1. Fabric substrate and sample deposition

One fabric substrate (100 % bleached white calico cotton, (article number 80070415004) was purchased from a commercial vendor (Spotlight, Australia). Fabric swatches (17.5 cm × 5 cm), sectioned into ten (Fig. 1) were prepared and treated with dual cycle ethylene oxide (EtO) by Steritech Pty Ltd (NSW, Australia) according to International Organization for Standardization (ISO) standard ISO 18385 [13] to render them DNA free. Donor biological fluids were stored at 4°C until use. One operator performed all biological fluid depositions and each recovery method for consistency. Deposition was conducted within the Evidence Recovery Unit of Forensic Science Queensland (FSQ), an operational forensic laboratory accredited under the International Organization for Standardization/International Electrotechnical Commission (ISO/IEC) standard, ISO/IEC 17025. Each biological fluid type and subsequent recovery methods were performed separately to prevent cross-contamination of the swatches. Each body fluid (10 μL, neat and 1/1000) were dispensed centrally per section after vortex mixing initially, and pipette mixing between each deposition to ensure homogeneity (n = 10 per biological fluid dilution per recovery method, plus 10 control direct aliquots, (n = 360). The treated swatches were left in a sterile cabinet overnight to dry.

Fig. 1.

Illustration of 175 mm × 50 mm cotton swatch divided into five equal segments (A, B, C, D, E) with two subsegments for each recovery method, (not to scale).

2.2. DNA recovery methods

Two samples were collected from each swatch per DNA recovery method (the single wet swabbing method assessed two different swabs). The four methods were applied to each swatch for uniformity, with methods rotating across the swatches ensuring the fabric morphology would not bias any one method. Two sub-sections were cut from the swatch and placed in petri dishes before recovery for the scraping method to prevent transfer of the biological material to other sections of the fabric swatches. Each recovered sample was transferred to a sterile screw cap 2 L microfuge tube and stored at −20°C until DNA extraction was performed.

The sampling devices (Fig. 2) included two swabs, and a tape-lift that are routinely processed for DNA testing at FSQ: the MWE Sterile Tubed Dryswab™ Rayon swab 1028 (hereafter referred to as the MWE swab), the Puritan® 6” Sterile Rayon Tipped Applicator w/Polystyrene Handle 25–806 (hereafter referred to as the Puritan swab) and the HazPak Tapelift Kit (Dandenong, Victoria Australia). The MWE swab is included in the Forensic Medical Examination Kit (FMEK) and the Puritan swab is used for evidence recovery by the Queensland Police Service (QPS).

Fig. 2.

The sampling devices compared in this study. Top to bottom: MWE DRYSWAB™ (A): Standard Plastic Rayon – Labelled Tube (MW1028). Puritan (B): Puritan 6″ Sterile Standard Rayon Swab w/Polystyrene Handle (25-806), HazPak tapelift (C), (packaged with gloves) and carbon steel surgical blade (D). All sampling equipment are DNA free forensic grade treated with ethylene oxide (A to C), or gamma irradiated (D).

The single wet swabbing method was performed for both EtO treated swabs: by wetting the tip with 100 μL of ultrapure water then rubbing one side in a circular motion 12 times before rotating to the other side and repeating. This method was chosen after experimentation with water quantity and sampling techniques indicated less water did not sufficiently wet the swabs and more rotation of the swabs resulted in fibre fraying.

For the excision method, a 5 mm × 5 mm section was excised from the material using a single-use surgical blade and transferred to the microfuge tube with the blade. For tape-lifting, the adhesive part of the EtO treated HazPak Tape-lift was applied over the stain and pressed consistently with moderate pressure, five times to ensure the whole area of adhesive was applied to the stain. This method was chosen due to results of a tape-lifting study finding no significant difference in median allele counts between one and 32 tape applications with increased pressing negatively affecting allele recovery [5]. The tape-lift was then rolled with the adhesive on the inside and placed into a tube. The scraping technique was chosen following a trial of forward and backward motions of the blade on the cotton fabric. The optimal technique for collecting sufficient fibres without gouging the fabric involved back scraping from each cardinal direction 20 times to lift the fibres, then forward scraping the surface until a small amount of fibre was evident and able to be transferred to the tube.

2.3. DNA extraction, quantification, polymerase chain reaction (PCR) amplification and STR profiling

Procedures were performed as per FSQ standard operating procedures (SOPs). DNA extractions were completed using the Maxwell® FSC Instrument (Promega) liquid handling platform (LHP) and the DNA IQ™ System (Promega) for semen samples. The blood and saliva sample extractions were performed with the QIASymphony SP Instrument (QIAGEN) using QIASymphony DNA Investigator Kit (QIAGEN). DNA quantification was completed using the QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific) and the Quantifiler™ Trio DNA Quantification Kit (Thermo Fisher Scientific). Amplifications were performed with the PowerPlex® 21 System (PP21) with 30 cycles. Capillary electrophoresis was completed using the Applied Biosystems™ 3500xL Genetic Analyser (Thermo Fisher Scientific) and resultant DNA profiles were analysed with GeneMapper™ ID-X Software v1.6 (Thermo Fisher Scientific).

2.4. Data analyses

Data analyses were performed in GraphPad Prism (version 10.4.1 (627)), GraphPad Software, California). The p-values were calculated with a One-Way ANOVA, followed by Tukey's post-hoc test. Where a statistically significant difference (p < 0.05) was determined between methods, an effect size was estimated with eta squared, (η²), calculated from the ANOVA sum of squares for the effect divided by the total sum of squares (η² = SSeffect/SS total). Effect sizes for η² are described as large (0.14), medium (0.06) and small (0.01) [14].

For each sampling method, the total DNA yield was calculated by multiplying the quantification results by the elution volume (60 μL). A total percentage DNA recovery was then estimated as follows:

$$\%Recovery=\frac{AverageDNAyieldfromfabricsamplecollectedwiththemethod}{AverageDNAyieldfromdirectextraction}\mathrm{x}100$$

The degradation index (DI) was calculated as a ratio of the small (SAT) and large (LAT) autosomal targets by the Quantifiler™ Trio DNA Quantification Kit. DNA degradation has been scored according to Vernarecci et al. as: 0 - 1.5 = no degradation; 1.5 - 4 = slightly degraded, 4 - 10 = moderate degradation and >10 indicates high degradation (or PCR inhibition) [15]. The profile quality was evaluated by two measures; the profile intensity, assessed by comparison of the average relative fluorescent units (RFU) at each locus and the profile balance, qualitatively inspected to ascertain if there are any variations between methods.

3. Results

All positive and negative controls performed as expected. For each biological material dataset, outliers were identified and removed using the robust regression and outlier removal (ROUT) method where the ROUT coefficient Q was equal to 1 % [16]. Most of the 1 in 1000 dilution samples returned undetermined DNA values, rendering the dataset unsuitable for statistical comparisons.

3.1. Quantitative comparison of efficiency between recovery methods

3.1.1. Blood

For the neat blood samples all ten replicates for each recovery method generated quantifiable DNA. Compared to the direct extraction of neat blood, the percentage DNA recovery was no more than 5 % for each method apart from excision at 59.8 %. Fig. 3a displays the mean total DNA yield in nanograms for each method. Both swabs’ mean total yields (MWE = 27.53 ng ± 9.56, Puritan = 26.09 ng ± 13.12) were slightly higher than tape-lifting (12.67 ng ± 3.43) or scraping (11.56 ng ± 6.16). However, no statistical significance was found between these methods. Excision was found to recover a statistically significant higher DNA yield (335.9 ± 84.47) compared to tape-lifting, scraping, MWE and Puritan swabs, F (5, 54) = 210.1, p < 0.0001, with a large effect size (ŋ² > 0.95). This finding is consistent with Kuffel et al. [10] when comparing excision with swabbing blood from cotton.

Fig. 3.

Total DNA yield per method for the recovery of (a) neat blood and (b) 1/1000 blood dilution from 100 % white calico cotton (n = 10 per recovery method, except for 1/1000 excision = 9). Circles represent each data point in the series, the line = median, x = mean and whiskers represent the upper and lower quartile. ***F (5, 54) = 210.1, p < 0.0001, with a large effect size (ŋ2 > 0.95).

The 1/1000 dilution of blood returned values for each excision replicate although one outlier was removed via the ROUT method. All other methods returned between two to four datapoints with the others undetermined due to insufficient DNA for quantification (Fig. 3b). The percentage DNA recovery appeared higher for all methods at this dilution, possibly due to fewer samples recovered. The MWE swab yielded a mean DNA quantity of 0.05 ng ± 0.04 and percent recovery of 28.8 % which was over four times the Puritan swab and the tape-lift. Scraping yielded 0.02 ng ± 0.02, and 9.29 %, around twice the amount than the Puritan swab (0.01 ng ± 0.001, 6.5 %), or tape-lift (0.01 ng ± 0.007, 5.67 %). However, excision exceeded all other methods with a DNA yield at 0.17 ng ± 0.17 and 90.2 % recovery, providing strong support that it is the most efficient DNA recovery method for blood from white cotton.

The DI for all neat blood samples did not exceed a mean value of 0.73, indicating that no degradation had occurred due to the different recovery methods. The mean DI was higher for the 1/1000 dilution samples. The direct extraction, Puritan swab and excision all ranged from 1.1 to 1.3 indicating no degradation. However, the results per sample varied considerably with the highest replicate DI for excision at 3.7 and direct extraction at 3.1, both in the slight degradation range (see supplementary material A for full data). The MWE swab, tape-lift and scraping samples’ means fell into the slight degradation category, with the highest individual sample result for scraping at 5.0, indicative of moderate degradation. There are too few samples at this dilution to perform statistical analyses or draw correlations between the methods and DI. The 1/1000 dilution sample deposition and recovery were performed a month after the neat samples, so the differences may be due to natural degradation over time of the donor blood despite their appropriate storage during that time.

3.1.2. Semen

Each replicate for the neat semen samples generated quantification data. However, two were identified as outliers by ROUT and removed from the analyses (one each from the excision samples and direct extraction). Compared to the direct extraction of neat semen, the percentage DNA recovery was very low at no more than 1.3 % for each method except for excision at 55.6 %. The mean total DNA yields (Fig. 4a) were as follows: MWE swab 3.79 ng ± 1.81, Puritan 2.44 ng ± 1.12, excision 283.2 ng ± 46.7, tape-lifting 7.42 ng ± 3.07 and scraping 7.04 ng ± 6.76. A higher statistically significant result was found for excision compared to the MWE, Puritan, tape-lifting and scraping methods, F (5, 52) = 276.4, p = <0.0001, with a large effect size (ŋ² = 0.96). No statistical difference was found between any other method comparisons.

Fig. 4.

Total DNA yield per method for the recovery of (a) neat semen and (b) 1/1000 semen dilution from 100 % white calico cotton (n = 10 per recovery method). Circles represent each data point in the series, the line = median, x = mean and whiskers represent the upper and lower quartile. ***F (5, 52) = 276.4, p < 0.0001, with a large effect size (ŋ2 = 0.96).

For the 1/1000 dilution, both swabs and the tape-lift recovered five samples each. Excision recovered results for all ten replicates and scraping generated seven results after one outlier was removed via the ROUT method. The mean DNA yield (Fig. 4b) and percentage DNA recovery for the MWE and Puritan swabs were 0.011 ng ± 0.006, 3.25 % and 0.007 ng ± 0.003, 2.33 %, respectively. Tape-lifting recovered 0.013 ng ± 0.008, 3.85 % and scraping recovered 0.024 ng ± 0.012, 7.3 %. Excision again outperformed all methods at 0.31 ng, 93 %.

The DI for all neat semen samples did not exceed a mean value of 0.8 with small variance between replicates, indicating that no degradation had occurred. The highest DI was for one scraping sample at 1.01. The mean DI was marginally higher for the 1/1000 dilution samples, ranging from 0.58 for the MWE swab to 2.4 for scraping. However, one scraping sample had a DI of 11.0, indicating high degradation which was far higher than the second highest scraping DI value at 2.58. This sample had a total SAT DNA yield greater than the other replicates, but this did not equate to greater allele recovery (8) or to contamination from exogenous alleles (see supplementary material B). Consequently, the sample was removed as an extreme outlier before generating the descriptive statistics and data visualisation. As with the blood results, the trend remains consistent between dilutions. One difference between them is that the swabs performed slightly better for DNA yield than the tape-lift or scraping for blood, but the reverse was observed for semen.

3.1.3. Saliva

The neat saliva results displayed an unexpected pattern where the DNA quantity is higher for the excision samples compared to the direct aliquot results. This was first considered an error. However, the pattern persisted with the 1/1000 dilution. No indication within the literature has been found to account for these results. However, it is possible that microbial growth and/or enzymatic activity had occurred within the saliva aliquots kept at 4°C overnight while the deposits on cotton swatches have ceased activity as they dried, thus preserving the DNA. The oral mucosa has a large microbial biota, with one study estimating that saliva contains approximately 1.7 × 10⁷ bacterial cells per millilitre with each epithelial cell hosting on average over 100 individual bacteria [17]. Further testing to confirm this would need to be undertaken.

For the neat saliva (Fig. 5a), each recovery method generated results for all ten replicates. There was a higher percentage DNA recovery for the swabs than for blood and semen; MWE (2.08 ng ± 0.74, 23.8 %), Puritan (1.54 ng ± 0.47, 17.6 %). Scraping recovered the least (1.06 ng ± 0.62, 12.1 %). Excision recovered the most DNA (19 ng ± 3.59) which was observed to be higher than the direct extraction (11.9 ng ± 2.34), equating to more than 1.5 times the yield and was statistically significant compared to the other recovery methods, F (5,54) = 49.06, p = < 0.0001, with a large effect size (ŋ² = 0.82). Tape-lifting recovered 5.20 ng ± 3.43, 59.5 %, which was significant compared to scraping (p = 0.0012), the MWE swab (p = 0.0145) and to the Puritan swab (p = 0.0046). As the tape-lifting results were significant, further testing was performed with Welch's t-tests to estimate the effect size between each method: tape-lift versus MWE swab F (9, 9 = 21.47), ŋ² = 0.45, tape-lift versus Puritan swab F (9, 9 = 52.80), ŋ² = 0.54 and tape-lift versus scraping, F (9, 9 = 30.44), ŋ² = 0.60. Therefore, all classed as large effect sizes, however, give an indication for ranking the methods: excision > tape-lifting > MWE > Puritan > Scraping.

Fig. 5.

Total DNA yield per method for the recovery of neat saliva (a) and 1/1000 saliva dilution (b) from 100 % white calico cotton. (n = 10 per recovery method). Circles represent each data point in the series, the line = median, x = mean, and the whiskers represent the upper and lower quartile. ***F (5,54) = 49.06, p < 0.0001, with a large effect size (ŋ2 = 0.82). **p < 0.05.

The 1/1000 dilution of saliva (Fig. 5b) did not amplify any of the replicates for either swab, tape-lift, or the scraping samples. Six replicates generated results for the direct extraction with a mean total DNA of 0.007 ng ± 0.003. Excision returned results for four replicates with a mean of 0.009 ng ± 0.004, representing 1.25 times the recovery as compared to the direct extraction.

The mean DI for the neat saliva samples ranged from 0.92 to 1.12 (n = 10 per method), indicating no degradation. The 1/1000 direct extraction replicate DIs ranged from 0.38 to 3.01 (n = 4), with a mean of 2.10, indicating slight degradation. However, three of those replicates failed to amplify any alleles. Three samples that generated one allele each failed to generate a DI. The sample that had a DI of 3.01 was the only sample that generated data for all the quantification data and amplified a single allele. The excision results generated a DI for four out of five that amplified alleles ranging from 0.53 to 3.15 with a mean of 1.48. Counterintuitively, the sample that produced the highest DI also amplified the most alleles (see supplementary material C).

3.1.4. STR profile quality

The DNA yields from the neat biological materials indicated that full profiles were likely to be generated for all recovery methods, therefore, these samples were not progressed to amplification. As the quantification results for the 1/1000 dilutions were close to, or lower than the limit of detection (LoD) of 0.0006 ng/μL [18] for most samples, this was determined to be the most informative level for genotyping. The large difference observed between excision and the other recovery methods at this level was also instructive. Mean percentage allele recovery and mean RFU values per biological fluid (1/1000 dilution) are presented in Fig. 6, Fig. 7, Fig. 8. As many replicates failed to produce amplicons, the mean value was calculated excluding those samples to give a more accurate reflection of the mean values per successful amplification. Full profiles concordant with the donor profile were found for all direct extraction replicates of semen and close to full profiles for blood, whereas saliva at this dilution produced only four alleles across four samples. Excision produced the greatest number of alleles for all biological fluids. Saliva direct samples again followed the pattern observed with DNA yield in that they generated a lower number of alleles than excision. A few samples amplified a single exogenous allele: three for blood, eight for semen and three for saliva. There did not appear to be any one recovery method that affected this result, and the exogenous alleles were omitted from the profile RFU calculations. Importantly, these alleles were not concordant with the researcher.

Fig. 6.

The percentage allele recovery (A) and mean RFU values (B) of 1/1000 dilution of blood recovery method comparison. Recovered replicates per method: MWE = 4, Puritan = 1, excision = 10, tape-lifting = 7, scraping = 9, direct extraction = 10.

Fig. 7.

The percentage allele recovery (A) and mean RFU values (B) of 1/1000 dilution of semen recovery method comparison. Recovered replicates per method: MWE = 6, Puritan = 7, excision = 10, tape-lifting = 9, scraping = 10, direct extraction = 10.

Fig. 8.

The percentage allele recovery (A) and mean RFU values (B) of 1/1000 dilution of saliva recovery method comparison. Recovered replicates per method: MWE = 0, Puritan = 0, excision = 5, tape-lift = 1, scraping = 0, direct extraction = 4.

3.1.5. Blood

The MWE swab generated four replicates with partial profiles (PPs) ranging from one to 31 alleles. Puritan recovered only one PP of 15 alleles. Excision recovered three full profiles and seven PPs ranging from 30 to 35 alleles. Tape-lifting recovered seven PPs with one to eight alleles and scraping generated nine PPs with one to five alleles. Furthermore, four samples had returned a quantification value, but no alleles were amplified. These samples consisted of three with a LAT only and one with a SAT only, and all below the 0.0006 ng/μL LoD threshold. Fig. 6 shows the allele recovery for blood as a percentage of the full profile and the mean RFU values. RFU values were highest for excision at 37723.1, followed by the MWE swab (3580.5), Puritan swab (1925), tape-lift (420.7) and scraping (457).

3.1.6. Semen

The MWE swab generated between 1 and 14 alleles per replicate (of five that returned results), with a mean of 4 alleles. Puritan produced 1-6 alleles across seven replicates, tape-lifting 1 to 15 alleles per replicate with an average of 6.6 and every replicate for scraping generated PPs ranging from two to 18 alleles. All excision samples produced full concordant profiles for every replicate, and all quantification targets generated results. Excision also had the highest mean RFU value (35836.5) compared to the other methods: MWE swab (564.5), Puritan swab (375.7), tape-lift (1018.8) and scraping (1407.3) (see Fig. 7B).

Across all recovery methods of the semen PPs, 17 were lacking a Y-target quantification value, but generated alleles concordant with the donor profile (from 1 to 11), which would not be sufficient to enter into the National Criminal Intelligence DNA Database (NCIDD) as the minimum recommendation is 12. These results could however be sufficient to use within the case for comparison to a suspect profile or for exclusionary purposes. Additionally, 21 samples reporting a SAT below threshold also amplified 1 to 16 alleles concordant with the donor profile. Furthermore, the scraping sample that produced 16 alleles had a quantification value of just 0.0004 ng/μL.

3.1.7. Saliva

Results indicated for both swabs and the scraping samples that the LoD had not been reached as STR profiling failed in all replicates. Only four replicates produced one allele each of the direct aliquot samples, and one tape-lifting replicate generated two alleles only. The same pattern emerged for the saliva profiling results as for DNA quantity; excision samples produced a higher mean allele recovery (2.8) and RFU (372) than the direct aliquots (1 and 90.75, respectively) and tape-lift (2 and 176, respectively). Five replicates reported quantification data without amplified alleles: two with a LAT only, and three with both the SAT and LAT data, all below the LoD.

4. Discussion

4.1. Blood

Experimental results indicate that the excision recovery method produced statistically greater DNA yields and profile percentage allele recovery than all other methods, which was supported by previous research [10]. Furthermore, the 1/1000 dilution of blood generated results for all 10 excision replicates, therefore, showing greater consistency of STR generation with the excision method. Moreover, at this dilution, fewer alleles were generated for the other methods. Therefore, diluted blood samples on cotton are most efficiently recovered using the excision method.

4.2. Semen

Semen recovery from cotton generated significantly higher DNA yields and RFU values for excision than the other methods, even though the method exhibited a great variation between the replicates. This variation was also displayed in the direct extraction of semen even though care was taken to mix the sample between each pipette aspiration. There was an observed separation of the seminal fluid, which could have resulted in some aliquots containing more of the viscous fraction than others and perhaps a greater spermatozoa count. It is important to note that the scraping and tape-lifting methods also generated 10 and 9 STR profiles from the replicates (see supplementary material B). This indicates the higher efficiency of the excision method is further supported as all 10 excision replicates produced full allele recovery (40 alleles), whereas scraping and tape-lifting produced 2-18 alleles and 1-15 alleles respectively.

Previous research has shown that spermatozoa can be retained within swab heads, with the mechanical and/or chemical properties of the swab fibres proposed as contributing factors. Mechanically, traditionally wound-fibre swabs such as cotton and rayon consist of a dense mass of fibres tightly spun around the shaft, creating a tortuous, high-surface-area network that physically traps cells and DNA limiting release for extraction [3,19]. In contrast, nylon flocked swabs comprise short fibres arranged perpendicular to the shaft in an open structure, promoting faster absorption and more complete sample release [20]. Benschop et al. [20] demonstrated this directly for semen, reporting more intact sperm nuclei, higher sperm cell counts, and greater DNA release, with only 9 % of total DNA retained in nylon flocked swabs compared with 30 % for cotton swabs. A chemical contribution to retention is also plausible. Cotton (natural) and rayon (synthetic) are cellulose-based fibres with broadly comparable texture and absorbance properties and may therefore interact similarly with biological material despite their different origins. The abundant surface hydroxyl groups of cellulose fibres can form hydrogen bonds with adsorbed biomolecules, potentially restricting DNA release from cellulose-based collection substrates during extraction [19,21]. However, spermatozoa lack the polysaccharide cell wall found in cotton fibres, which are more than 90 % crystalline cellulose [22]; instead, their plasma membrane is coated by a glycoprotein- and glycolipid-rich glycocalyx [23]. Any chemical interaction with cellulose fibres would therefore more plausibly involve hydrogen bonding between fibre hydroxyl groups and the glycan components of this glycocalyx; an interaction that has not yet been tested directly and warrants further study. Given the cellulosic similarity between the cotton substrate and the rayon swabs used in this study, the substantially greater allelic recovery achieved by scraping further supports fibre entrapment as the more likely explanation for the lower recovery obtained for swabs.

4.3. Saliva

The statistically high significance of the excision recovery method of saliva from cotton compared to wet swabbing is supported by other research [10]. Furthermore, the tape-lifting higher efficiency than wet-swabbing is also supported. Burmuzoska et al. [24] found the Lovell 3M tape recovered a significantly higher salivary DNA yield from cotton than the MWE rayon swab for a 1:50 dilution, but no significant difference for the 1:100 dilution, although both dilutions produced greater allele counts for the tape-lift [24]. However, the lack of significant differences between tape-lifting and swabbing neat saliva has been reported elsewhere for skin [25] and ammunition casings [26]. These findings may be connected to the substrate, but no other studies were found for recovering saliva from cotton comparing tape-lifts and swabs for analysis.

The present study was limited in scope to the validated protocols of FSQ, however future study of an expanded set of DNA extraction kits for saliva may result in better DNA preservation and extraction efficiencies, which could then be validated and implemented. DNA extraction studies for saliva have investigated various commercial kits, such as Oragene™ which has been shown to produce less fragmented DNA [27], and can preserve salivary DNA quality and quantity at room temperature for many months [28,29].

4.4. General method observations

Concerns regarding the scraping method were first raised in the Final Report of the 2022 Commission of Inquiry into Forensic DNA Testing in Queensland which highlighted the benefit of other more suitable methods for evidence recovery, explicitly citing health and safety, and contamination risks [12]. While no significant difference was found in total DNA yield between the tape lifting and scraping methods for blood or semen, a contamination risk with the scraping method was identified. On many occasions it was observed that the scraped fibres would statically ‘jump’ onto the researchers gloved hand or out of the Petri dish while trying to transfer them to the microfuge tube. Fibres were also witnessed separating from the aggregate collected and spreading across the Petri dish. In a sexual assault case where a pertinent area of underwear is sampled, this ‘jumping’ of fibres by scraping from another area such as the waist-band could lead to the cross contamination of exhibits or equipment, incorrect interpretation of DNA results, and miscarriages of justice such as in the case of Lukis Anderson [30]. Furthermore, it is often crucial in forensic casework to attribute discrete areas of staining to certain individuals relevant to multiple offender or complainant cases. Further still, issues at trial often arise in sexual assault cases regarding the attribution of one side of a surface to another, for example, the inside and outside of underpants. This is supporting evidence for establishing an alternative DNA recovery method to scraping, regardless of the DNA quantification and profiling results. However, the authors note that this phenomenon may be limited to the specific type of material-scraping interaction generating high static and warrants further testing with other fabrics, or other potential static mitigating processes.

In terms of general ease of use for each sampling method, wet-swabbing was simple and straight-forward, although snapping the head into a microfuge tube takes practice to ensure the correct stem length for robotic DNA extraction. This can be mediated by cutting the stem, but that could lead to potential contamination and increases labour. Tape-lifting was more labour intensive than swabbing or excision and it was challenging to roll the tape without touching the adhesive side, highlighting the importance of the provision of DNA-free gloves within the tape-lift kit. Excision was straightforward, although the necessity of holding down the fabric firmly to stop movement presents a risk while cutting the fabric with the blade. Some operators may find the method difficult to perform.

Finally, the scraping method was the most labour intensive to produce visible fibres for the fabric assessed in this study, also presenting an injury risk which raises concerns with regards to occupational work, health and safety, and requires considerable practice to adequately perform the technique. These risks are difficult to justify given the lower recovery associated with the technique.

In forensic casework, techniques should be utilised that maximise the chances of biological material and/or DNA recovery on the first examination. This streamlines investigations by the timely provision of forensic results thus enabling intelligence-based policing. Furthermore, critical sample material is also preserved for repeat testing and facilitates the application of further forensic techniques where necessary. Avoiding techniques with contamination and/or work, health and safety risks, offers opportunities to further expedite testing while removing risk to other irreplaceable exhibits or the examiner. Ultimately, the use of excision removes these risks while maximising DNA recovery and therefore, chances of successfully recovering an informative DNA profile which can then be immediately uploaded to a database for searching or compared to a person of interest profile.

4.5. Limitations

The scope of this study was limited to one common fabric substrate and three biological fluids to enable it to be completed within the timeframe of an Honours degree. However, as the literature has indicated, there is no one recovery method that is suitable to all substrates and biological material types (1). Further phases of this research will assess other common porous fabrics (poly-cotton, polyester fleece, nylon/spandex activewear and polyester carpet) to ascertain the most efficient recovery method for the biological fluids assessed here. Non-porous substrates should also be assessed as they are not conducive to the excision recovery method and research has shown that recovery methods for non-porous or solid porous substrates such as bricks and ammunition require different recovery techniques [[31], [32], [33]]. Furthermore, an assessment of saliva recovery from skin with the methods assessed here would provide value for case work where assaults include oral contact. Additionally, as the recovery of saliva failed to reach the LoD at the 1/1000 dilution, this warrants further assessment with a dilution between the two concentrations assessed here to investigate the differences more comprehensively between recovery methods. Finally, given its relevance to forensic casework, the inclusion of trace DNA containing cellular material is recommended.

A significant limitation to the present study was not assessing a male-female mixture. Future research of a mock sexual assault scenario is planned, comparing recovery methods and the standard DNA extraction with differential extraction. A preliminary sample set using neat saliva mixed with neat semen in a 9:1 ratio has been undertaken as part of the study (data not shown) which has also indicated excision to be the most efficient recovery method for DNA profiling.

The results of this research cannot be generalised to every forensic facility, however. They are explicitly linked to the DNA analysis, extraction, quantification, amplification, and capillary electrophoresis procedures validated at FSQ. Any researchers wishing to use this research as guidance must take those factors into consideration thus necessitating their own verification. Moreover, these results do not include male-female mixtures or trace DNA, and more research is necessary to include those sample types to optimise DNA recovery in sexual assault scenarios where mixtures are common.

5. Conclusion

The consistent trend observed for all three biological fluids tested at each dilution suggests that excision is the most efficient DNA recovery method for bleached white 100 % cotton substrate. The DNA yield was significantly greater for excision compared to all other recovery methods for each neat biological fluid and displayed higher quality STR profiles at the 1/1000 concentration. There were no statistical differences between either swab, tape-lifting or scraping methods for blood or semen. However, tape-lifting had a statistically significantly higher DNA yield compared to both swabs and scraping of neat saliva. Excision is thus the recommended recovery technique for biological fluid recovery from 100 % cotton fabric within the FSQ evidence recovery laboratory.

CRediT authorship contribution statement

Jane Whipp: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization, Visualization. Jeremy Watherston: Writing – review & editing, Supervision, Methodology, Conceptualization. Linzi Wilson-Wilde: Writing – review & editing, Supervision, Methodology, Conceptualization. Steve Smith: Writing – review & editing, Supervision. Dan N.O. Bonsu: Writing – review & editing, Supervision, Methodology, Conceptualization, Project administration.

Funding

Forensic Science Queensland funded this project in full. Jane Whipp was supported in part by Griffith University's Pro Vice Chancellor (Griffith Sciences) Academic Excellence Honours Scholarship.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article.