Comparison of swab types for collection and analysis of microorganisms

Abstract The human microbiome has begun to emerge as a potential forensic tool, with varied applications ranging from unique identification to investigative leads that link individuals and/or locations. The relative abundance of the combined DNA of the microbiome, compared to human nuclear DNA, may expand potential sources of biological evidence, especially in cases with transfer or low‐copy number DNA samples. This work sought to determine the optimal swab type for the collection and analysis of microorganisms. A bacterium (Proteus mirabilis) was deposited by pipette onto four swab types (cotton, flocked, dental applicators, and dissolvable), and extraction and real‐time PCR quantitation of the bacterial DNA were performed, which allowed for absolute microbial DNA recovery and comparison of yields across the four sampling substrates. Flocked swabs had the highest yield (~1240 ng) compared to the cotton swabs (~184 ng), dental applicators (~533 ng), and dissolvable swabs (~430 ng). The collection efficiency was further evaluated for cotton and flocked swabs using dried microbial samples spotted onto non‐porous surfaces (treated wood, glass, plastic, and tile). Flocked swabs performed consistently better across wood, glass, and tile, but showed decreased recovery from plastic. The cotton swabs failed in the recovery of P. mirabilis DNA across all surfaces. Knowing the appropriate sampling substrate will be useful as others continue to investigate the use of the microbiome as a forensics tool.

involve the collection of the biological material from a surface, followed by its subsequent release from the collection substrate, and analysis. Much research has been conducted on the optimal sampling substrate for use in traditional forensic DNA analysis and body fluid identification (Adamowicz et al., 2014;Luna, 2017;Viviano et al., 2018;Voorhees et al., 2006); however, the matter has not yet been extensively studied for the collection of the microbiome.
Swabbing and tape-stripping are comparable methods for sampling the microbiome (Ogai et al., 2018), but the potential inhibitory effects that adhesives can have on DNA extraction and downstream PCR amplification may make swabbing preferable. However, different types of swabs hold and release biological material differently (Bruijns et al., 2018). The objective of this study was to determine the optimal swab type for collection and analysis of the microbiome by comparing traditional cotton, nylon flocked, dissolvable swabs, and dental applicators. Despite being inefficient at releasing biological material during extraction processes (Adamowicz et al., 2014;Bruijns et al., 2018;Viviano et al., 2018;Voorhees et al., 2006), cotton swabs are widely available and used by the forensic community, even though other swab types may lead to better sample recovery. With perpendicular fibres and no internal mattress core, flocked swabs are designed for the effective collection and elution of samples (COPAN Diagnostics Inc, 2020). Dissolvable swabs, made from cellulose acetate, are soluble in buffers that contain chaotropic salts, like guanidinium thiocyanate used in commercially available nucleic acid extraction kits (Luna, 2017). Dental applicators can be brushes of various sizes but are typically composed of non-absorbent nylon flocking adhered to a spherical tip and used in the dental and make-up industries to apply various products (Safeco Dental Supply, 2020). Given the differences in the microbiome, and associated microbial DNA, compared to human nuclear DNA, it is also unknown if surface or sub-surface interactions will be similar (Alketbi & Goodwin, 2019;Verdon et al., 2013;Wood et al., 2017).
Thus, the collection efficiency from non-porous surfaces using the flocked and cotton swabs was also evaluated.

| Swab preparation
Proteus mirabilis is a bacterial species typically found in the gut microbiome and was used here as an available and representative bacterial component of the human microbiome. Proteus mirabilis was cultured, collected, washed, and pelleted via centrifugation to create a stock. Through prior quantitation and experimentation, it was determined that 10 µL of the uniformly mixed stock should result in DNA yields within the dynamic range of the qPCR standard curve (Wagner, 2021 Figure 1). Dental applicators were included due to their relatively small volume because it was thought that there may be fewer places for biological material to become trapped within the swab material (Safeco Dental Supply, 2020). In this case, the dental applicators used had a 2.0 mm head, while the traditional cotton swabs used were ~10.0 mm in length. The brand of dissolvable swabs used here was in development. According to the manufacturer, the provided prototypes contained ~20 mg total cellulase acetate fibre (pictured unshaven in Figure 1), but the fibre material entering the extraction reaction should be minimized to 5-7 mg, which required manual shaving of the dissolvable swabs to ~¼ the original size before use. As a positive control, eight replicates of the same P. mirabilis stock were added to sterile microcentrifuge tubes without a swab substrate. A negative control, consisting of a sterile microcentrifuge tube serving as a reagent blank, was included in each round of extraction and processed in the same manner as test samples.

| Surface preparation and swabbing
Non-porous surfaces utilized for sample collection included:  3 × 3 cm 2 on each surface designated for swabbing with flocked swabs and seven designated for swabbing with cotton swabs. The appropriate single swab was pre-moistened with sterile, deionized water (~100 µl) and used to collect the P. mirabilis from the corresponding surface section. Swabbing was performed by applying firm and constant pressure and repeatedly rotating the swab throughout, for 30 s. After drying (~2 h), swab heads were removed from the sticks and transferred to 2.0 ml microcentrifuge tubes for storage at −10°C until further processing. A positive manipulation control, consisting of P. mirabilis spotted onto cotton or flocked swabs, and negative control, consisting of a sterile microcentrifuge tube serving as a reagent blank, were included in each round of extraction and processed in the same manner as test samples.

| DNA extraction
Bacterial DNA was extracted using the Applied Biosystems™

| DNA quantitation
All sample and control extracts (2 μl per sample) were quantified

| Data analysis
Using the qPCR determined concentrations, total mass (in nanograms) of DNA recovered from each extract, averages, standard deviations (SD), and standard error of the means (SEM) were calculated across each swab type using Excel® (Microsoft, version 2101).
Statistical analysis was performed using the statistical software R (v 4.0.2). Between-group comparisons were conducted by ANOVA (α = 0.05) followed by a Tukey's Honest Significant Difference (HSD) test, as warranted.

| RE SULTS
There was a statistically significant difference in the total mass of microbial DNA recovered between the four swab types

| DISCUSS ION
While not currently as individualizing as other traditional forensic methods, the microbiome has begun to emerge as a potential forensic tool. Due to the relative abundance of bacteria and other microorganisms compared to human nuclear DNA, the microbiome could play a particularly important role in cases where evidence may have been transferred through skin contact but fingerprints or low amounts of human genetic material have been left behind. While the bacterium used here is primarily a component of the human gut microbiome, rather than the skin microbiome, it was used as a representative gram-negative bacillus, and its collection may still be relevant in certain forensic case scenarios that involve evidence in the form of faeces and urine.
When the P. mirabilis was spotted directly onto each swab type, the highest average mass of microbial DNA post-extraction was observed with the flocked swabs, followed by the positive manipulation control, dental applicators, dissolvable swabs, and finally cotton swabs. Generally, flocked and cotton swabs, which have been compared to each other numerous times in the forensic field for the recovery of human DNA, performed roughly as expected (Adamowicz et al., 2014;Bruijns et al., 2018;Viviano et al., 2018). Yet, the overall results seem counterintuitive in many respects. The flocked swabs unexpectedly yielded greater total recoveries than the positive manipulation control (comparatively 186% of the positive manipulation control recovery- Table 1), which contained no swab material.
This is interesting when one notes that similar recoveries to the positive manipulation control were observed with the other low surface area groups, that is, dissolvable swabs and dental applicators (64% and 80% of the positive manipulation control recovery, respectively- Table 1). If the only factor in elution efficiency was the It should be noted that several limitations were surrounding the use of the dissolvable swabs that may have contributed to the variability and the lower-than-expected yields from this group.
The dissolvable swabs used here were a prototype, which required F I G U R E 2 Total bacterial DNA yield following the direct deposit of sample onto the swab for each swab type tested (cotton, flocked, dissolvable, and dental applicators) compared to the positive manipulation control. Results displayed represent the average of the eight trials, with outliers removed, ±SEM. Letter designations represent Tukey's HSD comparisons: the same letter designation means results are not statistically different; when letter designations differ between groups, the p-value is < 0.05. manual shaving to ~ ¼ the original size before use, and extraction volumes were doubled for the dissolvable swabs so as not to sat-

urate the system (recommendations provided by Luna Innovations
Incorporated™, personal correspondence 30 November 2020).
These precautions may not have been sufficient to prevent the dissolved, or partially dissolved, swab material from interfering with the bead-based extraction process. Furthermore, the MagMAX™ DNA Multi-Sample Ultra 2.0 Kit, which was selected for DNA extraction because it may be used to isolate DNA from a variety of cell types, including bacteria, and has the requisite guanidinium salt component for dissolvable swabs, is not routinely utilized in forensic analysis.
Theoretically, the forensic use of the MagMAX™ DNA Multi-Sample Ultra 2.0 Kit should not pose problems, assuming that it is properly and thoroughly validated.
It is recognized that another limitation was the decision to compare only two of the four swab types in the surface study.
Flocked swabs were chosen for their comparatively high yields in the initial elution study. Even though higher yields in total mass were recovered from the dental applicators (2.9 × the amount of cotton) and dissolvable swabs (2.3 × the amount of cotton), yields between the three remaining swab types (dental, dissolvable, and cotton) were not statistically different from each other.
Considering swab volume, the dental applicators and dissolvable swabs displayed a better release efficiency than the cotton swabs; however, due to their smaller sizes, there was also a greater risk of saturation when being used to recover materials from surfaces.
Furthermore, the dental applicators were of the same general construction (flocked nylon fibres) as the flocked swabs. While cotton swabs were the least effective at releasing the P. mirabilis DNA, in terms of absolute quantity, it was suspected that they would be the most efficient at collecting biological material off of surfaces.
For these reasons, and their greater use and availability in the forensic community compared to dental applicators and dissolvable swabs, cotton swabs were chosen for further evaluation in the surface study.
The flocked swabs performed comparatively well across glass, tile, and wood surfaces, but the amount of microbial DNA recovered from plastic was significantly less (Table 2). This may be due to the polypropylene composition of the plastic surface. If, as hypothesized above, an increased swab surface area contributes to bacterial cell dehydration and lysis, thus aiding in DNA recovery, this might also be occurring when samples are deposited and allowed to dry on various surfaces. The adsorption to and denaturation of DNA by plastics, particularly polypropylene, is a well-documented phenomenon in the forensic DNA community that results in apparent DNA loss (Belotserkovskii & Johnston, 1996, 1997Gaillard & Strauss, 1998, 2000Kline et al., 2005;Lecerf & Le Goff, 2010;Wang et al., 2019).
Cotton swabs continued to perform poorly when surface swabbing was incorporated, demonstrating no or decreased recoveries. The decrease in recoveries seen with cotton swabs was consistent with what was observed for the flocked swabs, which showed as much as a 95% decrease in yield when comparing recoveries from swabs without surface sampling to recoveries obtained following surface sampling (Table 2).
All prepared swabs, non-porous surfaces, and surface collection swabs were allowed to air-dry for ~2 h following their respective preparations. As all swabs and surfaces throughout the study were exposed to the same environmental conditions, any contamination or effect from other aerosolized microbes should have been equally distributed across samples. While contamination with aerosolized microbes is a possibility, the negative results observed when using cotton swabs to collect bacterial samples from the various nonporous surfaces demonstrate that the potential effect from environmental contamination is negligible. This is a critical finding for the potential future forensic application of the microbiome as swabbed surfaces will never be completely protected from aerosol contamination between sample deposition and collection. From the forensic standpoint, systemic contamination in the form of contaminated reagents and sample-to-sample contamination pose bigger threats, which are typically controlled for with the processing of reagent blanks.

| CON CLUS IONS
While the human DNA extraction efficiency of various swab types has been a topic of much research (Adamowicz et al., 2014;Bruijns et al., 2018;Viviano et al., 2018;Voorhees et al., 2006), sample elution is only half the story. Here, we also examine the efficiency of flocked and cotton swabs for sample recovery from various nonporous surfaces. Additionally, except for work by Ogai et al. (2018) that compared cotton swabbing to tape lifting for collection of the microbiome, research on the optimal collection substrate for the microbiome in a forensic or non-medical setting is scarce. While future research could focus specifically on the transfer and collection of samples relating to the skin microbiome, these data support moving toward the use of flocked swabs, and away from cotton swabs, for the collection and analysis of bacterial samples relating to forensic use of the microbiome.

E TH I C S S TATEM ENT
None required. University.

CO N FLI C T S O F I NTE R E S T
None declared. Formal analysis (lead); Methodology (equal); Supervision (lead);

AUTH O R CO NTR I B UTI O N S
Writing-review & editing (lead).

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in this published article.