Swabs

Table 1 shows an overview of the five different swab materials that were evaluated: cotton, foam, rayon, polyester, and nylon flocked. These swabs were developed for biological (e.g., collection of bacteriologic specimens), cleaning or forensic purposes. Swabs #6‐#11 were provided by ITW Texwipe (Hoofddorp, NL), swabs #12 and #13 were provided by Copan (Italia, Brescia, Italy) and swab #14 was provided by bioTRADING Benelux B.V. Mijdrecht, NL.

DNA Samples

Multiple saliva samples from one donor were isolated using the GeneFiX™ Saliva‐Prep DNA Kit (provided by Isohelix Kent, UK) according to the manufacturers protocol. The concentration of double‐stranded DNA (dsDNA) of the original DNA extract obtained by the GeneFiX™ kit, and after extraction/recovery from the swabs, was determined with the AccuGreen™ High Sensitivity dsDNA Quantitation Kit (provided by Biotium, Fremont, CA, USA, via VWR, Amsterdam, NL). The fluorescence was measured using the Qubit 2.0 fluorometer (Thermo Fischer Scientific, Landsmeer, NL).

Swab Morphology

In order to determine the swab morphology, scanning electron microscopy (SEM) images were taken using a Jeol JSM (6010LA) SEM with an acceleration voltage of 10 kV and with a magnification fixed at 50× (Peabody, MA, USA).

Absorption Capacity

The absorption capacity (AC) was determined by placing the swab directly from the package for 1 min in a tube filled with TE‐buffer (10 mM Tris–HCl; 1 mM EDTA; pH 8.0; Merck, Darmstadt, Germany), where upon the tube was reweighed. With the known density of the TE‐buffer solution ρ (Supplemental Information), the absorption capacity (in μL) of the swabs was determined by:

(1)

where  is the difference in average weight (in g) of the tube filled with buffer () and the average weight (in g) of the tube with TE‐buffer after removing the swab (). Each swab was measured five times (n = 5). Details of the protocol are given in the Supplemental Information.

For the extraction and recovery efficiency experiments swabs with an absorption capacity above 100 μL were selected, to ensure that the swab absorbed a sufficient amount of sample and could be premoistened.

Extraction Efficiency

In order to determine the extraction efficiency (EE), 20 μL of DNA from a saliva sample isolated by the GeneFiX™ Saliva‐Prep DNA Kit, was pipetted directly onto each swab. After addition of the dsDNA to the swab, the swab head was cut off and placed in a 1.5 mL Eppendorf tube filled with 250 μL TE‐buffer and then vortexed for about 1 min. A negative control of each swab was taken by replacing the 20 μL of DNA with 20 μL of TE‐buffer. The extraction efficiency (in %) was calculated by:

(2)

where [sample] is the measured concentration of the DNA sample (in ng/μL), V_applied the volume (in μL) added to the swab (which is 20 μL, in this case), V_extraction the extraction volume (Supplemental Information) and [origin] the concentration of the original DNA extract (in ng/μL) as obtained with the GeneFix™ kit. Each swab was measured in triplicate (n = 3). More details for extraction and the following recovery protocols can be found in the Supplemental Information.

Recovery Efficiency

In order to determine the recovery efficiency (RE), 20 μL of DNA from a saliva sample isolated with the GeneFiX™ Saliva‐Prep DNA Kit, was pipetted onto a predefined area of a glass slide and dried for about 30 min at room temperature in a sterile laminar flow hood. Swabs were moistened by pipetting 50 μL of TE‐buffer onto the swab head. After the collection step (described in detail in the Supplemental Information), the swab head was cut off and placed in a 1.5 mL Eppendorf tube filled with 250 μL TE‐buffer and vortexed for about 1 min. A negative control was taken for each swab by replacing the 20 μL of DNA with 20 μL of TE‐buffer. RE was determined using an equation very similar to Eq. 2, Supplemental Information. Each swab was measured in triplicate (n = 3).

Swab Morphology

Scanning electron microscope images of a selection of swabs, to highlight the various design categories as already discussed, are shown in Fig. 1. Images of all tested swabs are presented in Figure S4 of the Supplemental Information. Swab #1 is the only cotton swab used in this research and consists of long wound fibers (Fig. 1A). Three wound rayon swabs, #3, #4, and #5, are tested in this research. The Transwab (#3 Fig. 1B) and the Transport swab (#5) are more densely packed than the Dryswab (#4). The polyester swabs, #7, #9, #10, and #11, show different types of packing (Fig. 1C,D). The Absorbond swab (#7) has a morphology of long wound fibers similar to the cotton swab. The other three swabs have a woven, more pad‐like structure. Swab #2, as well as #12 and #13 (Fig. 1E) are made of nylon‐flocked fibers, which are sprayed onto the swab. The three types of swab have a similar appearance. The nylon Microswab (#6) is a precision swab with a tip diameter of only 0.9 mm consisting of long and thin bundled filaments of nylon. #8 and #14 are foam swabs made of polyurethane and have a very open sponge‐like morphology (Fig. 1F).

Absorption Capacity

Most of the swabs show an absorption capacity well above 100 μL, Fig. 2. The Large Alpha polyester swab (#11) has a much larger swabbing area compared to the other swabs, which results in a capacity above 200 μL. Five swabs have a capacity lower than 100 μL, but note that the nylon Microswab (#6) and the polyester Honeycomb swab (#9) have a much smaller swabbing area than the other tested swabs. The rayon Transport swab (#5) and Absorbond polyester swab (#7) do not have an open structure, the fibers are closely packed (see the scanning electron microscope images in Figure S4 in the Supplemental Information). The variation in absorption capacity as found in Fig. 2 does not show a clear relation with swab morphology (see the scanning electron microscope images in Figure S4 in the Supplemental Information). This is due to the fact that the performance of a swab depends not only on swab tip material, but also on the volume and the porosity of the sorbent material, as well as on how tightly the material is wound onto the swab shaft 1.

Extraction Efficiency

As mentioned before, for the extraction efficiency experiments swabs with an absorption capacity above 100 μL were selected in order to ensure that the swabs absorb sufficient amounts of the sample. The extraction efficiency of these swabs, see Fig. 3 and Table 2, varied greatly with values of between 6% (#13) and 48% (#12), with the nylon flocked 4N6FLOQ swabs showing both the lowest and the highest percentage. It is rather striking that none of the swabs showed a value above 50%, which means that the largest part of the DNA sample remains on the swab. An important point that should be noted here is that in forensic practice it is common to use only one extraction step. Furthermore, it is common to use only a limited extraction volume, to avoid unacceptable high dilution. Inevitably therefore, an extraction efficiency of 100% is impossible, because part of the solution that contains a proportional amount of the dissolved DNA will remain on the swab. A maximum theoretical efficiency was calculated based on the ratio of extraction volume and total volume applied and based on the assumption that the DNA did not adsorb onto the swab surface but was completely dissolved in the buffer solution contained on the swab. In Table S1 of the Supplemental Information the maximum theoretical extraction efficiency is compared to the measured extraction efficiency, from which it is concluded that a substantial part of the DNA stayed behind on the swab. The polyester Alpha swab (#10) and the nylon flocked 4N6FLOQSwab (#12) showed the highest extraction efficiency when the extraction efficiency obtained was compared to the maximum extraction efficiency. Also notable is the low efficiency of swabs #4 and #13, whereas these swabs were specifically designed for forensic purposes.

The measured DNA concentrations for each swab were very similar, with the exception of swab #13, Table 2. Although the swabs were pressed against the tube to obtain as much DNA as possible, the extraction volume for most swabs was less than determined in the absorption capacity experiment, due to the 1 min. of vortexing. Therefore, it was important to take the extraction volume into account in order to determine an accurate extraction efficiency.

A substantial part of the DNA stayed behind on the swabs, resulting in a low extraction efficiency. In literature higher efficiencies were reported for direct inoculation of the swabs with spores. According to Rose et al., using an extraction buffer with a volume of 5 mL, the direct inoculation with Bacillus anthracis spores of cotton, macrofoam and rayon swabs yielded a percentage above 90% and for polyester above 80% 2. Rayon swabs resulted in an extraction efficiency of about 76% with direct inoculation with B. anthracis spores and sonication as the extraction method 10. Adamowicz et al. 17 applied cotton swabs and obtained an extraction efficiency of 50% for a buccal cell suspension and only 20% for a blood sample. Marshall et al. compared the X‐Swab of Diomics with the 4N6FLOQSwab, by applying 1, 2, and 5 ng quantities in a volume of 10 μL on the collection devices. The X‐Swab showed an extraction efficiency of 65%, 48%, and 60% for the aforementioned amounts of DNA, respectively. The 4N6FLOQSwab could recover 85%, 55%, and 32% for these amounts of DNA 19.

Recovery Efficiency

For the recovery efficiency experiments swabs with an absorption capacity above 100 μL were selected, in order to ensure that the swabs absorbed a sufficient amount of sample and could be premoistened. The recovery efficiency of most of these swabs (Fig. 4) was in the range of 20–40%, which meant that the majority of the sample could not be retrieved, a finding confirmed by comparison of the maximum theoretical recovery efficiency (Table S1, Supplemental Information) with the measured recovery efficiency. The nylon‐flocked 4N6FLOQSwab (#12) showed the best recovery efficiency, 47%, the nylon‐flocked 4N6FLOQSwab for crime scene samples (#13) showed the worst performance with an efficiency of only 2%. This was most likely caused by direct binding of DNA to the surface of this swab, in combination with a buffer that was unable to release the bound DNA effectively. However, it should be possible to recover this DNA by a second extraction step in a more suitable buffer. The manufacturer of the nylon flocked 4N6FLOQ Swabs advises the use of a Nucleic Acid Optimizer (NAO) to increase the recovery of DNA from the swab. The NAO is a semi‐permeable basket that is applied in combination with high‐speed centrifugation to collect the eluted DNA sample from the swab 24. Optimization of the recovery efficiency was not the goal of this study; therefore, such a method was not further investigated.

Again the polyester Alpha swab (#10) and the nylon flocked 4N6FLOQSwab (#12), together with the nylon flocked eSwab (#2), the rayon Transwab (#3) and the CleanFoam swab (#8), showed the highest extraction efficiency when the extraction efficiency obtained was compared to the maximum extraction efficiency (Table S1, Supplemental Information). As with the extraction efficiency, the recovery efficiency for the measured DNA concentrations did not differ much between the different swab types (Table 2).

DNA recovery efficiencies reported in the literature do vary somewhat, but are usually around 50%. Bacillus anthracis spores were collected from nonporous stainless steel surfaces by Rose et al., resulting in an efficiency above 40% for cotton and macrofoam swabs that were premoistened and vortexed for 2 min 2. For the polyester and rayon swabs an efficiency of around 10% was obtained under the same circumstances 2. Probst et al. 2 also studied Bacillus anthracis spores and obtained 49% recovery with an improved extraction protocol using nylon‐flocked swabs applied to a stainless steel surface. Using the old NASA protocol only about 30% could be recovered, but still higher than with a cotton swab, which only gave a recovery of about 13% 7. The recovery of Bacillus anthracis spores collected with a rayon swab from a stainless steel surface and from a painted wallboard surface resulted both in an efficiency of about 41% according to Brown et al. 10. Using optimal storage conditions Garvin et al. 16 obtained recovery efficiencies of a human saliva sample of 95% and 54% for cotton and viscose swabs, respectively. However, upon isolation the control sample only contained 52% of the predicted amount of DNA. Recovery with cotton and nylon swabs from a petri dish of a saliva sample turned out to be around 60%, an experiment in which a high quantity sample of DNA (more than 500 ng) was deposited 3. Edmonds et al. 8 compared liquid‐deposition and dry aerosol‐deposition of spores on glass. Cotton, dacron, rayon, and foam all gave an efficiency above 60% and 80% for the dry and liquid deposition, respectively. The reported recovery efficiency of the different swap types for bacterial spores and body fluids is in the same range as that for pure DNA (Table 2).

Additional Remarks

A NanoDrop instrument (NanoDrop1000, Thermo Fischer Scientific) can be used to quantify the concentration of DNA in a sample by measuring light absorption of a sample at 260 nm. RNA, ssDNA, and dsDNA all absorb at this wavelength and therefore this method is unable to discriminate between these types of nucleic acids. However, during a NanoDrop measurement it was found that an unknown compound that also absorbs at 260 nm must have leached from the majority of the swabs, Table 2. Therefore, this technique was discarded as a way to determine the extraction and recovery efficiency of swabs.

In order to check if contamination had taken place, negative controls were taken during the extraction and recovery experiments. Indeed some of these blanks gave a substantial concentration value (Table S2, Supplemental Information). A side cutter was used to cut swabs that did not have a suitable breaking point, and tweezers were used to pull out the swab from the tubes. To exclude these cutting tools as a source of DNA contamination, swabs were placed in 200 μL of TE‐buffer for 1 min and then pressed against the side wall to obtain as much extraction volume as possible. The same swabs used in the previous extraction and recovery experiments gave a positive value (Table S2, Supplemental Information) when measured with the Qubit assay. In order to validate if there was any DNA contamination on the swabs a multiple displacement amplification (MDA) reaction was carried out. MDA is an isothermal amplification reaction, which makes use of random hexamer primers that amplify all DNA present in a sample. The applied MDA protocol is described in the Supplemental Information, together with the amplification curve (Figure S5). The DNA concentration of the swab samples was determined prior to amplification with the AccuGreen™ High Sensitivity dsDNA Quantitation Kit and the Qubit assay (Table S2, Supplemental Information). After 1.5 h amplification none of the swab samples showed a positive signal, whilst the positive controls with a similar concentration as the swab samples (based on the Qubit measurements) did show a proper amplification curve. This suggests that the swab material itself caused a false positive reaction with the dye in the AccuGreen™ High Sensitivity dsDNA Quantitation Kit. It is possible that the dye binds to fibers released from the swabs into the extraction volume. The foam swab (#8) did not give a positive value in any of the blank controls.

The performance of several types of swabs may partially be explained by their chemical structure. Cotton and rayon have the same chemical structure (Figure S6, Supplemental Information), but differ in the degree of polymerization with values of 400–700 and 5000 for rayon and cotton, respectively. The polymer in these swab materials contains several O‐H groups, of which it is known that they form hydrogen bonds with nucleic acid chains. Nylon, a polyamide, contains N–H groups that are also known to form hydrogen bonds with nucleic acids. Polyester and polyurethane (foam) have polar C=O groups, which give only weak dipole‐dipole interaction 14, 25. It is likely though that the absorption capacity of the swabs, even more than on swab material composition, depends on swab tip dimensions and morphology. This is in line with Verdon et al. and Brownlow et al. who conclude that the amount of DNA collected is inversely proportional to the density of the fibers on the swab 1, 3.

Higher DNA recovery efficiencies were obtained by premoistening the swab with e.g., water 1, 3, 12, 23. The exact efficiency, and therefore the choice of most optimal swab material, is affected by the type of sample (epithelial cells, blood, or saliva), the porosity of the surface, the swab method (use of double swab technique or swab premoistening), swabbing medium (water, isopropanol or other), swabbing angle and the extraction protocol (vortexing or sonication, type of extraction kit) (1–3, 7, 10–12, 14, 17, 18, 23). It should be noted the recovery of more DNA would not necessarily result in good or complete DNA profiles 13.

In addition to DNA loss on the swab itself, determined by the extraction efficiency, some loss of DNA is always to be expected during swabbing, i.e., some DNA on a swabbed surface will not be picked up by the swab. This is most likely the reason that the recovery efficiency was lower for the foam (#8) and two of the nylon flocked swabs (#12 and #13). Note that the number of swabs tested per efficiency experiment was low (n = 5) and the standard deviation relatively high (for some experiments more than 10%). Most of the swabs tested were acquired through donation and therefore only a limited number of swab types was available for testing. In order to achieve a better statistical variance between experiments the efficiencies of two types, the nylon flocked 4N6FLOQSwab (#12) and the foam swab (#8), were tested with a larger data set (n = 10). The extraction and recovery efficiency of the nylon‐flocked 4N6FLOQSwab for this data set were 44.1 ± 7.3% and 40.6 ± 2.1%, respectively. In comparison to the n = 3 test the standard deviation of the extraction efficiency increased, while the recovery efficiency became lower. The n = 10 extraction and recovery efficiencies of the foam swab were 30.6 ± 0.5% and 23.8 ± 6.9%, respectively. Both standard deviations became lower in comparison to the n = 3 test. These standard deviation fluctuations seem to suggest that the variation in extraction and recovery efficiency within one batch of the same swab is relatively high. Nevertheless, it can still be concluded that for many swab types the largest fraction of the DNA sample remained on the swab and could not be retrieved with the chosen protocol.

Alternative options for sample collection that are now available include swabs made of a material that collects high amounts of cells/DNA and subsequently dissolves in the extraction buffer. The recently available X‐Swab from Diomics Corporation dissolves during incubation at 56°C for 1 h. Reported results for blood and saliva samples with this swab show a higher yield and average peak heights after profiling as compared to the 4N6FLOQSwab, whereas there was an indication that the dissolved swab material enhanced the PCR yield 19. Instead of swabs, tape has sometimes been used for sample collection in forensic cases 11, 13. These minitapes are completely soluble in water, show higher DNA concentrations and full profiles, but a drawback is that the extract is gel‐like, which makes pipetting impossible 13.