J Forensic Sci, 2015

doi: 10.1111/1556-4029.12765
TECHNICAL NOTE Available online at: onlinelibrary.wiley.com

CRIMINALISTICS

Michael S. Adamowicz,1 Ph.D.; Ren

ae D. Labonte,2 M.S.; and John E. Schienman,3 Ph.D.

The Potential of Cosmetic Applicators

as a Source of DNA for Forensic Analysis

ABSTRACT: Personal products, such as toothbrushes, have been used as both known reference and evidentiary samples for forensic DNA
analysis. This study examined the viability of a broad selection of cosmetic applicators for use as targets for human DNA extraction and short
tandem repeat (STR) analysis using standard polymerase chain reaction (PCR) conditions. Applicator types included eyeliner smudgers, pencils
and crayons, eye shadow sponges, mascara wands, concealer wands, face makeup sponges, pads and brushes, lipsticks and balms, and lip gloss
wands. The quantity and quality of DNA extracted from each type of applicator were examined by assessing the number of loci successfully
amplified and the peak balance of the heterozygous alleles in each full STR profile. While degraded DNA, stochastic amplification, and PCR
inhibition were observed for some items, full STR profiles were developed for 14 of 76 applicators. The face makeup sponge applicators
yielded the highest proportional number of full STR profiles (4/7).

KEYWORDS: forensic science, cosmetic applicators, lipstick, lip balm, forensic DNA analysis, missing persons cases, nontraditional refer-
ence samples

Deoxyribonucleic acid (DNA) has the potential to be recov-
ered from numerous types of evidentiary items and surfaces in a
variety of forensic casework. Probative DNA samples can be
collected and profiles developed from traditional sources such as
blood, body fluids, hair, and tissue to items commonly submitted
to forensic DNA laboratories such as firearms (1,2), garments
(3,4), and household items (5). Increases in the sensitivity of
forensic DNA analysis methods have opened up the range of
potential evidentiary items to include just about anything that a
person may come into physical contact with (6). Added to this
list are the various items that people use to groom themselves
such as combs, hairbrushes, razors, and toothbrushes, and prod-
ucts used to alter their appearance, such as cosmetics. The recov-
ery of DNA from personal products, such as unwashed clothing,
toothbrushes, hairbrushes and razors, as well as different types
of antemortem medical specimens can be important for use in
missing person’s cases or in identifying victims of mass disasters
as well (7–10). While traditional known reference samples such
as bloodstains or buccal swabs would be preferable to identify
disaster victim remains, they are often unavailable to generate a
DNA profile for the victims of these types of events. Personal
products have been used in an effort to obtain the known refer-
ence DNA profile of a person that can then be compared with
DNA profiles generated from human remains or other items of
evidence submitted in identification cases. The known reference
1
Forensic Science Department, Henry C. Lee College of Criminal Justice
& Forensic Sciences, University of New Haven, 300 Boston Post Road, West
Haven, CT.
2
Molecular Pathology, Bio-Reference Laboratories Inc., 481 Edward H.
Ross Dr., Elmwood Park, NJ.
3
Division of Scientific Services, Connecticut Department of Emergency
Services and Public Protection, 278 Colony Street, Meriden, CT.
Received 31 Jan. 2014; and in revised form 19 June 2014; accepted 10
July 2014.
DNA profiles from personal effects can also be used in the
Combined DNA Index System (CODIS) database or local data-
bases in an attempt to identify human remains that were not pre-
viously identified by other methods (11). Prior work has been
done on the recovery and typing of DNA from personal products
including toothbrushes (12) and lipsticks (13). These items were
shown to be potentially useful sources of DNA. However, one
study also discovered that certain elements within lipstick could
cause high levels of fluorescent artifacts when performing short
tandem repeat (STR) analysis, and different methodologies of
extraction were recommended to maximize future results (13). In
light of the potential for an increase in fluorescent artifacts and
other electrophoretic anomalies with lipsticks, along with the
possibility that toothbrushes may not always be available for an
individual or may produce mixture results, other personal prod-
ucts such as cosmetic applicators should be considered for use
in forensic DNA testing. As well as having some efficacy as
nontraditional known samples, cosmetic applicators may also be
evidentiary items in cases of burglary, theft, and other crimes.
The recovery of DNA from touched items has been examined
extensively, and results have determined that the amount of DNA
recovered from contacted surfaces can vary for an individual (14–
16). On average, humans shed approximately 400,000 skin cells
per day (17) and these, as well as cell-free nucleic acid (18), are a
significant source of evidential DNA. As cosmetic applicators
repeatedly come into contact with the skin of the owner, skin cells
will be picked up and carried on the applicators. DNA profiles
should be readily generated from this potential evidentiary source
if enough genetic material can be recovered from the object. As
with other personal products, challenges to obtaining DNA from
cosmetic applicators may include low levels of DNA on the
objects, mixed DNA samples, and the presence of compounds that
are inhibitory to the polymerase chain reaction (PCR).
Samples containing degraded or low quantities of DNA are
common to forensic DNA analyses, and maximizing the amount

© 2015 American Academy of Forensic Sciences 1

2 JOURNAL OF FORENSIC SCIENCES
of DNA recovered has been an area of importance in forensic sci-
ence. In the study conducted by Tanaka et al. (12), it was deter-
mined that the amount of DNA recovered from most of the
toothbrushes, which had been used for at least 3 months by the
user at the time of collection, was sufficient for personal identifica-
tion. In the study conducted by Webb et al. (13), a variety of
brands and colors of lip cosmetics that were in use by the owners
at the time of collection were swabbed and processed in an attempt
to recover DNA. Although it was concluded that the samples that
generated results were good sources for DNA collection, less than
half of the samples gave a full DNA profile, the remaining samples
were either mixed profiles or failed to generate a profile at all and
artifacts were observed in some of the profiles that were indepen-
dent of brand and color (13). The findings of these researchers
clearly indicate that the quantity of usable DNA collected from
personal products can vary widely and maximizing the number of
candidate items for testing is prudent. A number of factors can
affect the amount of DNA that is recovered on such samples,
including environmental conditions, handling time, the contacted
surface, and individual differences in cell shedding among people
(17). Therefore, sample collection and storage are important issues
as well when working with personal effects for victim identifica-
tion or as evidentiary material. In general, however, if more cells
are shed there is a better chance of DNA recovery and these cells
are more likely to adhere to porous surfaces that come into contact
with the cell-shedding surface (17). Because of this, there is a high
probability that cosmetic applicators, such as sponges, wands, and
brushes, used on a daily basis will pick up sufficient quantities of
DNA-bearing cells or cell-free nucleic acid upon contact with the
skin. One possible issue with using personal products as a source
for forensic DNA analysis is that they may have come into contact
with another person’s skin surface in the past as a result of han-
dling or shared usage, and the possibility of mixed profiles could
be a complicating factor. This was demonstrated in a previous
study on lip cosmetics in which partial mixtures were detected in
some samples (13). According to Wickenheiser (17), mixture pro-
files do result in these situations, but the profile generated is usu-
ally representative of the last person to come into contact with that
surface. The owner of a cosmetic applicator is likely to be the last
person to use it, and therefore, their single-source profile would
result from the processing of the item.
Polymerase chain reaction inhibitors have already been identi-
fied as an obstacle to developing DNA profiles from a wide vari-
ety of items (19,20). Inhibitors can be extracted along with DNA
from samples, and subsequently affect the PCR by a variety of
mechanisms including the inhibitor binding to the polymerase or
the inhibitor interacting with the DNA or polymerase in some way
(20). Dyes have been identified as a common class of PCR-inhibit-
ing molecule. The indigo dye in fabrics such as denim often results
in an extract that is dark blue in color and can interfere with the
amplification process, leading to no detectable PCR products (21).
Similar results have also been observed for lip cosmetics (13).
Considering that all cosmetics have some type of dye or colorant
in them, the potential for PCR inhibition is significant. Although
PCR-inhibited samples are a subject of concern, full STR profiles
have previously been obtained from samples that had undetected
internal positive control (IPC) values from real-time quantitative
PCR (Q-PCR) assays, or sample inhibition from dyes within the
extracts (22) indicating that amplification of DNA is still possible.
Sample dilution, possibly in combination with filtration, is a com-
mon method used to reduce the amount of PCR inhibitors in a
sample (23). However, valuable DNA can be lost by diluting, and
this is certainly undesirable when the starting quantities of DNA
may already be at low levels. Alternative approaches to eliminat-
ing PCR inhibitors include extracting the DNA from the samples
using validated magnetic bead protocols, such as the Promega
DNA IQTM (24) or AB PrepFilerTM (25) kits. A study by Cupples
et al. (26) demonstrated that when using the QuantifilerTM kit for
Q-PCR, if the concentration was reported by the assay as “unde-
tected,” there was a greater chance for not obtaining a DNA pro-
file; however, some partial STR results could still be gained. This
research aimed to find the cosmetic applicator(s) that yielded the
highest quality and quantity of DNA that could be recovered,
while identifying the types of applicators that are most likely to
degrade the DNA or contain PCR inhibitors such that they could
be identified as poor candidates for collection and testing.
A broad range of cosmetic applicators were collected to exam-
ine which would be the most efficient choice for use as a nontra-
ditional reference sample or as an evidentiary item in a forensic
investigation. Applicator types included eyeliner sponge smud-
gers, eyeliner pencils/crayons, eye shadow sponges, mascara
wands, face makeup sponges, face makeup brushes, nonsponge
face makeup pads, concealer wands, lipsticks, lip gloss and lip-
stick wands, and lip balms.
Materials and Methods
Known buccal swabs and cosmetic applicators were collected
from participants who volunteered to donate materials for this
study. All samples were collected according to the Institutional
Review Board (IRB)-approved ethical standards and methods.
Sample Collection and Preparation
Buccal swabs were collected from each of the participants
who contributed cosmetic applicators for reference sample com-
parison to the profiles developed from their respective items. All
of the participants were females with the exception of the donor
of the control liquid blood, who was male. Sample collection
varied depending on the cosmetic applicator type. The area most
likely to come into contact with the skin was determined by
observing an example of each type of item in use. A total of 76
separate cosmetic applicators of differing types and brands were
tested in this study. Once collected, if the sample was not pro-
cessed immediately, all items were stored at 20°C to preserve
the DNA on the applicators.
The eyeliner sponge smudgers were completely cut away from
their respective sticks with clean scissors for extraction. Samples
from eyeliners were collected by swabbing the pencil or crayon
area with a sterile cotton swab moistened with deionized water.
One of the collected eyeliners, sample 9-2, had no crayon left.
This item was swabbed around the rim that came into contact
with the eyelid during makeup application as well as being
swabbed along the inside of the applicator’s cap with a moist-
ened sterile cotton swab. Due to the different sponge shapes
used by the multiple brands tested, samples from the eye shadow
applicators were collected in one of two ways. The method was
determined by observing which part of the sponge came into the
most contact with the eyelid. Some of the eye shadow applica-
tors had the sponge portion completely removed from their
sticks, while others were sampled by cutting off just the tip of
the sponge portion. The upper half of all but one of the mascara
wand brushes tested was cut off with clean scissors where it
became more flexible. The exception was mascara wand 6-1,
which had a thinner brush stalk that was cut off of the wand in
its entirety. The face makeup sponge and nonsponge pad appli-
 ADAMOWICZ ET AL.
cators were cut into three vertical sections, and then, one of
those sections was cut laterally and half of that section was cut
into smaller pieces and used for testing. The vertical section
used for extraction was chosen by assessing which areas seemed
to have the most makeup product on them and were therefore
likely to have had more contact with the user’s skin. Samples
from face makeup brushes were collected by first examining the
bristles for the area that seemed to have the most makeup
adhered to the surface and then cutting sections of those bristles
for extraction. The tips of the bristles running across the top of
the brushes were cut off to a length of c. 1–1.5 cm. All con-
cealer, lip gloss, and lipstick wands had the sponge portion com-
pletely cut off of the sticks and used for DNA extraction.
Lipsticks and lip balms were processed by swabbing the top sur-
face of the used area of the product with a sterile cotton swab
moistened with deionized water. All of the lipsticks, whether
applied with a wand or from the product directly, were in neutral
shades. The collected swabs were all allowed to air-dry in a lam-
inar flow hood before extraction.
DNA Extractions
DNA from the cosmetic applicators was extracted using the
QIAampâ DNA Investigator Kit (Qiagen, Valencia, CA, USA)
and the protocol for the “Isolation of total DNA from surface and
buccal swabs” (27). All samples followed the volume specifica-
tions for a cotton or Dacron swab, and carrier RNA was not used
for this application. Many of the sample lysates did not pass
through the columns entirely during the first centrifugation step in
the 1-min spin time specified by the protocol. These samples were
further centrifuged until all of the lysate was observed to have
passed through the column. The samples were eluted in a final
volume of 100 lL of Buffer ATE. To maximize DNA yield, the
samples were incubated for 5 min at room temperature as opposed
to the 1 min specified by the manufacturer’s protocol (internal
validation, data not shown). DNA was extracted from reference
buccal swabs using the QIAampâ DNA Mini Kit following the
“Buccal swab spin” protocol and followed the volume specifica-
tions for a cotton or Dacron swab (28). The reference samples
were eluted in a final volume of 100 lL of Buffer AE.
Quantitation and Dilution
DNA extractions were quantified using an Applied Biosystems
(Foster City, CA, USA) 7500 real-time PCR system and the AB
QuantifilerTM Human DNA Quantification Kit according to the
manufacturer’s recommendations. Based on the quantitation
results, any sample that yielded a concentration of DNA that
was more than 1 ng/lL was diluted with sterile deionized water
down to c. 1 ng/lL prior to amplification.
Amplification—AmpFlSTRâ Identifilerâ PCR Amplification Kit
All samples were amplified using the Applied Biosystems
(Foster City, CA, USA) AmpFlSTRâ Identifilerâ kit following
the manufacturer’s protocols (29). The target mass of DNA for
amplification was 1 ng; however, many samples yielded concen-
trations of DNA that fell well below that value. In those cases,
the maximum volume of sample extract (10 lL) was used. The
amplification reaction final volume was 25 lL, and thermal
cycling was performed in an Applied Biosystems (Foster City,
CA, USA) GeneAmpâ PCR System 9700 using 9600 emulation
mode for 28 cycles.
. COSMETIC APPLICATORS AS A SOURCE OF DNA 3
Capillary Electrophoresis—3730 Genetic Analyzer
Sample fragment separation and detection was performed on an
Applied Biosystems (Foster City, CA, USA) 3730 Prismâ Genetic
Analyzer. Samples were prepared for injection in a mixture of
Applied Biosystems (Foster City, CA, USA) HiDiTM formamide
(8.7 lL/sample) and GeneScanTM 500 LIZTM size standard
(0.3 lL/sample). A total volume of 10 lL (1 lL sample or allelic
ladder and 9 lL formamide and size standard) was added to the
appropriate wells for injection. All injections were performed at
3 kV for 5 sec. Injection results were analyzed using Applied Bio-
systems (Foster City, CA, USA) GeneMapperâ ID v3.2 software.
Control Samples
A set of control cosmetic applicators were tested before begin-
ning any analysis on the used applicator samples collected from
participants. The control applicators were purchased from a beauty
supply store and had never previously been in contact with any
cosmetic product. The unused controls included face brush bris-
tles, a mascara wand, a sponge eye shadow applicator, and a face
makeup applicator sponge. These were intended to test for PCR
inhibition from the applicators themselves in the absence of any
cosmetic product and to gauge the relative levels of expected
DNA recovery after extraction. Each of the clean applicators was
spiked with 10 lL of fresh liquid human blood, and the blood was
allowed to dry. Another set of unused cosmetic applicators includ-
ing face brush bristles, a mascara wand, a sponge eye shadow
applicator, a face makeup applicator sponge, a lip balm, a lipstick,
and an eyeliner pencil were also processed after the addition of
previously unused makeup products to the applicator and 10 lL
of fresh liquid human blood. All of these samples were processed
in the same manner as the collected experimental samples from
the research participants. A fresh human blood baseline was cre-
ated for comparison by processing three duplicate samples of
10 lL of liquid blood from the same donor as that added to the
control samples. All of the control samples were extracted in the
same way as the experimental items, with the exception of the
liquid blood. It was extracted using the QIAampâ DNA Mini Kit
following the “Blood and body fluid spin” protocol and eluted in
a final volume of 100 lL of Buffer AE (30).
Additional Data
All participants filled out a questionnaire concerning the usage
and storage conditions associated with each donated item. Data
were collected regarding how long the applicator had been in
use, whether anyone other than the owner was known to have
used the item, whether the owner was the last person to have
used the applicator, and how the applicator was stored while in
the owner’s possession. After data collection, the history of the
applicators was reviewed to see whether time of use or time
since use had an effect on DNA recovery and whether access to
the applicator by other people correlated with the presence of
mixtures in the respective STR profiles that were developed.
Results
Control Samples
Three duplicate liquid human whole blood aliquots of 10 lL
each were extracted to establish a baseline for the recovery of
DNA from cosmetic applicators. The average quantity for the
4 JOURNAL OF FORENSIC SCIENCES
DNA recovered from the blood was 1.90  0.14 ng/lL.
Identical aliquots were then placed on clean, unused cosmetic
applicators and processed. The clean control applicators all gave
quantitation results, and all four yielded full STR profiles
(Table 1). Samples 1-A (mascara wand), 1-B (face brush), and
1-C (eye shadow sponge) had quantitation values similar to that
calculated for the whole blood control and developed STR pro-
files fully concordant with the control blood donor, indicating
that recovery of DNA from these applicators was not problem-
atic. Sample 1-D (face makeup sponge) yielded a quantitation
value lower than the whole blood control, but still high enough
to generate a full STR profile. The control applicators with
makeup product added and spiked with 10 lL of whole blood
also all gave detectable quantitation results (Table 2). Most val-
ues were lower than the clean applicators’ results; however, the
mascara wand, face brush, eye shadow sponge, face sponge, eye-
liner pencil, and lipstick all gave full STR profiles concordant
with the control blood donor. Only the lip balm sample failed to
develop complete results at all loci. It showed a concordant par-
tial profile with one allele dropping out in locus D2S1338.
Collected Makeup Applicators
All of the samples were processed through STR separation
and detection, even those with quantitation values below the
limit of detection. Electropherograms from each sample were
analyzed to see whether full, partial, or no DNA profiles were
obtained for the different cosmetic applicators (data not shown).
These profiles were also assessed for degradation, allele drop-
out, and potential mixtures. Any sample that generated a full
or partial profile was compared to the known reference samples
of the participant donors to determine whether the profiles were
concordant. The number of loci listed in items demonstrating
partial profiles corresponds to loci that match the reference
completely.
The quantitation and profile results for the eyeliner sponge
smudgers are shown in Table 3. All of the samples tested gave
low quantitation values and resulted in partial STR profiles in
which two were concordant with the contributors’ reference
swabs. Smudger 5-7 was nonconcordant with its respective refer-
ence, and questionnaire data indicated that persons other than
the owner had access to the eyeliner smudger and could have
been in contact with it while in the owner’s possession.
The quantitation and profile results for the eyeliner pencils
and crayons are shown in Table 4. Most of the samples gave
low-level quantitation results, but three of the samples generated
partial profiles concordant with the known contributors while
three did not yield any DNA profiles. Sample 9-2 had the high-
est yield of DNA and generated a full STR profile concordant
with the contributor reference. DNA was collected from this item
differently than the other eyeliner pencils and crayons as previ-
ously described. The higher yield of DNA for sample 9-2 may
TABLE 1––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for clean (no makeup product) control applicators spiked
with 10 lL of human blood.
Sample Code Description
1-A Mascara Wand
1-B Face Brush Cuttings
1-C Eye Shadow Sponge
1-D Face Sponge Cutting
be attributable to the variation in swabbing technique or may
have resulted from its specific history.
The eye shadow sponge, mascara wand, and concealer wand
applicators all proved to be poor sources of useful DNA
(Tables 5–7). The samples gave low or undetected quantitation
values, and none demonstrated full STR profiles. The DNA
extracted from these types of applicators was also commonly
degraded and of poor quality, as demonstrated by poor heterozy-
gous peak balance, allelic dropout in small loci, and complete
dropout in large loci (data not shown). The eye shadow sponge
applicators were all reported as having been in use for a least
1 year prior to collection. The mascara wands ranged from hav-
ing been used for c. 3 months to 1 year. The concealer wands
were all donated after 3 months or less time of use.
The face makeup sponge applicators proved to be better
sources of DNA (Table 8). While two samples (1-5 and 1-7)
had no quantifiable amounts of DNA and gave no STR profiles,
four of the other five samples yielded more than enough DNA
to generate full STR profiles with the remaining sample (5-5)
still showing a partial profile of 13 loci concordant with the
donor profile. The face makeup sponge applicators proved to
have the highest success rate of any cosmetic applicator type.
Degraded DNA was also observed in the profiles developed
from the face makeup sponges, and data collected from the ques-
tionnaires indicated that most of the applicators had been in use
for at least 2 months prior to collection.
The results of testing with face makeup brushes also indicated
that they were not always sources of high-quantity or high-qual-
ity DNA; however, they did sometimes yield full STR profiles
(Table 9). Only one sample (3-2) contained a sufficient quantity
of DNA to generate a full profile concordant with its reference,
and that profile showed degradation. Additionally, one sample
(8-3) was a mixture that was nonconcordant with its respective
reference. Data from the questionnaire for sample 8-3 indicated
that the donor did not believe that anyone else had been in con-
tact with or had used the product other than themselves. The
shortest time in use for the face makeup brushes was reported as
4 months (sample 11-1), while the longest was 3 years (sample
3-2). Among the nine face brushes collected, three of the respon-
dents said that someone else had access to their brushes,
although none had certain knowledge that the item had actually
been used by another person.
Nonsponge face makeup pad applicators gave mixed results
(Table 10). Two of the samples tested (8-6 and 10-8) did not
produce any STR results at all. Sample 10-9 developed a partial,
degraded STR profile concordant with its respective reference,
and the other two samples (1-8 and 20-2) showed full, concor-
dant DNA profile results. Three of the five samples collected (8-
6, 10-8, and 10-9) were reported as being in use for c. 1 year
with the fourth (1-8) and the fifth (20-2) having been used for c.
2 and 6 months, respectively, prior to being donated for this
study.
Quantity of
DNA (ng/lL) Profile Results (# Loci)
1.310 Full (16); Concordant with reference
1.050 Full (16); Concordant with reference
1.174 Full (16); Concordant with reference
0.367 Full (16); Concordant with reference
 ADAMOWICZ ET AL. . COSMETIC APPLICATORS AS A SOURCE OF DNA 5

TABLE 2––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for control applicators with unused makeup product added
and spiked with 10 lL of human blood.

Sample Code Description Quantity of DNA (ng/lL) Profile Results (# Loci)

2-A Mascara Wand 0.096 Full (16); Concordant with reference
2-B Face Brush Cuttings 0.536 Full (16); Concordant with reference
2-C Eye Shadow Sponge 0.176 Full (16); Concordant with reference
2-D Face Sponge Cutting 0.024 Full (16); Concordant with reference
2-E Lip Balm 0.008 Partial (15); Concordant with reference
2-F Eyeliner Pencil 0.171 Full (16); Concordant with reference
2-G Lipstick 0.593 Full (16); Concordant with reference

TABLE 3––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for eyeliner smudgers. The “time in use” and “time from
last use” values are those reported in the questionnaires filled out by the donors of each item.

Sample Code Description Quantity of DNA (ng/lL) Profile Results (# Loci) Time in Use Time from Last Use
â
1-3 Maybelline Eyeliner Smudger 0.007 Partial (4); Concordant with reference 5 months 2 months
5-7 Arbonneâ Eyeliner Smudger 0.009 Partial (5); Non-concordant with reference 2 weeks Unknown
9-3 Clinique Eyeliner Smudger 0.039 Partial (8); Concordant with reference 6 months 5 months

TABLE 4––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for eyeliner pencils and crayons. The “time in use” and
“time from last use” values are those reported in the questionnaires filled out by the donors of each item.

Time from
Sample Code Description Quantity of DNA (ng/lL) Profile Results (# Loci) Time in Use Last Use
1-1 MAC Pencil Eyeliner 0.002 No results 3 months 2 months
1-2 Maybellineâ Pencil Eyeliner 0.001 No results 4 months 2 months
1-6 MAC Pencil Eyeliner 0.003 Partial (6); Concordant with reference 2 months 2 months
5-3 Mary Kayâ Crayon Eyeliner 0.004 Partial (4); Concordant with reference 2 weeks Unknown
5-4 Eyeliner Pencil* 0.017 Partial† (10): Concordant with reference 2 weeks Unknown
5-6 Arbonneâ Crayon Eyeliner 0.001 No results 2 weeks Unknown
9-2 Clinique Crayon Eyeliner 0.993 Full (16); Concordant with reference 6 months 5 months
13-4 Eyeliner Pencil* 0.023 Full (16); Concordant with reference 4 years 1 day
*Product brand could not be identified.
†
Samples demonstrated degraded profiles.

TABLE 5––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for eye shadow sponges. The “time in use” and “time from
last use” values are those reported in the questionnaires filled out by the donors of each item.

Time from
Sample Code Description Quantity of DNA (ng/lL) Profile Results Time in Use Last Use
2-1 Lanc^ome Eye Shadow Sponge 0.001 No results 5 years 8 months
8-4 Clinique Eye Shadow Sponge 0.127 Partial* (10); Mixed 1 year 1 week
8-5 Almayâ Eye Shadow Sponge 0.016 No results 1 year 1 week
10-4 Eye Shadow Sponge† 0.001 No results 1 year Unknown
10-5 Eye Shadow Sponge† 0.001 No results 1 year Unknown
10-6 Eye Shadow Sponge† 0.002 No results 1 year Unknown
10-7 Eye Shadow Sponge† Undetected No results 1 year Unknown
10-10 World Class International Undetected No results 1 year Unknown
Eye Shadow Sponge
13-8 Eye Shadow Sponge† 0.017 Partial* (7); Mixed 1 year 1 day
*Samples demonstrated degraded profiles.
†
Product brand could not be identified.

The quantitation and profile results for lipsticks and lip balms
were similar (Tables 11 and 12). All samples yielded low
quantities of DNA, and only one sample from each type of
applicator (20-1 and 11-2) produced a full STR profile concor-
dant with its respective reference. Partial profiles, often
degraded, were commonly seen with these samples. The longest
reported time in use for any of these products was c. 1 year
(samples 8-2, 8-7, 8-8, and 20-1), while the shortest usage per-
iod was described as 3 weeks (sample 7-3).
Lip gloss and lipstick wands collected in this study were
shown to be better sources of DNA than the lipsticks and lip
balms (Table 13). Although two samples gave no profile
results, the other five applicators resulted in partial or full
profiles, some degraded, but all concordant with their respec-
tive references. The usage times of these applicators prior to
collection had the widest range, with one of the samples
(13-6) having been used for only 3 days prior to collection.
The rest of the wands were collected after having been in use
6 JOURNAL OF FORENSIC SCIENCES

TABLE 6––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for mascara wands. The “time in use” and “time from last
use” values are those reported in the questionnaires filled out by the donors of each item.

Time from
Sample Code Description Quantity of DNA (ng/lL) Profile Results (# Loci) Time in Use Last Use
1-4 Maybellineâ Mascara Wand 0.002 No results 3 months 2 months
2-2 Estee LauderTM Mascara Wand Undetected No results 7 months 2 months
2-3 L’Orealâ Mascara Wand Undetected No results 9 months 4 months
2-4 L’Orealâ Mascara Wand 0.001 No results 9 months 4 months
3-1 MAC Mascara Wand 0.004 No results 6 months Unknown
6-1 Maybellineâ Define-A-Lash Undetected No results 6 months 6 months
Mascara Wand
7-1 L’Orealâ Hydrofuge Mascara Wand 0.006 No results 3 months 1 day
8-1 CoverGirlâ Lashblast 0.005 Partial (X Only) 1 year 1 week
8-9 Clinique Mascara Wand Undetected No results 1 year 1 week
10-1 Maybellineâ Define-A-Lash 0.013 Partial* (5); Concordant 1 year Unknown
Mascara Wand with reference
*Samples demonstrated degraded profiles.

TABLE 7––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for concealer wands. The “time in use” and “time from
last use” values are those reported in the questionnaires filled out by the donors of each item.

Profile Results
Sample Code Description Quantity of DNA (ng/lL) (# Loci) Time in Use Time from Last Use
5-2 L.A. Colorsâ Concealer 0.002 No results 2 weeks 2 weeks
Wand
5-8 Concealer Wand* Undetected No results 2 weeks 2 weeks
5-9 Concealer Wand* Undetected No results 2 weeks 2 weeks
6-2 Concealer Wand* 0.001 No results 3 months 3 months
6-3 Physicians Formula Undetected Partial† (9); Nonconcordant with reference 3 months 3 months
Concealer Wand
*Product brand could not be identified.
†
Samples demonstrated degraded profiles.

TABLE 8––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for face makeup sponges. The “time in use” and “time
from last use” values are those reported in the questionnaires filled out by the donors of each item.

Time from
Sample Code Description Quantity of DNA (ng/lL) Profile Results (# Loci) Time in Use Last Use
1-5 Face Makeup Sponge* Undetected No results 3 months 2 months
1-7 Face Makeup Sponge* Undetected No results 3 months 2 months
1-9 Face Makeup Sponge* 0.106 Full (16); Concordant with reference 4 months 2 months
4-1 L’Orealâ Face Makeup 0.453 Full† (16); Concordant with reference 2 months 6 days
Sponge
5-5 Face Makeup Sponge* 0.057 Partial† (13); Concordant with reference Unknown 2 weeks
5-10 Face Makeup Sponge* 0.952 Full† (16); Concordant with reference Unknown 2 weeks
7-4 Wet N Wildâ Face Makeup 0.659 Full (16); Concordant with reference 6 months 1 day
Sponge
*Product brand could not be identified.
†
Samples demonstrated degraded profiles.

for 1 week (sample 13-3) up to 1 year (samples 8-10, 8-11,
and 16-3).
To assess the potential presence of PCR inhibitors and the
overall quality of the recovered DNA, the heterozygous peak
balance (HPB) was calculated at each locus for all full profiles
obtained for each type of applicator, as well as the control sam-
ples. At each heterozygous locus, sister alleles were considered
balanced if the peak heights were ≥70% of each other. This
HPB value was determined based on the previous validation data
for the individual instrument and amplification kits used in this
laboratory for this study (data not shown). The unused control
applicators without makeup products added showed heterozy-
gous peak balance at all loci for the mascara wand, face sponge,
and eye shadow sponge. The face brush sample was not bal-
anced at every locus. The unused control applicators with
makeup products added showed heterozygous peak balance at all
loci for the face sponge and eyeliner pencil only, whereas all of
the other applicators demonstrated imbalance in at least one
locus. The results of the HPB analysis for the used cosmetic
applicators are shown in Fig. 1. Only the lip gloss and lipstick
wands had heterozygous peak balance at all of the loci in every
full DNA profile generated. The eyeliner, face makeup sponges,
and face makeup nonsponge categories all had sample applica-
tors that yielded full, concordant profiles, but not every one of
those profiles showed complete balance among the heterozygous
alleles. The face makeup brushes, lipstick, and lip balm samples
each included one donated applicator that yielded a full STR
profile, but none of those profiles had HPBs that consistently
 ADAMOWICZ ET AL. . COSMETIC APPLICATORS AS A SOURCE OF DNA 7

TABLE 9––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for face makeup brushes. The “time in use” and “time
from last use” values are those reported in the questionnaires filled out by the donors of each item.

Time from
Sample Code Description Quantity of DNA (ng/lL) Profile Results (# Loci) Time in Use Last Use
3-2 Face Makeup Brush* 0.168 Full† (16); Concordant with reference 3 years Unknown
5-1 Face Makeup Brush* 0.006 Partial† (5); Concordant with reference Unknown 2 weeks
8-3 Clinique Face Makeup Brush 0.030 Partial† (2); Mixed Nonconcordant with reference 1.5 years 1 week
10-2 Face Makeup Brush* 0.016 Partial† (4); Concordant with reference 1 year 1 year
10-3 Face Makeup Brush* Undetected No results 1 year 1 year
11-1 L’Orealâ Face Makeup Brush 0.019 Partial (14); Concordant with reference 4 months 1 day
11-3 Face Makeup Brush* 0.001 No results 1 year 1 day
13-7 Face Makeup Brush* 0.005 No results 1 year 1 day
14-1 Face Makeup Brush* 0.001 No results 1 year 1 month
*Product brand could not be identified.
†
Samples demonstrated degraded profiles.

TABLE 10––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for face makeup nonsponge pads. The “time in use” and
“time from last use” values are those reported in the questionnaires filled out by the donors of each item.

Profile Results Time from

Sample Code Description Quantity of DNA (ng/lL) (# Loci) Time in Use Last Use
1-8 Revlonâ Color Stay Pressed 0.031 Full (16); Concordant 3 months 2 months
Powder Pad with reference
8-6 L’Orealâ Loose Powder Pad Undetected No results 1 year 1 week
10-8 CoverGirlâ Pressed Powder Pad 0.023 No results 1 year 1 year
10-9 CoverGirlâ Pressed Powder Pad 0.028 Partial* (10); Concordant with reference 1 year 1 year
20-2 Pressed Powder Pad† 0.674 Full* (16); Concordant with reference 6 months 2.5 months
*Samples demonstrated degraded profiles.
†
Product brand could not be identified.

TABLE 11––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for lipsticks. The “time in use” and “time from last use”
values are those reported in the questionnaires filled out by the donors of each item.

Quantity of Time from

Sample Code Description DNA (ng/lL) Profile Results (# Loci) Time in Use Last Use
7-2 Rimmel London Lipstick 0.001 No results 6 months 1 week
8-7 Clinique Lipstick 0.001 No results 1 year 1 week
8-8 Clinique Lipstick 0.004 Partial (X Only) 1 year 1 week
13-2 L’Orealâ Lipstick 0.019 Partial (15); Concordant with reference 1 month 1 day
17-2 Stilaâ Lip glaze 0.002 Partial (3); Concordant with reference Unknown Unknown
20-1 Avonâ Lipstick 0.009 Full* (16); Concordant with reference 1 year Unknown
*Samples demonstrated degraded profiles.

met the criteria. The eyeliner smudgers, eye shadow sponges,
concealer wands, and mascara wands all failed to yield a single
full STR profile, and heterozygous allele peak balance across the
profiles was poor at many loci.
Discussion
This study was designed to determine which types of cosmetic
applicators, if any, were better candidates for retrieving high-
quantity and high-quality human DNA for STR testing. There
are a wide variety of applicator types and an enormous variety
of brands and colors of products associated with the applicators.
We attempted to sample a selection of applicator types that was
large enough to establish trends in the efficacy of using one type
over another. At least one sample from each category of cos-
metic applicator yielded, at the minimum, a partial STR profile,
indicating that DNA could be extracted and typed from all of
the different kinds of applicators. There were, however, some
applicator categories that yielded better results than others, and
at least three of the types demonstrated results that indicate they
would be poor choices for forensic DNA testing.
Data collected from the study’s participants about the history
of each sample item indicated that 100% of the respondents
believed that they were the last person to have used the cosmetic
prior to donation. They also reported that c. 31% of the sample
items could potentially have come into contact with another per-
son and that c. 20% of the other applicators were known to have
been used by a person other than the donor. While just over one
half of the applicators were either known to have been used by
another person or another person had access to them, only two
items in these categories actually demonstrated a mixed or non-
concordant STR profile. Sample 5-7, an eyeliner smudger, was
reported as having no known user other than the donor, but it
was noted that other people had access to this item. This item
developed a partial profile that was not that of the donor, nor
did the profile match any of the other experimental samples,
contributor references, or researchers. Sample 13-8, an eye sha-
dow sponge, was known to have had a user other than the donor
8 JOURNAL OF FORENSIC SCIENCES

TABLE 12––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for lip balms. The “time in use” and “time from last
use” values are those reported in the questionnaires filled out by the donors of each item.

Time from
Sample Code Description Quantity of DNA (ng/lL) Profile Results (# Loci) Time in Use Last Use
7-3 Chapstickâ Cherry Lip Balm 0.006 Partial (X Only) 3 weeks 1 day
8-2 Lumeneâ Lip Balm 0.003 No results 1 year 1 week
11-2 Chapstickâ Lip Balm 0.114 Full* (16); Concordant with reference 5 months 1 day
12-1 Chapstickâ Cherry Lip Balm 0.012 Partial* (2); Concordant with reference Unknown Unknown
12-2 Chapstickâ Lip Balm 0.003 Partial* (3); Concordant with reference Unknown Unknown
12-3 Lip Balm† 0.003 Partial (2); Nonconcordant with reference Unknown Unknown
13-1 Chapstickâ Lip Balm 0.003 No results 3 months Unknown
*Samples demonstrated degraded profiles.
†
Product brand could not be identified.

TABLE 13––Cosmetic applicator description, DNA yield, and AmpFlSTRâ Identifilerâ profile results for lip gloss and lipstick wands. The “time in use” and
“time from last use” values are those reported in the questionnaires filled out by the donors of each item.

Quantity of Time from

Sample Code Description DNA (ng/lL) Profile Results (# Loci) Time in Use Last Use
8-10 NARSâ Lip Gloss Wand Undetected No results 1 year 1 week
8-11 NARSâ Lip Gloss Wand 0.259 Full* (16); Concordant with reference 1 year 1 week
9-1 Revlonâ Liquid Lipstick Wand 0.424 Full* (16); Concordant with reference 6 months 2 months
13-3 Maybellineâ Wet Shine Diamonds 0.038 Full (16); Concordant with reference 1 week 1 day
Liquid Lipstick Wand
13-6 New York Colorâ Lip Wand 0.003 No results 3 days 1 day
15-4 Mary Kayâ Lip Gloss Wand 0.025 Partial (11); Concordant with reference Unknown Unknown
16-3 Clinique Lip Gloss Wand 0.003 Partial* (7); Concordant with reference 1 year 1 day
*Samples demonstrated degraded profiles.

FIG. 1––Percentage of full DNA profiles for each used applicator type that
had all heterozygous sister alleles >70%. Partial profile results were not
assessed.
and yielded a partial, mixed profile consistent with the reference
and another female. All of the other items that resulted in mixed
or nonconcordant profiles were believed to have been used by,
and accessible to, the donor only. These results support the con-
clusion that the DNA profile developed is usually that of the last
person to contact a cosmetic item.
A set of control cosmetic applicators were processed to see
whether the applicator type alone or applicator with added
makeup affected the amount of DNA recovered from the sample
or the ability to obtain a full DNA profile with heterozygous
peak balance. The clean, unused applicators all yielded usable
quantities of DNA and gave full profiles with some peak height
imbalance showing in the face brush sample. The imbalance was
not severe and would not preclude the profile from being used
for interpretation. Once makeup was added to the unused appli-
cators and they were spiked with blood, the recovered quantities
decreased. However, all but the lip balm still gave full STR pro-
files. The lip balm swab had allele dropout at one locus
(D2S1338). After re-amplifying this sample, allele dropout was
still present and was most likely the result of low quantities of
input DNA to the PCR causing stochastic amplification. In addi-
tion to diminished DNA yield, the presence of makeup products
impacted the quality of the extracted DNA. Heterozygous peak
imbalance was seen in five of the seven applicator types tested.
The STR results for the mascara wand and eye shadow sponge
were both quantitatively and qualitatively poorer when makeup
was present. Based on previous work, these results were
expected (13,21). The addition of dyes, waxes, and other constit-
uent molecules in makeup products will affect the efficiency of
the processes included in STR profiling. Despite these impedi-
ments, six of the seven controls with makeup products still
yielded full, concordant profiles. These results demonstrated that
the applicators alone did not interfere with DNA extraction,
quantitation, or amplification procedures, and the addition of
makeup products, while deleterious to the process, would not
preclude these types of samples from being considered as items
for forensic DNA analysis.
Initial expectations were that all of the applicators that were
of a sponge or pad type would show the highest quantities and
quality of extracted DNA. This was anticipated due to the nature
of the applicators and their method of use. The amount of skin
surface area sponges and pads contact is usually larger than the
other applicators, thus resulting in a greater chance of capturing
nucleated epithelial cells and trapping them within the surface of
the applicator for easy retrieval from the matrix of the sponge or
pad. These expectations were partially substantiated. The eye
 ADAMOWICZ ET AL.
shadow applicators, while being relatively wide sponges that are
drawn across the eyelids, showed poor DNA typing results.
Extracted quantities of DNA were very low for all of the sam-
ples, and only degraded, mixed profiles resulted. As the eye sha-
dow makeup products often contained highly vibrant and/or dark
pigments, there was a concern that these samples contained
higher levels of PCR-inhibiting dyes. Even though they had a
sponge tip c. 7–9 mm in length and 1–2 mm wide, the concealer
wands were not expected to be particularly rich sources of DNA
as they are usually only dabbed onto an area briefly and there-
fore would not pick up a lot of skin cells due to minimal con-
tact. This was verified as four of the five samples tested gave no
STR results. Only one of the concealer wands developed an
STR result, and it was a degraded partial profile that was not
concordant with the reference. The eyeliner smudgers were simi-
lar to the concealer wands as they were also small sponge appli-
cators. The smudgers were differently shaped, however, being c.
1–3 mm long by 2–3 mm wide. There were concerns associated
with the smudgers due to their size and the dyes and waxy com-
ponents of the eyeliners and the eye shadows that the smudgers
come into contact with on the skin surface. All of these could
possibly interfere with DNA extraction or amplification. The
eyeliner smudgers yielded slightly better results than expected.
All three samples gave very low levels of extracted DNA; how-
ever, two of the three samples gave partial STR profiles concor-
dant with the donors. The lipstick and gloss wands had sponge
tips, often with hairlike projections designed to transfer the cos-
metic product but also potentially useful for collecting and trap-
ping epithelial cells. Testing with these applicators gave mixed
results. DNA yields ranged from undetected to 0.424 ng/lL and
STR results were equally diverse, with two samples showing no
amplification product and three samples yielding full, concordant
profiles. The quality of the full profiles derived from the lipstick
and gloss wands was consistently high, with 100% of the full
profiles showing heterozygous allele peak balance of ≥70% at
all loci. While the results were highly variable, when detectable
quantities of DNA were extracted from the face makeup sponges
they yielded higher levels than any of the other applicator types
with an average quantitation value of 0.445  0.377 ng/lL.
Testing with the face makeup sponges also developed the high-
est percentage of full STR profiles for any of the applicators
tested. Four of the seven samples collected (57%) gave full, con-
cordant profiles with three of those four profiles (75%) meeting
peak balance criteria. The face makeup nonsponge pad applica-
tors resulted in two of five applicators (40%) yielding full DNA
profiles concordant with the references, but only one of the two
full profiles generated had heterozygous peak balance at all loci.
The average yield of DNA was also less than that of the similar
face sponge applicator. These results indicated that the non-
sponge pad applicator was not as good of a candidate for testing
as the sponge variety, but could still result in useful profiles.
Both the sponge and nonsponge pad face makeup applicators have
the most contact with the face and are also touched by the hands
of the user, unlike the applicators with sticks or wands. For these
reasons, they were expected to yield the highest quantities of type-
able DNA. Challenges, however, were found in processing these
sample types. Their size requires that they have to be cut into subi-
tems and that sampling can lead to missing an area on the applica-
tor that contains the bulk of the available DNA. The face sponges
and pads also retain large amounts of makeup product which can
make extraction difficult due to physically clogging membrane-/
column-type extraction systems or creating imprecise interfaces in
organic solvent methods. Another challenge encountered during
. COSMETIC APPLICATORS AS A SOURCE OF DNA 9
the extraction procedure with the face makeup sponges occurred
with samples 1-5 and 1-7. Both of these sponges broke up when
placed in the spin column and very little liquid was collected from
these samples, which later gave no DNA profile results. There
appears to be more variability in the construction of the face
sponges than the other types of applicators, as no others displayed
a tendency to disintegrate.
The brush applicators shared some similarities to the face
sponges and pads. They cover a large area of the face when
used, and they also have a large surface area to collect DNA
from. Overall, the makeup brushes did not prove to be reliable
sources of high-quality or high-quantity DNA. Only one brush
of nine (11%) tested gave a full profile, and DNA degradation
was commonly observed with this type of applicator.
The eyeliner pencils and crayons and mascaras wands were all
initially anticipated to give the least amount of DNA profile
results. This was due to the way that these applicators are used.
They come into contact with the least amount of skin and for
brief periods of time compared to the other applicators tested in
this study. Mascara wands, in particular, were expected to have
the lowest yields because they usually have minimal contact
with skin and the makeup product has a thick, waxy consistency
that may interfere with DNA extraction procedures as well as
the potential presence of PCR inhibitors from the black dyes in
the mascara. The results for the mascara wands supported the
expectations. The samples had very low quantities of DNA
recovered, and no full profiles were developed. The DNA
extraction step was often problematic with the mascara wands,
with column clogging a common occurrence. The kit manufac-
turer’s suggestions were used, and the columns were centrifuged
at full speed for longer periods of time until the liquid passed
through the column into the collection tube. It is possible that
this problem reduced the efficiency of DNA binding to the sil-
ica, resulting in undetectable or low quantitation results and the
lack of DNA profiles. The eyeliner pencils and crayons gave
better results, in general, than the mascara wands with two of
eight samples (25%) yielding full DNA profiles concordant with
the references. While most quantities of recovered DNA were
still low and most profiles were only partial, item 9-2 had a
much higher yield. It should be noted that the collection method
for 9-2 was different than the other applicators of its type as
described earlier. The crayon portion of item 9-2 was absent, so
the interior area where the product is held in place was swabbed,
rather than the crayon itself. These results show that, while not
ideal, enough DNA can be collected from the eyeliner pencils
and crayons to generate full STR profiles at least some of the
time. Also, collection should be carried out around the rim of
the product where it meets the body of the pencil, rather than
being swabbed off of the product itself.
The lipstick and lip balm applicator types gave similar results
to one another. Both kinds ranged from samples yielding no
results to examples with full, concordant profiles. The lipsticks
tested resulted in one of six samples (16%) and the lip balms
showed one of seven (14%) yielding full DNA profiles concor-
dant with the references. With a single exception (lip balm sam-
ple 11-2), the detected quantities were low and the quality of the
extracted DNA from all of the samples was also low. Evidence
of DNA degradation was widely present, and none of the full
profiles had heterozygous peak balance across all loci.
There was no clear trend in the length of use of each item and
the ability to obtain a full DNA profile across all of the types of
applicators (Tables 3–13). According to the participant question-
naires, some of the items that gave full STR profiles were in use
10 JOURNAL OF FORENSIC SCIENCES
for longer or shorter periods of time than items that had no
detectable quantities of DNA or produced no STR results when
amplified. An example of this was found in the lipstick and lip
gloss wands where sample 8-11 was reported as being in use for
c. 1 year and yielded a full STR profile, while sample 13-6 was
only in use for 3 days and yielded no detectable amplification
products (Table 13). Conversely, the nonsponge pad sample 1-8
was in use for c. 3 months and developed a full profile, while
sample 10-8 was used for c. 1 year and showed no results
(Table 10). Items that were used closest to the date of collection
tended to give better profile results than those with longer time
gaps between the last use and donation. The lip balm sample
11-2 was reported as having been used the day before it was
collected and gave a full STR profile, while sample 8-2 had a 1-
week period from last use and showed no detectable STR profile
(Table 12). Similar results were seen with other applicator types;
however, there were a few examples of items that had not been
used for longer periods of time than others, yet still generated
better profile results. Time of collection is an important factor
whenever biological evidence is tested; however, the time since
the last usage of the cosmetic applicators could also be relevant.
Degraded DNA, mixed STR profiles, and PCR inhibitors are
constant challenges to successfully typing evidentiary samples,
and cosmetic applicators demonstrated all of these issues to
some extent. Cosmetic applicators are normally stored at room
temperature when in a person’s possession and are often kept
for months or years, sometimes with lengthy periods between
uses. Nuclear DNA present on these items is therefore often
present in a degraded state unless there has been heavy or very
recent usage. Many of the items processed in this study did
result in degraded DNA profiles. As cosmetic applicators are
sometimes also shared by more than one person, mixtures were
also a concern. Three of the samples processed showed evi-
dence of low-level mixtures in the stochastic range. In addition
to sample 13-8, sample 8-3 was a face makeup brush and sam-
ple 8-4 was an eye shadow sponge applicator that both devel-
oped degraded, mixed, partial profiles. As with sample 8-3, the
questionnaire associated with sample 8-4 stated that no one
other than the donor had used it before. While mixtures were
observed in STR profiles developed from cosmetic applicators,
they only occurred in three of the 76 samples tested (c. 3.9%).
These data indicate that applicator samples will often yield a
single-source DNA profile (c. 96% in this study). In some sam-
ples, the resulting extracts were colored rather than clear, indi-
cating that dyes were not completely removed from the purified
DNA. One sample, in particular face makeup brush 10-3, had a
liquid extract tinted blue. This sample had a quantitation value
of “undetected,” and its Q-PCR internal positive control (IPC)
was also undetected, indicating PCR inhibition. Although this
result was expected to occur in more of the applicators pro-
cessed because of the dye and wax components within the
products, 10-3 was the only item to show no value for its Q-
PCR IPC. While the AB AmpFlSTRâ Identifilerâ kit was used for
this research, the latest generation STR kits, such as the Applied
Biosystems (Foster City, CA, USA) AmpFlSTRâ Identifilerâ Plus
and GlobalFilerTM or Promega (Madison, WI, USA) PowerPlexâ
16 HS and Fusion kits, have improved performance in the
presence of PCR inhibitors. The use of these kits is likely to
further mitigate, or eliminate entirely, the already low impact level
of PCR inhibitors on the rate of generating DNA profiles from
cosmetic applicators. The quantity of recovered DNA and the level
of degradation are more significant than PCR inhibition in
developing STR profiles from cosmetic applicators.
Conclusion
Personal effects such as cosmetic applicators can be helpful
tools in generating reference DNA profiles for missing persons or
victims of mass disasters as nontraditional known references and
can also be submitted to forensic DNA laboratories as evidentiary
samples. The challenges of mixed samples and PCR inhibitors are
still present in cosmetic applicators, but data indicate that they are
not so severe as to suggest that these items should be eliminated
from consideration for testing. Additionally, while many of the
samples yielded low quantities of DNA, often leading to no detect-
able amplification products or partial profiles, 14 samples (18%)
gave full profiles concordant with the donors’ profiles. After ana-
lyzing the DNA profile results from 76 separate samples in 11 cat-
egories of cosmetic applicators, this study found that the face
makeup sponges were the most efficient cosmetic applicator type
for forensic DNA typing and identification purposes. They were
followed by lip gloss and lipstick wands, nonsponge face makeup
pad applicators, eyeliner pencils and crayons, lipsticks, lip balms,
and brushes, respectively. The least efficient cosmetic applicator
types for DNA analysis were the eyeliner smudgers, eye shadow
sponges, mascara wands, and concealer wands, none of which
yielded a full STR profile.
Acknowledgments
The authors would like to thank all of the donors who partici-
pated in this project for their generosity in giving us the cos-
metic products that were tested. We would also like to thank
Todd Bille for reviewing early versions of this manuscript and
his helpful suggestions, as well as the anonymous reviewers
whose helpful and relevant comments on the submission signifi-
cantly improved the paper.
References
1. Schyma C, Madea B, Courts C. Persistence of biological traces in
gun barrels after fatal contact shots. Forensic Sci Int Genet
2013;7:22–7.
2. Richert NJ. Swabbing firearms for handler’s DNA. J Forensic Sci
2011;56:972–5.
3. Bright J-A, Petricevic SF. Recovery of trace DNA and its application to
DNA profiling of shoe insoles. Forensic Sci Int 2004;145:7–12.
4. Linacre A, Pekarek V, Swaran YC, Tobe SS. Generation of DNA pro-
files from fabrics without DNA extraction. Forensic Sci Int Genet
2010;4:137–41.
5. Abaz J, Walsh SJ, Curran JM, Moss DS, Cullen J, Bright J-A, et al.
Comparison of the variables affecting the recovery of DNA from com-
mon drinking containers. Forensic Sci Int 2002;126:233–40.
6. Goray M, Eken E, Mitchell RJ, van Oorschot RAH. Secondary DNA
transfer of biological substances under varying test conditions. Forensic
Sci Int Genet 2010;4:62–7.
7. Leclair B, Fregeau CJ, Bowen KL, Fourney RM. Enhanced kinship
analysis and STR-based DNA typing for human identification in mass
fatality incidents: the Swissair flight 111 disaster. J Forensic Sci
2004;49:939–53.
8. National Institute of Justice. Mass fatality incidents: a guide for
human forensic identification. Washington, DC, June 2005; http://
www.ojp.usdoj.gov/nij/pubs-sum/199758.htm (accessed July 2013-Octo-
ber 2013).
9. Alonso A, Martin P, Albarran C, Garcia P, de Simon LF, Iturralde MJ,
et al. Challenges of DNA profiling in mass disaster investigations. Croat
Med J 2005;46:540–8.
10. Prinz M, Carracedo A, Mayr WR, Morling N, Parsons TJ, Sajantila A,
et al. DNA Commission of the International Society for Forensic
Genetics (ISFG): recommendations regarding the role of forensic genet-
ics for disaster victim identification (DVI). Forensic Sci Int Genet
2007;1:3–12.
 ADAMOWICZ ET AL.
11. Budowle B, Bieber FR, Eisenberg AJ. Forensic aspects of mass disasters:
strategic considerations for DNA-based human identification. Leg Med
2005;7:230–43.
12. Tanaka M, Yoshimoto T, Nozawa H, Ohtaki H, Kato Y, Sato K, et al.
Usefulness of a toothbrush as a source of evidential DNA for typing.
J Forensic Sci 2000;45:674–6.
13. Webb LG, Egan SE, Turbett GR. Recovery of DNA for forensic analysis
from lip cosmetics. J Forensic Sci 2001;46:1474–9.
14. van Oorshot RAH, Jones MK. DNA fingerprints from fingerprints. Nat-
ure 1997;387:767.
15. Ladd C, Adamowicz MS, Bourke MT, Scherczinger CA, Lee HC. A sys-
tematic analysis of secondary DNA transfer. J Forensic Sci
1999;44:1270–2.
16. Phipps M, Petricevic S. The tendency of individuals to transfer DNA to
handled items. Forensic Sci Int 2007;168:162–8.
17. Wickenheiser RA. Trace DNA: a review, discussion of theory, and appli-
cation of the transfer of trace quantities of DNA through skin contact.
J Forensic Sci 2002;47:442–50.
18. Quinones I, Daniel B. Cell free DNA as a component of forensic evi-
dence recovered from touched surfaces. Forensic Sci Int Genet
2012;6:26–30.
19. Akane A, Matsubara K, Nakamura H, Takahashi S, Kimura K. Identifi-
cation of the heme compound copurified with deoxyribonucleic acid
(DNA) from bloodstains, a major inhibitor of polymerase chain reaction
(PCR) amplification. J Forensic Sci 1994;34:362–72.
20. Opel KL, Chung D, McCord BR. A study of PCR inhibition mechanisms
using real time PCR. J Forensic Sci 2010;55:25–33.
21. Larkin A, Harbison SA. An improved method for STR analysis of blood-
stained denim. Int J Legal Med 1999;112:388–90.
. COSMETIC APPLICATORS AS A SOURCE OF DNA 11
22. Koukoulas I, O’Toole FE, Stringer P, van Oorschot RAH. QuantifilerTM obser-
vations of relevance to forensic casework. J Forensic Sci 2008;53:135–41.
23. Bourke MT, Scherczinger CA, Ladd C, Lee HC. NaOH treatment to
neutralize inhibitors of Taq polymerase. J Forensic Sci 1999;44:1046–50.
24. Promega Corporation. DNA IQTM DNA isolation system small sample
casework technical bulletin. Madison, WI: Promega Corporation, 2006.
25. Brevnov MG, Pawar HS, Mundt J, Calandro LM, Furtado MR, Shewale
JG. Developmental validation of the PrepFilerTM forensic DNA extraction
kit for the extraction of genomic DNA from biological samples. J Foren-
sic Sci 2009;54:599–607.
26. Cupples CM, Champagne JR, Lewis KE, Cruz TD. STR profiles from
DNA samples with “undetected” or low QuantifilerTM results. J Forensic
Sci 2009;54:103–7.
27. QIAGEN. QIAampâ DNA investigator handbook. Valencia, CA: Qiagen,
2010;14–7.
28. QIAGEN. QIAampâ DNA mini and blood mini handbook, 3rd edn.
Valencia, CA: Qiagen, 2010;37–8.
29. Applied Biosystems. AmpFlSTRâ Identifilerâ PCR amplification kit
user’s manual. Foster City, CA: Applied Biosystems, 2006.
30. QIAGEN. QIAampâ DNA mini and blood mini handbook, 3rd edn.
Valencia, CA: Qiagen, 2010;27–9.
Additional information and reprint requests:
Michael Adamowicz, Ph.D.
Department of Forensic Science
University of New Haven
300 Boston Post Rd.
West Haven, CT 06516
E-mail: madamowicz@newhaven.edu