1. technology and computing

Forensic Science International: Genetics 5 (2011) 484–492
Contents lists available at ScienceDirect
Forensic Science International: Genetics
journal homepage: www.elsevier.com/locate/fsig
Integrated sample cleanup and capillary array electrophoresis microchip
for forensic short tandem repeat analysis
Peng Liu a, James R. Scherer b, Susan A. Greenspoon c, Thomas N. Chiesl b, Richard A. Mathies a,b,*
a
UCSF/UC Berkeley Joint Graduate Group in Bioengineering, University of California, Berkeley, CA 94720, USA
Department of Chemistry, University of California, Berkeley, CA 94720, USA
c
Virginia Department of Forensic Science, Richmond, VA 23219, USA
b
A R T I C L E I N F O
A B S T R A C T
Article history:
Received 22 July 2010
Received in revised form 7 October 2010
Accepted 7 October 2010
A twelve-lane capillary array electrophoresis (CAE) microsystem is developed that utilizes an efficient
inline capture injection process together with the classical radial microfabricated capillary array
electrophoresis (mCAE) format for high-sensitivity forensic short tandem repeat (STR) analysis. Biotinlabeled 9-plex STR amplicons are captured in a photopolymerized gel plug via the strong binding of
streptavidin and biotin, followed by efficient washing and thermal release for CE separation. The analysis
of 12 STR samples is completed in 30 min without any manual process intervention. A comparison
between capture inline injection and conventional cross injection demonstrated at least 10-fold
improvement in sensitivity. The limit-of-detection of the capture-CAE system was determined to be 35
haploid copies (17–18 diploid copies) of input DNA; this detection limit approaches the theoretical limits
calculated using Poisson statistics and the spectral sensitivity of the instrument. To evaluate the
capability of this microsystem for low-copy-number (LCN) analysis, three touch evidence samples
recovered from unfired bullet cartridges in a pistol submerged in water for an hour were successfully
analyzed, providing 53, 71, and 59% of the DNA profile. The high-throughput capture-CAE microsystem
presented here provides a more robust and more sensitive platform for conventional as well as LCN and
degraded DNA analysis.
ß 2010 Elsevier Ireland Ltd. All rights reserved.
Keywords:
Forensic human identification
PCR cleanup
Lab-on-a-chip
Microfabrication
Capillary electrophoresis
Genetic analysis
1. Introduction
Driven by escalating backlogs and the challenges of analyzing
evidence with a wide range in DNA quantity and quality [1–4],
forensic scientists are always seeking techniques that can improve
short tandem repeat (STR) analysis for better throughput, cost,
sensitivity, and reliability. While process automation has advanced
significantly through the application of robotics [5,6] and capillary
electrophoresis (CE) [7], the translation of the process into a
nanoliter scale in an integrated microfluidic format is more
desirable due to its potential for reducing reagent and time
consumption, enhancing the sensitivity and reliability, and
eliminating the risk of sample contamination and mix-up [8–
11]. Towards this goal, multi-lane microfabricated capillary array
electrophoresis (mCAE) system has been developed for high-
* Corresponding author at: Department of Chemistry, 307 Lewis Hall, MS 1460,
University of California, Berkeley, CA 94720, USA. Tel.: +1 510 642 4192;
fax: +1 510 642 3599.
E-mail addresses: ramathies@berkeley.edu, rich@zinc.cchem.berkeley.edu
(R.A. Mathies).
1872-4973/$ – see front matter ß 2010 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.fsigen.2010.10.009
throughput STR typing [12,13]. The integration of on-chip CE with
PCR was successfully demonstrated for real-time forensic human
identification [14,15]. Sample processing steps, such as DNA
extraction, were also conducted on-chip [16,17]. These achievements validate the feasibility of conducting high-performance
forensic STR analysis on integrated microfabricated devices, but
further process optimization of integration for routine forensic
investigations is needed.
The most valuable advantage provided by microfabrication
technology is the ability to integrate additional functions into the
STR analysis process, which are critical to enhance the performance, but not economical in conventional formats. For example,
current STR analysis bypasses a post-PCR cleanup step for routine
sample processing in order to save time, cost and excessive sample
handling. However, direct electrokinetic injection from high-salt
PCR products introduces an injection bias against large DNA
fragments and only a small fraction of the PCR products (<1%) are
injected for analysis, resulting in reduced sensitivity [10]. Although
post-PCR purification and concentration prior to CE analysis
improves the sensitivity, the extra time and cost consumption, and
increasing risks of sample contamination have limited its adoption
in forensic investigations [18–20]. Microfabrication technology
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
485
allows us to explore the integration of PCR sample cleanup with
electrophoretic analysis on a single device to overcome the
problems of direct electrokinetic injection without introducing
additional issues.
A variety of on-chip purification, concentration and separation
methods have been developed, including membrane filtration
[21,22], sample stacking [23,24], and sample extraction [25].
Although these approaches have demonstrated improvements in
sensitivity, the delicate operations are incompatible with highthroughput on-chip integration. An oligonucleotide-based gel
capture method was developed in our group for DNA sequencing
[26] and genotyping [27], but would be difficult to optimize for a
high-order multiplex capture process. To address this issue, we
have developed an integrated STR sample cleanup and separation
microdevice and method that employs a streptavidin capture gel
chemistry coupled to a simple direct-injection geometry [28]. STR
products having one dye-labeled strand and the other labeled with
biotin are efficiently captured and concentrated in a photopolymerized streptavidin gel plug, followed by washing and thermal
release for separation. An entire analysis can be complete in about
40 min. Compared to conventional microchip CE with a crossinjector, the fluorescence intensity was improved 14–19-fold for 9plex STR products. However, to make this device and method
practically useful and cost-effective for routine forensic work, the
scaling of this technology and method to high-density array
structures is essential.
In this study, we developed a 12-lane capture-capillary array
electrophoresis (CAE) microdevice that combines the efficient
streptavidin-capture-gel chemistry [28] presented earlier with an
improved automated process and the classical design of highthroughput mCAE systems [29]. To evaluate the capture method for
high-throughput forensic DNA profiling, we performed 9-plex STR
analyses from standard genomic DNA. The fluorescence signals
obtained using the capture-CAE device were compared with those
using conventional cross injection under the same condition. The
limit of detection of this system was also determined for forensic
applications. Finally, we tested the ability of the capture-CAE
microsystem to process and improve the analysis of touch
evidence collected from unfired bullet cartridges that were
removed from a pistol submerged in water. This study is a
significant step towards the practical application of this integrated
capture-separation process in forensic investigation.
2. Materials and methods
2.1. Microchip design
The design of the 12-lane capture-CAE microdevice is presented
in Fig. 1. On a 4 in. glass wafer, twelve 10-cm-long separation
channels sharing a common anode are grouped into six doublets,
similar to our previous mCAE devices [12]. Each doublet includes
two capture gel inline injectors with two sample wells and one
shared cathode and one waste well. The capture gel inline injector
is a 500-mm-long double-T channel junction with a tapered
structure for PCR product cleanup, concentration, and inline
injection. The tapered structure is designed to keep the capture gel
in place during gel loading. All features were isotropically etched to
a depth of 40 mm and a final width of 160 mm using the same wet
etching method as described previously [12]. Prior to use, the
microchannels are coated with 0.25% polyDuramide (pDuramide)
dynamic coating polymer to minimize DNA absorption to the
channel walls and electroosmotic flow during electrophoresis [30].
The coating procedure consists of 1 M HCl incubation for 15 min, DI
water flush, and pDuramide incubation for an hour. After
treatment, the chips are flushed with water, and then dried with
vacuum.
Fig. 1. Schematic of the 12-lane capture-CAE microdevice. (A) A total of 12
electrophoretic separation channels coupled with capture gel inline injectors are
arranged on a 4 in. glass wafer. (B) Each doublet includes two capture gel inline
injectors and two sample wells sharing one cathode and one waste well. (C)
Expanded view of the gel capture region. A constriction was fabricated at the top of
the capture region to keep the capture gel in place during the gel loading process.
(D) Expanded view of the hyper-turn structure in the separation channels.
2.2. Compact scanner system
A photograph of the instrument used to perform analyses is
shown in Fig. 2A. The system has dimensions 12 in. 12 in. 8 in.,
and it can be used as either a bench-top or portable instrument.
The instrument contains a 488-nm diode laser (100 mW, Sapphire
488, Coherent, Santa Clara, CA), a confocal optical system with a
rotary objective for detecting four different fluorescence signals,
pneumatics for control of on-chip microvalves [14,31], four PCR
temperature control systems [14,15], and four high voltage power
supplies for electrophoresis. A LabVIEW graphical interface
(National Instruments, Austin, TX) developed in-house is used to
486
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
filter and collected at a rate of 5000 data points per revolution of
the objective using the DAQ board.
During operation, a polydimethlysiloxane (PDMS) elastomer
well is secured on top of the cathode and waste wells on the chip to
create continuous buffer reservoirs. The microdevice is then placed
onto the 6 in. heated stage on the top of the instrument and held in
place with a Plexiglas manifold and vacuum supplied by the
instrument. The manifold contains Pt electrodes that are positioned within the reservoirs on the microchip for the application of
high voltages during electrophoresis. Further design details and
schematics can be found at http://www.cchem.berkeley.edu/
ramgrp/scanner.
2.3. Streptavidin capture gel preparation
A 500 mL streptavidin gel solution, containing 5% (v/v)
acrylamide and bis-acrylamide (19:1, Bio-Rad, Hercules, CA),
8 M Urea, 1 TTE (500 mM Tris, 500 mM TAPS acid and 100 mM
EDTA), 2 mg/mL streptavidin-acrylamide (Invitrogen, Carlsbad,
CA), 0.0006% riboflavin (w/v) and 0.125% TEMED (v/v), is prepared
in an opaque 2-mL scintillation vial with Teflon closure (National
Scientific, Rockwood, TN) following the method developed
previously [28]. To form a capture gel, the photopolymerization
solution was first loaded into channels by vacuum. A viscous 5%
LPA solution was loaded into each reservoir to stop hydrodynamic
flow. Using a UV exposure setup installed on a Nikon inverted
microscope (TE2000U, Nikon) and photomask [28], a 500-mm
capture gel plug is formed in the double-T channel junctions. The
polymerization of each gel plug is complete in 5 min, 12 plugs are
finished in 1 h. Un-reacted solution was evacuated out of the
channels and replaced with 1 TTE buffer.
2.4. Short tandem repeat typing
Fig. 2. The second-generation CAE scanner system. (A) Photograph of the
12 in. 12 in. 10 in. analysis system. (B) Schematic assembly of the four-color
confocal detection system with a rotary objective. Design details and schematics
will be found at http://www.cchem.berkeley.edu/ramgrp/scanner.
control the system through a NI 6259 OEM multifunction DAQ
board (National Instruments, Austin, TX). This system has the
flexibility of accommodating one 150-mm diameter wafer in
different throughput (12, 24, 48 or 96 channel) configurations
[12,32]. With the incorporation of a temperature-controlled stage,
pneumatics, and PCR heater control, this system is capable of
performing mCAE analysis alone [12], cleanup and separation of
off-chip amplified samples [28], and even fully integrated on-chip
PCR amplification and STR analysis [14,15].
The optical system is shown in Fig. 2B. The laser beam is
reflected by two dielectric mirrors up through a dichroic mirror
(Chroma, Rockingham, VT, Z488bpxr). It then enters the hollow
shaft stepper motor, is displaced 7 mm by the rhomb, and is
focused into the center of the microplate channel. The low mass of
the rhomb objective assembly (48 g) allowed us to use a small
hollow shaft rotary motor (Empire Magnetics, Inc. U17-7). The
clear aperture of the objective was 2.8 mm, which was small
enough to pass through the hollow shaft of the stepper motor.
Fluorescence collected by the objective returns through the
stepper motor shaft and is reflected by the dichroic mirror into
a confocal assembly. The light is focused on a 200 mm pinhole with
a 20 mm fl achromat lens, and collimated into a 0.7 mm diameter
beam with a 5 mm fl achromat lens that enters the 4-color PMT
(Hamamatsu H9797 fitted with special sequential dichroic beam
splitter optics, 537dclp, 570dc, 595dclp, Z488bpxr, Chroma). The
converted electrical signals are processed using a 5-Hz low-pass
For proof of concept, a 9-plex autosomal STR typing system
developed previously based on the primer sequences and fluorescence dye labeling scheme used in the PowerPlex1 16 System
(Promega, Madison, WI) was employed to test the capture-CAE
system [28]. To enable the post-PCR cleanup and inline injection, the
unlabeled primers were replaced with biotin-labeled primers (IDT,
Coralville, IA). The STR loci included in the 9-plex system are
amelogenin for sex typing and 8 CODIS core STR loci (D3S1358,
TH01, D21S11, D5S818, D13S317, D7S820, vWA and D8S1179).
To facilitate allele calls, a biotin-modified sizing standard was
constructed in-house by mixing a series of purified PCR amplicons
with different fragment lengths (60, 80, 95, 120, 140, 160, 172, 250,
275, and 350 bp). These fragments are amplified from pUC19,
producing products having one strand labeled with ROX and the
other labeled with biotin. This sizing standard can be co-captured
and separated with PCR products for allele size calibration. Due to
the limited amount of the custom-built sizing standards, the crossinjection experiments employed a commercially available sizing
standard (ILS 600, Promega).
Genomic standard DNA 9947A and 9948 were purchased from
Promega (Madison, WI) and diluted in deionized water (DI water)
according to the requirements of the studies. The PCR mixture
prepared for 9-plex STR typing is comprised of 1 Gold ST*R buffer
(50 mM KCl, 10 mM Tris–HCl (pH 8.3), 1.5 mM MgCl2, 0.1% Triton
X-100, 160 mg/mL BSA, 200 mM each dNTP) (Promega), the primer
mixture, 0.16 U/mL AmpliTaq Gold DNA polymerase (Applied
Biosystems, Foster City, CA), and DI water. All PCR samples were
amplified using a PTC-200 thermocycler (MJ Research, Waltham,
MA) according to the manufacturer’s protocol for the PowerPlex1
16 kit.
In the capture-CAE studies, PCR products are first mixed with
Hi-Di formamide (Applied Biosystems) in a ratio of 9–1, and then
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
loaded into the sample wells for analysis. When a sizing standard is
included, the mixing ratio of PCR products, sizing standard, and
formamide is 8:1:1. In the cross-injection experiments, the sample
preparation recipe is kept the same, but an extra heating step at
95 8C for 3 min is performed to denature the PCR products prior to
loading.
2.5. Touch evidence preparation
The touch evidence samples recovered from unfired bullet
cartridges were prepared by the Virginia Department of Forensic
Science (VDSF) following the procedure described previously [33].
Briefly, bullet cartridges were picked up out of box and loaded into
a pistol by a volunteer with no additional or excessive handling of
the bullet cartridges (‘‘real conditions’’). This pistol was then
submerged in water for an hour. After collection from the
submerged weapon, each cartridge was swabbed using the double
swab technique [34]. Cartridges were each swabbed with a sterile
damp swab containing approximately 40 mL of Type I (ultrapure)
water, followed by a second sterile dry swab. Swabs were allowed
to air dry before storing at room temperature. DNA samples were
extracted following the VDSF BioMek1 2000 Automation Workstation Procedures Manual for Large volume samples. DNA
samples with 2.5% Sarkosyl (Sigma, St. Louis, MO), 1 TNE (Tris
NaCl EDTA) buffer, 0.39 mmol dithiothreitol (DTT, Sigma), 1.25 mg
Proteinase K (Sigma), and Type I water in a 500 mL volume were
digested overnight at 56 8C. The cuttings were then placed in spin
baskets to collect all of the associated liquid. DNA IQTM lysis buffer
with DTT (1.0 mL of DNA IQTM lysis buffer, DTT at 75 mM final
concentration) and 8 mL DNA IQTM resin were added to each
sample in a 2 mL tube. The samples were vortexed for 30 s,
incubated at room temperature for 5 min, then vortexed again for
487
30 s. Samples were then purified utilizing the BioMek1 2000
Automation Workstation (Beckman Coulter, Pasadena, CA), following the VDFS BioMek1 2000 Automation Workstation Procedures
Manual (http://www.dfs.virginia.gov/manuals/forensicBiology/
index.cfm). DNA quantitation was performed using the Plexor HY
System (Promega) on a Stratagene Mx3005PTM Real-Time PCR
System (Cedar Creek, TX) according to the manufacturer’s specifications with a recalibration of standard DNA concentrations. The
samples were shipped to Berkeley for analysis on the capture-CAE
device. PCR amplifications were performed with 4 mL input DNA in
12.5 mL reaction volume using the protocol described above.
2.6. Capture-CAE device operation
The capture and separation process illustrated in Fig. 3 is similar
to our previous work [28], but has been significantly modified to
minimize manual intervention during analysis. Following the
photopolymerization of the capture gel plugs in the chip, a
separation matrix (5% LPA with 8 M Urea in 1 TTE) is loaded from
the anode to the waste and from the cathode to the sample
reservoirs to form a matrix-capture-matrix gel sandwich structure
in the capture inline injection regions. The tapered structure in the
capture region ensures that the capture gel plug will be retained.
Off-chip amplified PCR sample is loaded into sample wells and
injected through the capture bed under an electric field of 25 V/cm
at room temperature for 10 min. PCR products are bound
efficiently to the capture gel via the biotin–streptavidin interaction
to form a tightly concentrated plug. After capture, an electric field
of 25 V/cm is applied from the cathode to the waste reservoirs to
wash all the uncaptured sample contents through the gel to the
waste for 5 min. The channels above the capture gel are then
cleaned by applying the same electric field from the cathode to the
Fig. 3. Fluorescence images of the capture-CAE operation process. (A) A 500-mm capture gel plug containing streptavidin is formed in the capture region by
photopolymerization. (B) Off-chip amplified PCR product is loaded into the sample well and injected through the capture bed using an electric field of 25 V/cm at room
temperature for 10 min. (C) The captured DNA products are then washed under the same electric field for 5 min. (D) An electric field of 25 V/cm is then applied from the
cathode to the sample reservoir to clear unbound sample above the gel. (E) Next, a backwash step is carried out to wash any unbound materials in the separation channel
below the gel towards the waste. (F) The microdevice is equilibrated at 67 8C for 1 min to thermally release the fluorescently labeled DNA strands, which are electrophoresed
at 250 V/cm towards the anode for separation analysis.
488
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
sample. To clean the separation channels, a backwash step is
carried out to wash any unbound materials from the anode
towards the waste. Finally, the microdevice is equilibrated at 67 8C
for 1 min to thermally release the fluorescently labeled DNA
strands, which are electrophoresed at 250 V/cm towards the
anode. The overall process can be completed in 30 min. After each
run, all the gels and solutions in the chip are pushed out using
water and the channels are cleaned using 1 M NaOH and piranha
(7:3 H2SO4:H2O2) for 10 min to prevent run-to-run carryover
contamination.
To plot the STR traces, the four-color fluorescence data
underwent baseline and color crosstalk corrections using BaseFinder 6.1.16, and plotted using Microcal Origin. To determine the
allele repeat numbers, the four-color fluorescence data are first
converted to binary format and appended with proper header
information by a custom LabVIEW program. The preprocessed data
files are then analyzed for allele calling using the MegaBACETM
Fragment Profiler 1.2 (GE Healthcare, Piscataway, NJ) program
which performs baseline and color cross-talk correction.
2.7. Cross injection operation
The cross-injection analyses are also performed on the same
capture-CAE microdevice with no capture plugs and 5% LPA
separation gels in all channels at 67 8C. The PCR samples loaded in
the sample wells are electrophoretically injected towards the
waste reservoir by applying an electric field of 100 V/cm for 60 s
while floating the anode and cathode in order to create an injection
plug. A separation field of 250 V/cm is then applied between the
cathode and the anode to effect the separation.
3. Results and discussion
3.1. Automated high-throughput analysis
In a high-throughput capillary array electrophoresis system,
human intervention in the analytical process should be completely
eliminated. While an integrated capture and capillary electrophoresis system utilizing the strong binding of streptavidin and biotin
has been previously developed by our group, the complicated
manual operation of this system, which includes three buffer
exchanges, is the limiting factor for scaling up to high-throughput
system. Several modifications were made here to achieve the
automation of the process in the 12-lane capture-CAE system: (i)
Separation matrix (5% LPA with 8 M urea) is loaded into both sides
of the capture gel and PCR samples are only pipetted into the
sample reservoirs. In the previous system, samples filled the entire
injection channel on the top of the capture gel plug. As a result,
with the previous design, the subsequent replacement of the
sample with a washing buffer is required. In the new design, a
simple push-back step by applying voltage from the cathode to the
sample reservoir is enough to eliminate excess sample in the
channels. (ii) Urea is mixed into the capture gel to lower the
denaturing temperature of the double-strand PCR amplicons. In
the previous system, a formamide washing step with two buffer
exchanges was performed to facilitate the complete sample release
for separation at 67 8C. By incorporating 8 M urea into the capture
gel plug as well as the separation matrix, this washing step is
successfully eliminated. (iii) Another function of the formamide
washing in the single-lane system was to improve the separation
resolution by stacking the sample plug during sample release for
separation. We found a similar resolution can be successfully
obtained by simply shortening the capture gel plug from 1 mm to
500 mm. Through the modifications described above, the on-chip
STR cleanup, capture and separation system has been successfully
scaled up to a 12-lane system and the entire process does not
require any manual intervention. The total time for the captureCAE process is less than 30 min, including 20-min post-PCR
purification and <10-min separation. Compared to the off-chip
post-PCR purification methods, such as Qiagen MinElute, which
has a 30-min procedure with multiple manual operations and
centrifugations [18], our on-chip capture-CAE system is rapid,
easy to operate, and more reliable due to the elimination of
human intervention which may cause contamination or sample
mix-up.
To test the chip design and the operational protocol for highthroughput forensic STR typing, we analyzed 9-plex STR samples
amplified from 50 copies of 9947A standard genomic DNA on the
12-lane capture-CAE microdevice. The same amplified sample was
loaded into all 12 sample reservoirs and analyzed simultaneously.
As shown in Fig. 4, STR profiles were successfully obtained from all
the 12 lanes in 30 min. The average percentage standard deviation
of the allele signal-to-noise (S/N) ratios is 18.8% and the average
standard deviation of the allele migration time is 9.8 s. The
variations are similar to those found in conventional chip-based
mCAE separations, and can be effectively corrected by incorporating sizing standards [29]. These results demonstrate that the
current chip design and operational protocol are compatible with
high-throughput integrated STR sample cleanup and analysis. The
average separation resolution of the THO1 allele 8 and 9.3 (7 bp
difference) was also calculated to be 3.2 0.1. It is possible to detect
the rare 1-bp variants (such as THO1 9.3/10) using the current chip
design. Furthermore, the resolution can be improved by simply
extending the length of the separation channels to 15 cm.
3.2. Comparison with cross-injection method
We also conducted a comparison study between conventional
cross injection and the capture inline injection under the same
conditions to evaluate the sensitivity improvement provided by
the capture-CAE microsystem. Since the design of the 12-lane
capture-CAE device is similar to the mCAE system, both the capture
inline injection and the conventional cross injection can be
performed on the same microchip, which allows us to directly
compare these two methods. In this study, the STR samples
amplified from 50 copies of 9948 standard genomic DNA were
analyzed using both methods. As shown in Fig. 5, the allele S/N
ratios in the 9-plex STR profile obtained using the capture inline
injection were improved 12.1 1.8-fold over the cross injection.
Similar peak balance across the DNA profile was maintained for both
analysis methods. The excess primers observed in the cross-injection
separation profiles were completely removed by the purification
process, demonstrating the effectiveness of sample cleanup, concentration and inline injection.
3.3. Limit of detection
A sensitivity study was carried out to evaluate the limit of
detection (LOD) of the capture-CAE microsystem by using 9-plex
STR samples amplified from serially diluted 9947A standard
genomic DNA. Fig. 6A plots the percent allele detection as a
function of haploid copy numbers of the DNA template and the S/N
ratios of each allele at the 9 loci for the 35-copy samples are shown
in Fig. 6B. All 13 expected STR alleles were successfully and
reproducibly detected with as few as 35 haploid copies of DNA for
each locus (105 pg template DNA). The average S/N ratio is 30.0
and the lowest is 5.0 from X allele in amelogenin. This sensitivity is
slightly lower than the previous single-lane system (25 copies) due
to the change of the sample loading method. Since the samples are
only loaded into the sample reservoirs instead of the entire
injection channels, the injection into the capture region is less
efficient. Nevertheless, the limit of detection is still improved
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
489
Fig. 5. Comparison of STR profiles obtained using the capture inline injection and
the conventional cross injection methods under the same conditions. STR samples
are amplified from 0.15 ng of 9948 standard DNA using the 9-plex STR typing
system. (A) The top trace was obtained by cross injection using the capture-CAE
device. The bottom trace was generated on the same device using the capture inline
injection procedure. (B) Graph of the S/N ratio improvement on each allele. The
average improvement by using the capture inline injection over the cross injection
is 12.1 1.8-fold.
Fig. 4. (A) STR profiles of 0.15-ng standard DNA obtained from the 12-lane captureCAE microdevice. The sample cleanup, concentration, and CE analysis were finished
within 30 min. All the traces from the 12 lanes are plotted on the same signal
intensity scale. The average percentage standard deviation of the allele signal-tonoise (S/N) ratios is 18.8% and the average standard deviation of the allele migration
time is 9.8 s. (B) Expanded view of one of the 9-plex STR traces. The dramatically
reduced primer peaks show the effectiveness of the sample cleanup procedure
which eliminates the their interference with smaller STR fragment detection.
the copy numbers of each chromosome present in the PCR reaction
conform to the Poisson distribution, resulting in a variation of the
PCR amplification of each allele [36]. To determine the contribution
of template loading variation to the limit-of-detection results, the
fraction of the DNA profile as a function of input DNA copy number
was calculated theoretically based on the expectation from the
Poisson distribution.
The probability that an allele can be detected by the system at a
given input DNA copy number can be expressed using the Poisson
cumulative probability function shown in Eq. (1):
Pðx; lÞ ¼ 1 x1 l i
X
e l
i¼0
significantly compared to commercial CE instruments and crossinjection based microchip platforms (50 copies) [12].
When doing a limit-of-detection study in the low-copy-number
(LCN) region (100 pg or 33 copies of each locus) [1,35], one
challenge is the occurrence of stochastic effects, which include
heterozygote imbalance, stutter, and drop-out of one or both
alleles during PCR amplification. One of the primary factors causing
these stochastic effects is template loading variation: when a low
concentration of genomic DNA is introduced into a PCR reaction,
i!
(1)
where l is the input copy number of each locus (0.5l for
heterozygous alleles), e is the base of the natural logarithm, and x is
the system detection limit expressed in copy number [37]. The
average fraction of the full DNA profile is determined by the
average peak numbers obtained and divided by the total allele
number, as shown in Eq. (2):
P
P¼
Pi
100%
nt
(2)
490
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
Fig. 6. The limit-of-detection test of the capture-CAE microsystem. (A) Fraction
of STR profiles obtained using the capture-CAE device for 9947A standard DNA
as a function of input DNA copy number. Full STR profiles were reliably obtained
from as low as 35 copies of DNA template. The solid line shows the theoretical
fraction of 9-plex profiles fits to the experimental data. The dashed lines are the
individual Poisson distributions for detection limits of 1, 3, 8, and 15 haploid
copies. (B) Average allele S/N ratios for the 35-copy DNA sample on the captureCAE device.
P
where Pi is the sum of the probabilities of detecting all the alleles
and nt is the total number of alleles analyzed. The sensitivity of the
instrument is 20 pM for FAM, 50 pM for JOE, and 100 pM for
TMR (data not shown), which translates to approximately 2 pM
FAM, 5 pM JOE, and 10 pM TMR before capture gel concentration.
Since one single template generates about 108 DNA fragments in an
on-chip monoplex PCR [38], it is reasonable to assume that 107
DNA fragments will be generated for each allele in a 9-plex STR
amplification which may have lower PCR efficiency due to reagent
competition. Therefore, in a 25 mL reaction volume, the detection
limits of the capture-CAE system are estimated to be 3 copies of
DNA template for FAM-labeled, 8 copies for JOE-labeled, and 15
copies for TMR-labeled alleles. As shown in Fig. 6A, the
experimental detection profile vs. DNA concentration corresponds
well with the various theoretical detection profiles, demonstrating
that the statistical loss of template does play a significant role in
determining the fall-off from 100% profile detection at the lowest
concentrations. The individual detection profiles for detection
limits of 1, 3, 8, and 15 copies were also plotted vs. input copy
number for comparison.
Fig. 7. Analyses of three touch evidence samples recovered from unfired bullet shells
which were submerged in water for an hour. In a 12.5-mL PCR reaction volume, 4–5
copies of DNA template were amplified using the 9-plex STR typing system. Top traces
in each sample were generated using the cross-injection method on the capture-CAE
device – no peaks were obtained. In contrast, the lower traces show that 52.9%, 70.6%,
and 58.8% of the alleles can be successfully detected using the sample cleanup,
concentration, and inline injection procedure on the same chip. Biotin-labeled sizing
standards, shown in red (bottom traces), were co-captured and electrophoresed to
facilitate the calculation of allele repeat numbers. (For interpretation of the references
to color in this figure legend, the reader is referred to the web version of the article.)
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
491
Table 1
Alleles obtained and called using the capture-CAE system.
Sample
D3
Expected alleles
Shell case 1
Shell case 2
Shell case 3
16,
*
*
TH01
17
*
*
6,
*
*
D21
9.3
*
29,
*
*
D5
30
*
12,
*
13
*
*
D13
D7
Amel
13
*
*
*
10,
11
*
*
X,
*
*
*
vWA
Y
*
*
17,
*
*
*
D8
18
*
*
*
13,
*
*
Percent profile (%)
14
53
71
59
(*) Allele above calling threshold.
*Possible drop-in allele with a repeat number of 32.
We have demonstrated that LCN analysis using the capture-CAE
system approaches the theoretical limits determined by Poisson
statistics and our spectral sensitivity. The plots in Fig. 6A show that
an improvement in our sensitivity by a factor of 3 as well as the use
of energy-transfer dye labeling (which eliminates the sensitivity
variation between different dyes) [39] would enable nearly full
profile detection at 10 input haploid copies, but beyond this limit
purely statistical effects that are independent of detection
sensitivity preclude the possibility of obtaining a full STR profile.
These statistical effects are best circumvented by moving towards
digital or discrete single cell analysis methods that guarantee full
genomic content [40].
3.4. Touch evidence analysis
A DNA profile may be obtained by swabbing items that have
been handled by a suspect, even in the absence of visible evidence
[33,41,42]. The ability to obtain this touch DNA evidence has
evolved over recent years, and has become a routine practice in
forensic investigations in numerous countries around the world.
Although touch evidence does not always result in low copy
number or low template DNA, the limited quantity of DNA is one of
the greatest challenges faced by forensic scientists. Needless to say,
an instrument with enhanced sensitivity, but also with integrated
functionality that reduces the threat of contamination or sample
mix-up, will benefit the DNA analysis of touch evidence or LCN
samples.
To test the capability of the capture-CAE system for low-copynumber DNA analysis, we analyzed touch evidence provided by the
Virginia Department of Forensic Science. In total, three samples,
each of which was extracted from a single unfired bullet cartridge
that was touched by the same person and retrieved from a pistol
submerged in water, were amplified separately and tested on the
capture-CAE system. Sample concentrations were determined
using real-time PCR (4.1, 4.6, and 4.8 pg/mL) to guide the input
DNA quantity for the following PCR amplification. 9-plex STR
products were amplified from each sample by loading 5 copies of
DNA template in 12.5 mL reaction volumes (equal to 10 copies in
25 mL). As demonstrated in Fig. 7, 53%, 71%, and 59% of the 9-plex
STR profiles were successfully obtained by using the sample
cleanup, concentration and analysis method on the capture-CAE
device. Table 1 summarizes the allele calls of these three samples
together with the expected alleles. The average fraction of the DNA
profile detected is 61%, which is in accordance with the LOD study
described above. With the aid of biotin-labeled sizing standards,
these alleles were successfully recognized and called for their
repeat numbers. Possible drop-in alleles as well as amplification
artifacts were observed in this study due to the LCN and
compromised sample quality. As a comparison, the same amplified
samples were analyzed using the cross injection method under the
same experiment settings, revealing consistent blank profiles. The
DNA yields of the unfired cartridges from the control, unsubmerged pistol were approximately 3 times that of the
cartridges retrieved from the submerged pistol (17.9 22.6 pg
and 5.9 8.1 pg, respectively; n = 15), demonstrating the extremely
challenging nature of these samples (unpublished data, S. Green-
spoon). This study dramatically validates the advantages of the
capture-CAE system for the analyses of ‘‘touch’’ or low-copy-number/
low template DNA analysis.
4. Conclusions
We have successfully demonstrated a 12-lane capture-CAE
microsystem that employs a photopolymerized streptavidin-gel
capture process in a high-throughput format for rapid purification,
concentration and separation of biotin-modified STR amplicons.
The process of analyzing 12 STR samples required no manual
intervention and can be completed in less than 30 min. The signal
intensity has been improved at least 10-fold compared with
conventional microchip CE with a cross-injector, allowing the
detection of full 9-plex STR profiles from as few as 35 copies of
input DNA. This enhanced sensitivity enabled the analysis of touch
evidence from unfired bullet cartridges retrieved from a pistol
submerged in water. The capture-CAE microsystem hence provides
a reliable and robust platform for forensic STR typing of LCN and
degraded DNA due to its seamless integration of analytical steps,
automated operation process, and 100% efficient sample analysis.
This system has the potential to become a routine way to reliably
analyze DNA samples in forensic investigations.
In the future the development of UV exposure systems for rapid
photopolymerization at multiple spots would speed up the
formation of gel capture plugs and enable even higher throughput
capture-CAE devices. Additionally, this inline capture injection
process can be incorporated into an on-chip PCR-CE system,
leading to the realization of a highly sensitive, contamination-free
fully integrated system to explore more challenging studies in
forensics, such as STR typing from single cells.
Acknowledgements
We thank Samantha Cronier and Brian S. Cho for providing
biotin-labeled sizing standards. This project was supported by
Grant No. 2007-DN-BX-K142 awarded by the National Institute of
Justice, Office of Justice Programs, U.S. Department of Justice.
Points of view in this document are those of the authors and do not
necessarily represent the official position or policies of the U.S.
Department of Justice. RAM discloses a financial interest in
IntegenX, a company that is developing elements of the
technologies presented here.
References
[1] P. Gill, J. Whitaker, C. Flaxman, N. Brown, J. Buckleton, An investigation of the rigor
of interpretation rules for STRs derived from less than 100 pg of DNA, Forensic Sci.
Int. 112 (2000) 17–40.
[2] J.M. Butler, Y. Shen, B.R. McCord, The development of reduced size STR amplicons
as tools for analysis of degraded DNA, J. Forensic Sci. 48 (2003) 1054–1064.
[3] P. Gill, R. Sparkes, R. Pinchin, T. Clayton, J. Whitaker, J. Buckleton, Interpreting
simple STR mixtures using allele peak areas, Forensic Sci. Int. 91 (1998) 41–53.
[4] R.A. Wickenheiser, D.N.A. Trace, A review, discussion of theory, and application of
the transfer of trace quantities of DNA through skin contact, J. Forensic Sci. 47
(2002) 442–450.
[5] S.A. Greenspoon, J.D. Ban, K. Sykes, E.J. Ballard, S.S. Edler, M. Baisden, B.L.
Covington, Application of the BioMek (R) 2000 laboratory automation worksta-
492
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
P. Liu et al. / Forensic Science International: Genetics 5 (2011) 484–492
tion and the DNA IQ (TM) system to the extraction of forensic casework samples, J.
Forensic Sci. 49 (2004) 29–39.
S.A. Greenspoon, K.L.V. Sykes, J.D. Ban, A. Pollard, M. Baisden, M. Farr, N. Graham,
B.L. Collins, M.M. Green, C.C. Christenson, Automated PCR setup for forensic
casework samples using the normalization wizard and PCR setup robotic methods, Forensic Sci. Int. 164 (2006) 240–248.
J.M. Butler, E. Buel, F. Crivellente, B.R. McCord, Forensic DNA typing by capillary
electrophoresis using the ABI Prism 310 and 3100 genetic analyzers for STR
analysis, Electrophoresis 25 (2004) 1397–1412.
P. Liu, R.A. Mathies, Integrated microfluidic systems for high-performance genetic
analysis, Trends Biotechnol. 27 (2009) 572–581.
A. Manz, N. Graber, H.M. Widmer, Miniaturized total chemical-analysis systems –
a novel concept for chemical sensing, Sens. Actuators B: Chem. 1 (1990) 244–
248.
D.J. Harrison, K. Fluri, K. Seiler, Z.H. Fan, C.S. Effenhauser, A. Manz, Micromachining a miniaturized capillary electrophoresis-based chemical-analysis system on a
chip, Science 261 (1993) 895–897.
K.M. Horsman, J.M. Bienvenue, K.R. Blasier, J.P. Landers, Forensic DNA analysis on
microfluidic devices: a review, J. Forensic Sci. 52 (2007) 784–799.
S.H.I. Yeung, S.A. Greenspoon, A. McGuckian, C.A. Crouse, C.A. Emrich, J. Ban, R.A.
Mathies, Rapid and high-throughput forensic short tandem repeat typing using a
96-lane microfabricated capillary array electrophoresis microdevice, J. Forensic
Sci. 51 (2006) 740–747.
J.H. Aborn, S.A. El-Difrawy, M. Novotny, E.A. Gismondi, R. Lam, P. Matsudaira, B.K.
Mckenna, T. O’Neil, P. Streechon, D.J. Ehrlich, A 768-lane microfabricated system
for high-throughput DNA sequencing, Lab Chip 5 (2005) 669–674.
P. Liu, T.S. Seo, N. Beyor, K.J. Shin, J.R. Scherer, R.A. Mathies, Integrated portable
polymerase chain reaction-capillary electrophoresis microsystem for rapid forensic short tandem repeat typing, Anal. Chem. 79 (2007) 1881–1889.
P. Liu, S.H.I. Yeung, K.A. Crenshaw, C.A. Crouse, J.R. Scherer, R.A. Mathies, Realtime forensic DNA analysis at a crime scene using a portable microchip analyzer,
Forensic Sci. Int.: Genet. 2 (2008) 301–309.
J.M. Bienvenue, N. Duncalf, D. Marchiarullo, J.P. Ferrance, J.P. Landers, Microchipbased cell lysis and DNA extraction from sperm cells for application to forensic
analysis, J. Forensic Sci. 51 (2006) 266–273.
K.M. Horsman, S.L.R. Barker, J.P. Ferrance, K.A. Forrest, K.A. Koen, J.P. Landers,
Separation of sperm and epithelial cells in a microfabricated device: potential
application to forensic analysis of sexual assault evidence, Anal. Chem. 77 (2005)
742–749.
P.J. Smith, J. Ballantyne, Simplified low-copy-number DNA analysis by post-PCR
purification, J. Forensic Sci. 52 (2007) 820–829.
T. Lederer, G. Braunschweiger, P. Betz, S. Seidl, Purification of STR-multiplexamplified microsamples can enhance signal intensity in capillary electrophoresis,
Int. J. Legal Med. 116 (2002) 165–169.
K.A. Dugan, H.S. Lawrence, D.R. Hares, C.L. Fisher, B. Budowle, An improved
method for post-PCR purification for mtDNA sequence analysis, J. Forensic Sci.
47 (2002) 811–818.
A.V. Hatch, A.E. Herr, D.J. Throckmorton, J.S. Brennan, A.K. Singh, Integrated
preconcentration SDS-PAGE of proteins in microchips using photopatterned
cross-linked polyacrylamide gels, Anal. Chem. 78 (2006) 4976–4984.
A.E. Herr, A.V. Hatch, D.J. Throckmorton, H.M. Tran, J.S. Brennan, W.V. Giannobile,
A.K. Singh, Microfluidic immunoassays as rapid saliva-based clinical diagnostics,
Proc. Natl. Acad. Sci. U.S.A. 104 (2007) 5268–5273.
M.J. Gong, K.R. Wehmeyer, P.A. Limbach, F. Arias, W.R. Heineman, On-line sample
preconcentration using field-amplified stacking injection in microchip capillary
electrophoresis, Anal. Chem. 78 (2006) 3730–3737.
[24] H. Yang, R.L. Chien, Sample stacking in laboratory-on-a-chip devices, J. Chromatogr. A 924 (2001) 155–163.
[25] Z.C. Long, Z. Shen, D.P. Wu, J.H. Qin, B.C. Lin, Integrated multilayer microfluidic
device with a nanoporous membrane interconnect for online coupling of solidphase extraction to microchip electrophoresis, Lab Chip 7 (2007) 1819–1824.
[26] R.G. Blazej, P. Kumaresan, S.A. Cronier, R.A. Mathies, Inline injection microdevice
for attomole-scale Sanger DNA sequencing, Anal. Chem. 79 (2007) 4499–4506.
[27] N.M. Toriello, C.N. Liu, R.A. Mathies, Multichannel reverse transcription-polymerase chain reaction microdevice for rapid gene expression and biomarker analysis,
Anal. Chem. 78 (2006) 7997–8003.
[28] S.H.I. Yeung, P. Liu, N. Del Bueno, S.A. Greenspoon, R.A. Mathies, Integrated sample
cleanup-capillary electrophoresis microchip for high-performance short tandem
repeat genetic analysis, Anal. Chem. 81 (2009) 210–217.
[29] B.M. Paegel, C.A. Emrich, G.J. Weyemayer, J.R. Scherer, R.A. Mathies, High throughput DNA sequencing with a microfabricated 96-lane capillary array electrophoresis bioprocessor, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 574–579.
[30] S.A. Greenspoon, S.H.I. Yeung, K.R. Johnson, W.K. Chu, H.N. Rhee, A.B. McGuckian,
C.A. Crouse, T.N. Chiesl, A.E. Barron, J.R. Scherer, J.D. Ban, R.A. Mathies, A forensic
laboratory tests the Berkeley microfabricated capillary array electrophoresis
device, J. Forensic Sci. 53 (2008) 828–837.
[31] W.H. Grover, A.M. Skelley, C.N. Liu, E.T. Lagally, R.A. Mathies, Monolithic membrane valves and diaphragm pumps for practical large-scale integration into glass
microfluidic devices, Sens. Actuators B: Chem. 89 (2003) 315–323.
[32] C.A. Emrich, H.J. Tian, I.L. Medintz, R.A. Mathies, Microfabricated 384-lane capillary array electrophoresis bioanalyzer for ultrahigh-throughput genetic analysis,
Anal. Chem. 74 (2002) 5076–5083.
[33] K.M. Horsman-Hall, Y. Orihuela, S.L. Karczynski, A.L. Davis, J.D. Ban, S.A. Greenspoon, Development of STR profiles from firearms and fired cartridge cases,
Forensic Sci. Int.: Genet. 3 (2009) 242–250.
[34] D. Sweet, M. Lorente, J.A. Lorente, A. Valenzuela, E. Villanueva, An improved
method to recover saliva from human skin: the double swab technique, J. Forensic
Sci. 42 (1997) 320–322.
[35] B. Budowle, A.J. Eisenberg, A. van Daal, Validity of low copy number typing and
applications to forensic science, Croat. Med. J. 50 (2009) 207–217.
[36] P. Taberlet, S. Griffin, B. Goossens, S. Questiau, V. Manceau, N. Escaravage, L.P.
Waits, J. Bouvet, Reliable genotyping of samples with very low DNA quantities
using PCR, Nucleic Acids Res. 24 (1996) 3189–3194.
[37] D. Schamhart, J. Swinnen, K.H. Kurth, A. Westerhof, R. Kusters, H. Borchers, C.
Sternberg, Numeric definition of the clinical performance of the nested reverse
transcription-PCR for detection of hematogenous epithelial cells and correction
for specific mRNA of non-target cell origin as evaluated for prostate cancer cells,
Clin. Chem. 49 (2003) 1458–1466.
[38] N.M. Toriello, E.S. Douglas, N. Thaitrong, S.C. Hsiao, M.B. Francis, C.R. Bertozzi, R.A.
Mathies, Integrated microfluidic bioprocessor for single-cell gene expression
analysis, Proc. Natl. Acad. Sci. U.S.A. 105 (2008) 20173–20178.
[39] S.H.I. Yeung, T.S. Seo, C.A. Crouse, S.A. Greenspoon, T.N. Chiesl, J.D. Ban, R.A.
Mathies, Fluorescence energy transfer-labeled primers for high-performance
forensic DNA profiling, Electrophoresis 29 (2008) 2251–2259.
[40] Z. Yong, R. Novak, J. Shuga, M.T. Smith, R.A. Mathies, High-performance single cell
genetic analysis using microfluidic emulsion generator arrays, Anal. Chem. 82
(2010) 3183–3190.
[41] R.A.H. vanOorschot, M.K. Jones, DNA fingerprints from fingerprints, Nature 387
(1997) 767–1767.
[42] M.K. Balogh, J. Burger, M. Bender, P.M. Schneider, K.W. Alt, STR genotyping and
mtDNA sequencing of latent fingerprint on paper, Forensic Sci. Int. 137 (2003)
188–195.