Transgenic Res

DOI 10.1007/s11248-015-9864-x

ORIGINAL PAPER

Next-generation sequencing is a robust strategy for the high-

throughput detection of zygosity in transgenic maize
Leonie Fritsch Rainer Fischer
Christoph Wambach Max Dudek
Stefan Schillberg Florian Schroper

Received: 30 September 2014 / Accepted: 22 January 2015

Springer International Publishing Switzerland 2015

Abstract Simple and reliable, high-throughput tech-
niques to detect the zygosity of transgenic events in
plants are valuable for biotechnology and plant
breeding companies seeking robust genotyping data
for the assessment of new lines and the monitoring of
breeding programs. We show that next-generation
sequencing (NGS) applied to short PCR products
spanning the transgene integration site provides accu-
rate zygosity data that are more robust and reliable
than those generated by PCR-based methods. The
Max Dudek: Deceased.
Electronic supplementary material The online version of
this article (doi:10.1007/s11248-015-9864-x) contains supple-
mentary material, which is available to authorized users.
L. Fritsch  R. Fischer  S. Schillberg  F. Schroper (&)
Fraunhofer Institute for Molecular Biology and Applied
Ecology IME, Forckenbeckstrasse 6, 52074 Aachen,
Germany
e-mail: florian.schroeper@ime.fraunhofer.de
NGS reads covered the 50 border of the transgenic
events (incorporating part of the transgene and the
flanking genomic DNA), or the genomic sequences
flanking the unfilled transgene integration site at the
wild-type locus. We compared the NGS method to
competitive real-time PCR with transgene-specific
and wild-type-specific primer/probe pairs, one pair
matching the 50 genomic flanking sequence and 50 part
of the transgene and the other matching the unfilled
transgene integration site. Although both NGS and
real-time PCR provided useful zygosity data, the NGS
technique was favorable because it needed fewer
optimization steps. It also provided statistically more-
reliable evidence for the presence of each allele
because each product was often covered by more than
100 reads. The NGS method is also more suitable for
the genotyping of large panels of plants because up to
80 million reads can be produced in one sequencing
run. Our novel method is therefore ideal for the rapid
and accurate genotyping of large numbers of samples.

R. Fischer Keywords Breeding  Genotyping  High

Institute for Molecular Biotechnology, RWTH Aachen throughput  Real-time PCR  Transgenic plants
University, Worringerweg 1, 52074 Aachen, Germany

C. Wambach  M. Dudek
Euregio Analytic Biochem GmbH, Oleftal 12,
53937 Schleiden, Germany Introduction

S. Schillberg
Phytopathology Department, Institute for Phytopathology
and Applied Zoology, Justus-Liebig University Giessen,
Heinrich-Buff-Ring 26-32, 35392 Giessen, Germany
The cultivation of transgenic crops increased to 433
million acres in 2013 (ISAAA 2014), and food and
animal feed products derived from these crops are

123

transported all over the world. It is therefore necessary
to test plants for the stable incorporation of transgenes
during product development and after commercial
cultivation, to ensure regulatory compliance and to
guarantee traceability. These tests are often designed
to report the presence or absence of a given transgene
sequence, but for biotechnology companies, breeders
and farmers, it is also important to determine the
zygosity of transgenic events.
Many different techniques can be used to confirm
the presence of a transgene and, more importantly, the
zygosity of the transgenic locus. Conventional tech-
niques include Southern or dot blot analysis, standard
PCR and in situ hybridization (Gachet et al. 1998;
Ahmed 2002), but these can be expensive, labor-
intensive and difficult to scale up for large numbers of
samples. Real-time PCR is an alternative approach
that allows the rapid detection of transgene sequences,
but its accuracy for the detection of zygosity is a
matter of controversy (Bubner and Baldwin 2004;
Bubner et al. 2004). Most real-time PCR techniques
also require that the DNA sample is normalized to an
endogenous reference gene (German et al. 2003;
Bubner et al. 2004) or a standard curve (Schmidt and
Parrott 2001; Gadaleta et al. 2011), both of which are
also labor-intensive procedures.
Next-generation sequencing (NGS) is based on the
generation of large numbers of short sequence reads
which can be assembled into larger contigs to
sequence entire genomes. NGS can also be used for
census-based sequencing approaches to provide accu-
rate quantitative data, e.g. RNA-seq (Yockteng et al.
2013) to measure gene expression levels and rese-
quencing to measure the frequency of specific alleles
in a population (Metzker 2010). We applied the same
census-sequencing principle to the analysis of zygos-
ity in transgenic plants and compared the accuracy and
robustness of NGS and real-time PCR.
First, we used a competitive real-time PCR with two
individual primer/probe pairs in one reaction for the
simultaneous detection of transgenic and/or wild-type
events. Event-specific real-time PCRs have been
described before although not using the strategy we
have developed (Zimmermann et al. 2000; Hernandez
et al. 2003). The advantage of this approach is, that no
endogenous reference or standard curve is required,
because the amplification curves provide an unambig-
uous yes/no result, showing a clear distinction from the
123
Transgenic Res
negative controls. Therefore, when many samples are
analyzed, calibration and optimization steps can be
omitted. Second, we applied ion-torrent sequencing to
pooled competitive endpoint PCR products, meaning
that the PCR products were mixed before sequencing.
We generated reads representing the different alleles,
providing discrete sequence information aligned with
each reference to avoid false positives. The coverage
(usually more than 100 reads per product) ensured that
the sequencing data were statistically sound and reduced
the impact of potential sequencing errors. Although
NGS is used for genotyping in medical research, e.g. for
the detection of cancer signatures (Goya et al. 2010;
Network 2013), there have been few reports of such
applications in plants and our strategies and targets were
also different (Franssen et al. 2011; Deschamps et al.
2012; Mascher et al. 2013; Teixeira et al. 2014).
We validated and compared the two approaches
using wild-type and transgenic material of known
zygosity. Our results suggest that although the real-
time PCR approach was successful, the NGS based
approach is easier to implement because it needs fewer
optimization steps than real-time PCR. Therefore, if
there is access to Next-generation sequencing tech-
nology, this approach is much more suitable for the
processing of large numbers of samples.
Materials and methods
Kernels and genomic DNA
Homozygous and heterozygous transgenic maize (Zea
mays) kernels containing the NK603 transgene (Mons-
anto, Creve Coeur, Missouri) were provided by
Euregio Analytik BioChem GmbH (Schleiden, Ger-
many), who received the kernels from KWS Saat AG
(Einbeck, Germany). DNA was extracted from the
kernels using the Nucleospin Plant II kit (Machery-
Nagel, Duren, Germany) according to the manufac-
turers protocol. Maize DNA containing the DAS-
59122-7 transgene (Dow AgroSciences, Indianapolis,
USA) was provided by Euregio Analytik BioChem
GmbH (Schleiden, Germany), who received the
extracted and purified homozygous or heterozygous
DNA from Dow AgroSciences. All DNA samples
were dissolved in elution buffer (50 mM Tris/HCl pH
8.5) to a final concentration of 3060 ng/ll.
Transgenic Res
Primers and probes: general aspects
Primers and probes were designed using CLC Main
Workbench v6.9.1 (CLC Bio, Qiagen, Venlo, Nether-
lands) and synthesized by Tib Molbiol (Berlin, Ger-
many). The sequences and the location of the
fluorescent dyes are listed in Table 1 (online resource).
The reference sequences for primer design were
obtained from the patents assigned to Bing et al.
(2011) for the DAS-59122-7 transgenic event and
Behr et al. (2002) for the NK603 transgenic event. In
the case of NK603, additional sequencing results
based on long-range PCR products covering the
complete transgene were used because the patent does
not contain sufficient sequence data for primer design
and additional upstream and downstream sequence
were needed (data not shown). The same primer pairs
were used to amplify products for detection with the
Agilent 2100 Bioanalyzer and for NGS.
Long-range PCR to generate samples for Sanger
sequencing
Long-range PCR was carried out on a Veriti 96 well
thermocycler (Life Technologies, Carlsbad, USA)
using the Long-range PCR Kit (Qiagen). Each reaction
comprised 2 U Taq DNA polymerase, 500 lM dNTPs
and 0.4 lM each primer (Table 2, online resource)
and *30 ng of template DNA made up to 50 ll with
the buffer supplied in the kit. The template was
denatured for 3 min at 93 C and then amplified (35
cycles at 93 C for 15 s, 60 C for 30 s and 68 C for
5 min), followed by indefinite storage at 8 C. Each
reaction was carried out in triplicate. The products
were purified using the Nucleospin Gel and PCR
cleanup kit (MachereyNagel) and *40 ng of each
PCR product (determined by measuring the absor-
bance at 260 nm) was sequenced from the 50 -end using
an Applied Biosystems 3730 DNA Analyzer and the
BigDye Terminator v3.1 Cycle Sequencing Kit (Life
Technologies).
Standard PCR to generate NGS templates
NGS templates were prepared by PCR on a Veriti
96-well thermocycler (Life Technologies) using the
Expand High Fidelity PCR System (Roche, Rotkreuz,
Switzerland), Axygen 8-strip tubes (Thermo Fisher
Scientific, Waltham, USA) and 8-lid flat cap strips
(Sarstedt, Numbrecht, Germany). Each reaction con-
tained 3.5 U Taq/Tgo DNA polymerase enzyme mix,
500 lM dNTPs, 0.5 lM of each primer (Table 3,
online resource) with the appropriate adapter (Table 4,
online resource) and barcode (Table 5, online
resource), and 3060 ng of template DNA made up
to 25 ll with the buffer supplied in the kit. The
template was denatured at 95 C for 2 min and then
amplified (30 cycles at 95 C for 30 s, 55 C for 30 s
and 72 C for 30 s), followed by a final elongation step
for 4 min at 72 C and indefinite storage at 8 C. PCR
products were sorted by size using the Agilent High
Sensitivity DNA Kit on an Agilent 2100 Bioanalyzer
according to the manufacturers instructions (Life
Technologies).
Real-time PCR
Real-time PCR was carried out on a 7500 Real-time
PCR cycler (Life Technologies) using the Platinum
Quantitative PCR SuperMix-UDG (uracil-DNA gly-
cosylase) with ROX reference dye (Life Technolo-
gies) according to the manufacturers instructions. The
samples were amplified in semi-skirted 96-well plates
with QPCR cap strips (Peqlab, Erlangen, Germany).
After several optimization steps (data not shown) to
determine the probe concentration (100, 200, 300 and
400 nM), the primer concentration (200, 400, 600 and
800 nM), the amount of template (7.515, 1530 and
3060 ng) and the elongation temperature (60, 58, 55
and 53 C) the final reactions contained 400 nM of the
transgene-specific primers, 600 nM of the wild-type-
specific primers, 200 nM of the transgene-specific
fluorogenic probe, 300 nM of the wild-type-specific
fluorogenic probe and 1530 ng of template DNA
made up to 50 ll final reaction volume with 19 QPCR
SuperMix. The reaction was heated to 50 C for 2 min
(UDG incubation), then to 95 C for 2 min (polymer-
ase activation) followed by 40 cycles of 95 C for 15 s
and 55 C for 30 s. Data were collected in single-
acquisition mode during the 55 C phase. The trans-
gene-specific and wild-type-specific primers and
probes were added to all reactions, but two different
primer-probe mixes were used, one specific to the
DAS-59122-7 transgene and insertion site and one
specific to the NK603 transgene and insertion site
(Table 1, online resource). Each sample was assayed
in triplicate and a no-template negative control for
each primer-probe mix was also processed under the
123

same conditions. The data were analyzed using
OriginPro v8.1 (Northampton, USA).
Next-generation sequencing
Barcoded PCR products were purified and diluted
using the Ion Library Equalizer Kit and then
sequenced in a pool on an Ion Torrent Sequencer
(Life Technologies). The results were analyzed using
Lasergene Genomic Suite software (DNA Star, Mad-
ison, USA) and the SeqMan NGen program was used
to assemble the NGS reads on the reference sequences
described above, for subsequent analysis with SeqMan
Pro. The barcodes shown in Table 5 (online resource)
were used to differentiate among the samples.
Results and discussion
Primer design and initial testing
We developed our NGS-based zygosity detection
procedure using two well-established transgenic
maize events as models, namely DAS-59122-7 (Dow
AgroSciences, Indianapolis, USA) and NK603 (Mons-
anto Creve Coeur, Missouri). Primers were designed
to amplify transgene-specific products overlapping the
50 border of the transgenic locus and corresponding
wild-type products at the unfilled site (Table 3, online
resource). In each case, the forward primer was
specific for the 50 flanking sequence at the integration
site and was therefore common to both the transgene-
specific and wild-type-specific reactions. The reverse
primers were specific to the 50 sequence of the
transgene or to the genomic sequence downstream of
the transgene integration site (Fig. 1). Primers with
RW primer RW primer
FW primer TG specific WT specific
Maize genome Transgene Maize genome
Fig. 1 The primer binding sites used to generate NGS
templates were similar in both transgenic events. The forward
primer binds upstream of the transgene, one reverse primer
binds within the transgene and another reverse primer binds
downstream of the transgene. Arrows show primers and the
direction of elongation. Horizontal bars show stretches of DNA
with green bars representing genomic DNA and orange bars
representing the transgene. FW forward, RW reverse, WT wild-
type, TG transgene. (Color figure online)
123
Transgenic Res
attached adapters (Table 4, online resource) and
barcodes (Table 5, online resource) were used to
minimize sample preparation steps by generating PCR
products linked with two adapters and one sample-
specific barcode (Fig. 2).
Amplification of the entire transgene locus by the
forward primer and the wild-type-specific reverse
primer was prevented by limiting the elongation time
in the cycling program to 30 s. This is insufficient time
to generate the full-size products ([7,000 bp for the
DAS-59122-7 event and [4,000 bp for the NK603
event) because the polymerization rate
is * 1,500 bp/min (Roche 2011) and non-specific
single strands are degraded, as confirmed by the
amplification of single bands representing the trans-
genic allele (Fig. 3). Reference sequences were
obtained from the corresponding patents for each
event (Behr et al. 2002; Bing et al. 2011) and, in the
case of NK603, also by sequencing the long-range
PCR products covering the complete transgene
because the patent does not contain sufficient
sequence data for primer design (data not shown).
PCR products were size fractionated by capillary
electrophoresis using the Agilent 2100 Bioanalyzer to
confirm the reactions were successful and determine
whether the zygosity could already be detected by
electrophoresis (Fig. 3). The same PCR products were
later used for NGS. The DAS-59122-7 samples
generated two specific products differing in length
by 77 bp, the larger band representing the wild-type
allele and appearing fainter than the transgene-specific
product in heterozygous (HZ) samples. The NK603
samples also generated two products but only one
template
DNA
forward primer barcode
P1-adapter reverse primer A-adapter
forward primer reverse primer
P1-adapter PCR product A-adapter
barcode
Fig. 2 PCR procedure to generate NGS templates. The NGS-
ready adapters are attached to the PCR primers before
amplification. An additional barcode is attached to the reverse
primer for sample identification after sequencing. The PCR
yields a product containing two adapters and one barcode
Transgenic Res

(a)

Negative control

Negative control
FW primer RW primer

WT NK603
TG NK603

HZ NK603
WT DAS
TG DAS

HZ DAS
Maize genome Transgene Maize genome
Ladder

NK603
DAS
Integration site
(b)
FW primer RW primer
400 bp
300 bp
200 bp Maize genome

Fig. 3 Electrophoresis of PCR products used for NGS. The Fig. 4 The primer binding sites for real-time PCR were similar in
expected product sizes are: DAS-59122-7 transgene-specific both transgenic events. Horizontal bars show stretches of DNA
product, 260 bp; DAS-59122-7 wild-type-specific product, with green bars representing genomic DNA and orange bars
337 bp; NK603 transgene-specific product: 312 bp; NK603 representing the transgene. Arrows show primers and the direction
wild-type-specific product, 314 bp. The heterozygous samples of elongation. The fluorogenic probe binds between the forward
should yield both products. The Agilent High Sensitivity DNA and reverse primers (not shown). a In the transgenic allele, the
ladder was used to provide size standards. The violet and green forward primer binds upstream of the transgene and the reverse
bands mark the upper and lower sizes, respectively. TG primer binds within the transgene. b In the wild-type allele, the
homozygous transgenic, WT homozygous wild-type, HZ het- forward primer bind upstream of the transgene integration site
erozygous. (Color figure online) (orange arrow) and the reverse primer binds downstream of the
transgene integration site. All four primers are added to one
reaction. FW forward, RW reverse. (Color figure online)

band was visible because the bands representing the

transgenic and wild-type alleles differed in length by reactions regardless of event and genotype was
only 2 bp. Therefore, the zygosity could already be designed to bind to the 50 genomic flanking sequence.
determined by comparing electrophoretic DNA bands For both events, the transgene-specific probe and
if the allele-specific products differed sufficiently in reverse primer were designed to bind to the transgene
length, but it can be difficult to design PCR primers sequence. In the case of NK603, the wild-type-specific
that achieve visual separation. The possibility of reverse primer and probe were designed to bind
primers annealing to random sequences in the genome downstream of the integration site, whereas in the case
is also not ruled out with this approach, which may of DAS-59122-7 the probe annealed across the inte-
explain the uncharacterized faint bands that were gration site and the reverse primer annealed down-
present in some of the samples, e.g. in the heterozy- stream of the integration site (Fig. 4 and Table 1,
gous NK603 (Fig. 3) sample. This shows that inaccu- online resource).
rate results could arise from the detection of random The amplification curves generated by these prim-
amplification products or through contamination with ers matched the expected targets, showing exclusively
foreign DNA. Additionally, an assay involving 11 a positive wild-type-specific amplification curve in the
samples would take 35 min to complete and the cost wild-type sample, a transgene-specific amplification
would scale with increasing sample numbers making it curve in the transgenic sample and both curves in the
impractical for large-scale analysis (Vitale 2000). heterozygous sample (Fig. 5). The signal intensity
varied substantially among the reactions and some
Zygosity detection by real-time PCR amplification reactions produced a rather flattened
curve. Nevertheless, the difference between genuine
Two discrete real-time PCRs for the wild-type (WT) amplification curves and the negative control was
and transgenic (TG) alleles were combined in a single clear. The wild-type-specific amplification curves also
reaction to avoid the need for an endogenous reference showed a negative control-like signal in the homozy-
or standard curve and internal calibration steps. One gous transgenic sample, as did the transgene-specific
wild-type-specific primer pair and the corresponding amplification curve in the homozygous wild-type
probe, and one transgene-specific primer pair and the sample. Therefore, false positive results could be
corresponding probe, were therefore mixed in a single ruled out by comparing the curves to these anticipated
reaction vessel (Fig. 4). The forward primer in all negatives (Fig. 5a, b, e, f). We were therefore able to

123

(a) (b)
(c) (d)
(e) (f)
(g) (h)
Fig. 5 Real-time PCR amplification curves of the DAS-59122-
7 transgenic event (ad) and the NK603 transgenic event (e
h) using the corresponding primer and probe set corresponding
to the transgene and unfilled insertion site. The solid line
represents the transgene-specific FAM reporter dye, and the
dashed line represents the wild-type-specific Cy5 reporter dye.
Only cycles 2040 are shown. a DAS-59122-7 genotyping with
the wild-type template. b DAS-59122-7 genotyping with the
homozygous transgenic template. c DAS-59122-7 genotyping
with the heterozygous template. d DAS-59122-7 no-template
negative control. e NK603 genotyping with the wild-type
template. f NK603 genotyping with the homozygous transgenic
template. g NK603 genotyping with the heterozygous template.
h NK603 genotyping with the no-template negative control.
DRn describes the magnitude of the signal under each set of PCR
conditions and is based on the signal produced by a reaction
containing all components minus the signal produced by a no-
template control
123
Transgenic Res
detect the DAS-59122-7 and NK603 transgenes and
determine the zygosity of the DNA samples.
The controversy surrounding the use of real-time
PCR to detect zygosity reflects the ability of the
technique to differentiate between two-fold differ-
ences when only one primer/probe pair is used
(Bubner and Baldwin 2004; Bubner et al. 2004; Prior
et al. 2006). We addressed this drawback by using
separate primer/probe combinations for each allele,
which also avoided the need to calculate the abun-
dance of each product against a calibration or standard
curve or to carry out efficiency corrections for
individual reactions thus reducing time and costs
(Bubner and Baldwin 2004; Bubner et al. 2004;
Gachon et al. 2004). The ability to omit calibration
steps makes the assay more suitable for larger-scale
applications because more samples can be processed
in a short time, but overall the real-time PCR approach
is still limited by the throughput of current real-time
PCR machines which can accommodate microtiter
plates containing up to 384 wells. This means that
approximately 380 samples plus controls can be tested
in the time required for assay setup, pipetting and the
amplification itself (2 h). Although the results were
satisfactory and confirmed the zygosity of each
transgenic event, the amplification curves were not
equally steep and could be difficult to categorize e.g.
in case of the heterozygous NK603 sample (Fig. 5g).
Therefore a second approach using NGS was tested to
determine whether more robust data could be
acquired.
Zygosity detection by NGS
The original PCR products, that were sorted by size on
the Bioanalyzer (Fig. 3) were prepared for NGS and
the reads were aligned either to the 50 end of the
transgene or to the unfilled integration site in the wild-
type genomic DNA. The alignment of the NGS reads
to these reference sequences provided unambiguous
specificity for the expected genotype (Fig. 6). The
homozygous transgenic samples representing each
event exclusively showed reads that were aligned to
the corresponding transgene sequence and no reads
aligned with the genomic sequence around the unfilled
integration site (Fig. 6a, d). The wild-type samples
only showed reads that were aligned with the wild-
Transgenic Res

(a) (b) preferential amplification of certain targets (Le et al.

800 800 2009). The lower number of reads for the wild-type
No. of reads

No. of reads
600 600 PCR product in the DAS-59122-7 heterozygous sam-
400 400 ple (Fig. 6c) was anticipated by the fainter band
200 200 representing the wild-type amplification product in the
0 0 heterozygous DAS-59122-7 sample (Fig. 3). This
TG WT TG WT
difference in amplification efficiency could be reduced
(c) (d) by optimizing the PCR conditions. In our case, such
No. of reads 250 time-consuming work was not necessary because we
400
No. of reads

200 sought a yes/no result for the presence of each

300 150
transgene, which can be achieved independently of
200 100
differences in the read number. Differences in read
100 50
number among the homozygous samples occur
0 0
TG WT TG WT because of minor differences in the concentration of
the PCR products or read errors/low quality reads
(e) (f)
250
excluded from the final dataset. This often occurs
250
when automatically trimming or filtering out reads
No. of reads
No. of reads

200 200
150 150
100 100
50 50
0 0
TG WT TG
Fig. 6 Zygosity of the DAS-59122-7 (ac) and NK603 (d
f) transgenic events determined by NGS, showing the number of
reads aligned to the corresponding transgenic and wild-type
reference sequences. a DAS-59122-7 reads on the homozygous
transgenic template. b DAS-59122-7 reads on the corresponding
wild-type template. c DAS-59122-7 reads on the heterozygous
template. d NK603 reads on the homozygous transgenic
template. e NK603 reads on the corresponding wild-type
template. f NK603 reads on the heterozygous template.
TG = reads aligned to the corresponding 50 transgene sequence;
WT = Reads aligned to the corresponding wild-type genomic
sequence flanking the transgene integration site
type sequence surrounding the transgene integration
site, and there were no reads that aligned with the
transgene itself (Fig. 6b, e). The heterozygous sam-
ples showed reads aligning with both the transgene
and the wild-type sequence flanking the insertion site
(Fig. 6 c, f). Although the number of aligned reads
differed among the samples, robust and unambiguous
zygosity data were generated by the alignment of more
than 100 reads to each reference sequence in most
samples. In the heterozygous samples, the number of
reads aligned to the wild-type and transgenic sequence
differed because of the unequal efficiency of ampli-
fication during the synthesis of the PCR products. This
problem often occurs during multiplex PCR and
reflects artifacts such as primer dimerization and the
which, for example, are polyclonal or have an off-
scale signal on the Ion Torrent server (Life Technol-
ogies 2014). False positives, e.g. caused by the random
annealing of PCR primers, could be ruled out through
WT
the large number of similar reads and the unambiguous
alignment to the reference sequences.
Even the contamination of NGS samples with
external DNA after the PCR step would not affect the
results because the reads were aligned and sorted by
the barcode attached during the PCR, allowing the
identification of differentially-barcoded samples
pooled before sequencing and the exclusion of
sequences lacking a barcode (Fig. 2). The sample-
specific use of barcodes allows plant material to be
screened for several different transgenes in a single
sequencing run while also reporting their zygosity.
This was verified by the use of three different barcodes
assigned to three maize DNA samples differing in
zygosity, but using the same barcode for both events.
Barcode A01 was assigned to the homozygous trans-
genic maize, A02 to the homozygous wild-type maize
and A03 to heterozygous samples (Table 5, online
resource). The barcode is added to distinguish between
samples after the sequencing run and was assigned to
the samples of known zygosity as an example. All
samples were sequenced together and the alignment of
reads to the reference sequence showed only the
sequence of the targeted PCR products. Furthermore,
it was not necessary for the two endpoint PCRs of each
transgenic event to be equivalent in efficiency. As
shown in Fig. 6c and f, the numbers of reads aligned to
the two different references were not equal.

123

Technically, only one read is necessary to detect a
given PCR product and therefore identify the corre-
sponding allele. Nevertheless, a higher number of
reads is desirable to make the data more robust.
Distinct values for the number of reads required in this
kind of assay are, to our knowledge, not yet defined,
but it is clear that a greater depth of reads would reduce
the error rate significantly (Sims et al. 2014).
Our approach is advantageous because there is no
need to normalize or equalize the PCRs to produce
robust zygosity data. Other studies involving the
detection of transgene copy number by NGS have
suggested that depth normalization is required
(Medvedev et al. 2010; Boeva et al. 2011; Klam-
bauer et al. 2012; Duan et al. 2013). However, it is
possible to generate 80 million reads per run so our
assay is suitable for the high-throughput screening
of a large number (possibly thousands) of samples
containing one or more transgenes in one experi-
ment, allowing the detection of each event and the
corresponding zygosity. Our approach also reduces
the need for time-consuming sample preparation
steps such as fragmentation, sizing and adapter
ligation. The NGS technique we use involves the
ligation of two sequencing adapters and one barcode
for sample identification to DNA molecules with a
maximum length of 400 bp. The adapters (Table 4,
online resource) and barcodes (Table 5, online
resource) are added during the PCR step and the
products are short enough to be sequenced directly
(Fig. 2), therefore these additional sample prepara-
tion steps after the PCR are omitted.
In summary, our novel NGS-based method could be
widely adopted as a tool for high-throughput transgene
detection and zygosity determination in transgenic
plants and their products because a large number of
samples can be processed simultaneously. Neverthe-
less, this assay requires that the exact integration site
and adjacent genomic sequences are known, as well as
the transgene copy number. This assay is therefore
suitable to provide zygosity information about trans-
genic events with a known integration site and copy
number, e.g. to monitor crossing/breeding programs
and to ensure traceability of the corresponding trans-
gene. However it is unclear whether this assay will be
suitable for tetraploid plants such as potato and
hexaploid crops such as wheat. In such cases, the
PCR equalization and the normalization of reads may
once again become necessary because the ratio of
123
Transgenic Res
reads will need to be determined in order to assign the
zygosity of each sample correctly.
Acknowledgments We acknowledge Raphael Soeur and
Claudia Hansen (Fraunhofer IME, Aachen, Germany) for
performing the next-generation sequencing and for the helpful
discussions about sample processing and data management. We
also acknowledge Euregio Analytik BioChem GmbH for
cooperation in the project and for providing the template DNA
material. We acknowledge Dow AgroSciences and KWS Saat
AG for making the transgenic kernels/DNA available for the
project. We acknowledge Dr. Richard M Twyman for editing
the manuscript. This work is supported by the BMBF
(0316075B).
References
Ahmed FE (2002) Detection of genetically modified organisms
in foods. Trends Biotechnol 20:215223. doi:10.1016/
S0167-7799(01)01920-5
Behr C, Heck G, Hironaka C, You J (2002) Nucleotide
sequences for use in the generation of herbicide resistant
plants. US20020013960A1
Bing JW, et al (2011) Corn event DAS-59122-7 and methods for
detection thereof. US007875429B2
Boeva V, Zinovyev A, Bleakley K, Vert JP, Janoueix-Lerosey I,
Delattre O, Barillot E (2011) Control-free calling of copy
number alterations in deep-sequencing data using GC-
content normalization. Bioinformatics 27:268269. doi:10.
1093/bioinformatics/btq635
Bubner B, Baldwin IT (2004) Use of real-time PCR for deter-
mining copy number and zygosity in transgenic plants.
Plant Cell Rep 23:263271. doi:10.1007/s00299-004-
0859-y
Bubner B, Gase K, Baldwin IT (2004) Two-fold differences are
the detection limit for determining transgene copy numbers
in plants by real-time PCR. BMC Biotechnol 4:14. doi:10.
1186/1472-6750-4-14
Deschamps S, Llaca V, May GD (2012) Genotyping-by-
sequencing in plants. Biology 1:460483. doi:10.3390/
biology1030460
Duan J, Zhang JG, Deng HW, Wang YP (2013) Comparative
studies of copy number variation detection methods for
next-generation sequencing technologies. PloS one
8:e59128. doi:10.1371/journal.pone.0059128
Franssen SU, Shrestha RP, Brautigam A, Bornberg-Bauer E,
Weber AP (2011) Comprehensive transcriptome analysis
of the highly complex pisum sativum genome using next
generation sequencing. BMC Genom 12:227. doi:10.1186/
1471-2164-12-227
Gachet E, Martin GG, Vigneau F, Meyer G (1998) Detection of
genetically modified organisms (GMOs) by PCR: a brief
review of methodologies available. Trends Food Sci
Technol 9:380388. doi:10.1016/S0924-2244(99)00002-3
Gachon C, Mingam A, Charrier B (2004) Real-time PCR: what
relevance to plant studies? J Exp Bot 55:14451454.
doi:10.1093/jxb/erh181
Transgenic Res
Gadaleta A, Giancaspro A, Cardone M, Blanco A (2011) Real-
time PCR for the detection of precise transgene copy
number in durum wheat. Cell Mol Biol Lett 16:652668.
doi:10.2478/s11658-011-0029-5
German MA, Kandel-Kfir M, Swarzberg D, Matsevitz T, Granot
D (2003) A rapid method for the analysis of zygosity in
transgenic plants. Plant Sci 164:183187. doi:10.1016/
S0168-9452(02)00381-3
Goya R et al (2010) SNVMix: predicting single nucleotide
variants from next-generation sequencing of tumors. Bio-
informatics 26:730736. doi:10.1093/bioinformatics/
btq040
Hernandez M, Pla M, Esteve T, Prat S, Puigdomenech P, Ferr-
ando A (2003) A specific real-time quantitative PCR
detection system for event MON810 in maize YieldGard
based on the 30 -transgene integration sequence. Transgenic
Res 12:179189
ISAAA (2014) International service for the acquisition of agri-
biotech applications. http://isaaa.org/contactus/default.asp.
Accessed 27 June 2014
Klambauer G, Schwarzbauer K, Mayr A, Clevert DA, Mitter-
ecker A, Bodenhofer U, Hochreiter S (2012) cn.MOPS:
mixture of Poissons for discovering copy number varia-
tions in next-generation sequencing data with a low false
discovery rate. Nucleic Acids Res 40:e69. doi:10.1093/nar/
gks003
Le T, Hidalgo Ashrafi E, Paul N (2009) Enhancing multiplex
PCR efficiency using Hot Start dNTPs. BioTechniques
47:972973
Life Technologies (2014) Trimming and filtering. Life-Tech-
nologies. http://137.226.86.55/ion-docs/Technical-Note
Filtering-and-Trimming. Accessed 27 Nov 2014
Mascher M, Wu S, Amand PS, Stein N, Poland J (2013)
Application of genotyping-by-sequencing on semicon-
ductor sequencing platforms: a comparison of genetic and
reference-based marker ordering in barley PloS one
8:e76925. doi:10.1371/journal.pone.0076925
Medvedev P, Fiume M, Dzamba M, Smith T, Brudno M (2010)
Detecting copy number variation with mated short reads.
Genome Res 20:16131622. doi:10.1101/gr.106344.110
Metzker ML (2010) Sequencing technologiesthe next gen-
eration. Nat Rev Genet 11:3146. doi:10.1038/nrg2626
Network TC (2013) Corrigendum: comprehensive genomic
characterization defines human glioblastoma genes and
core pathways. Nature 494:506. doi:10.1038/nature11903
Prior FA, Tackaberry ES, Aubin RA, Casley WL (2006)
Accurate determination of zygosity in transgenic rice by
real-time PCR does not require standard curves or effi-
ciency correction. Transgenic Res 15:261265. doi:10.
1007/s11248-005-4024-3
Roche (2011) Expand high fidelity PCR system, 20th edn.
Roche Diagnostics, Mannheim (Germany)
Schmidt M, Parrott W (2001) Quantitative detection of trans-
genes in soybean [Glycine max (L.) Merrill] and peanut
(Arachis hypogaea L.) by real-time polymerase chain
reaction. Plant Cell Rep 20:422428. doi:10.1007/
s002990100326
Sims D, Sudbery I, Ilott NE, Heger A, Ponting CP (2014)
Sequencing depth and coverage: key considerations in
genomic analyses. Nat Rev Genet 15:121132. doi:10.
1038/nrg3642
Teixeira RT, Fortes AM, Pinheiro C, Pereira H (2014) Com-
parison of good- and bad-quality cork: application of high-
throughput sequencing of phellogenic tissue. J Exp Bot.
doi:10.1093/jxb/eru252
Vitale D (2000) Comparing the Agilent 2100 Bioanalyzer per-
formance to traditional DNA analysis techniques. Agilent
Technologies, Santa Clara (USA)
Yockteng R, Almeida AM, Yee S, Andre T, Hill C, Specht CD
(2013) A method for extracting high-quality RNA from
diverse plants for next-generation sequencing and gene
expression analyses. Appl Plant Sci. doi:10.3732/apps.
1300070
Zimmermann A, Luthy J, Pauli U (2000) Event specific trans-
gene detection in Bt11 corn by quantitative PCR at the
integration site. LWT-Food Sci Technol 33:210216.
doi:10.1006/fstl.2000.0637
123