![DNA amplification with the primer pair 2550F and 2718R [19]](https://www.researchgate.net/profile/Herbert-Fuhrmann/publication/51038472/figure/tbl1/AS:668669231239174@1536434576129/DNA-amplification-with-the-primer-pair-2550F-and-2718R-19_Q320.jpg)
Content uploaded by Herbert Fuhrmann
Author content
All content in this area was uploaded by Herbert Fuhrmann on Sep 17, 2014
Content may be subject to copyright.
ORIGINAL PAPER
Gender determination of fertilized unincubated chicken eggs
by infrared spectroscopic imaging
Gerald Steiner &Thomas Bartels &Allison Stelling &
Maria-Elisabeth Krautwald-Junghanns &
Herbert Fuhrmann &Valdas Sablinskas &Edmund Koch
Received: 13 January 2011 / Revised: 20 March 2011 / Accepted: 21 March 2011 / Published online: 9 April 2011
#Springer-Verlag 2011
Abstract Each year, billions of day-old layer chicks are
produced in the world. Since only female chicks are reared
for egg production, the chicks must be sexed and the
unwanted male layer chicks are culled. The culling of male
chicks is a serious problem, both in terms of animal welfare
and waste disposal. The germinal disc in fertilized but
unincubated eggs contains already several thousands of
blastoderm cells. The cellular DNA in birds is different for
male and female chicks. The difference in DNA content
between male and female chicks is around 2% and is
measurable by Fourier transform infrared (FT-IR) spectros-
copy. In this study, small amounts of blastoderm cells from
22 chicken eggs were characterized by attenuated total
reflection FT-IR spectroscopic imaging and classified by
linear discriminant analysis. Polymerase chain reaction
(PCR) was used as a reference method to determine the
gender. The spectroscopic results demonstrate that male
blastoderm cells exhibit a higher content of DNA than cells
from female blastoderm. The spectroscopic-based gender
determination led to the same result as the PCR analysis.
FT-IR spectroscopic imaging allows the gender determina-
tion of unincubated eggs within a few seconds based on the
accurate determination of the different DNA contents in
blastoderm cells of both sexes.
Keywords Spectroscopy .IR spectroscopy .Bioanalytical
methods
Introduction
Each year, approximately 280 million male day-old chicks
from layer strains are culled in the European Union [1].
Currently, after hatching, the gender of the day-old chick is
determined by vent sexing [2], feather sexing, or plumage
color sexing [3]. These methods rely on the visual
identification of sex based on gender-specific appearance
of feathers or cloacal morphology [4]. After gender
discrimination, the male day-old chicks are culled by
carbon dioxide exposure or maceration. From the animal
welfare point of view, the killing of cockerels just after
hatch is not acceptable [5]. Clearly, there is a need for new
techniques that avoid the breeding of "male eggs" [6]. One
approach is to determine the gender of 16- to 18-day-old
embryos by the quantification of the hormone levels in the
allantoic fluid [4,7]. However, this method requires that the
fertilized egg be incubated for at least 16 days to develop an
embryo. Furthermore, it is quite difficult to take a sample
Published in the special issue Biophotonics with Guest Editors Jürgen
Popp and Reiner Salzer.
G. Steiner (*):A. Stelling :E. Koch
Faculty of Medicine Carl Gustav Carus,
Clinical Sensoring and Monitoring,
Dresden University of Technology,
Fetscherstr. 74,
01307 Dresden, Germany
e-mail: gerald.steiner@tu-dresden.de
T. Bartels :M.-E. Krautwald-Junghanns
Faculty of Veterinary Medicine, Clinic for Birds and Reptiles,
University of Leipzig,
An den Tierkliniken 17,
04103 Leipzig, Germany
H. Fuhrmann
Faculty of Veterinary Medicine,
Institute of Physiological Chemistry, University of Leipzig,
An den Tierkliniken 1,
04103 Leipzig, Germany
V. Sablinskas
Faculty of Physics,
Department of General Physics and Spectroscopy,
University of Vilnius,
01108 Vilnius, Lithuania
Anal Bioanal Chem (2011) 400:2775–2782
DOI 10.1007/s00216-011-4941-3
from a well-defined point within the egg without degradation
of the embryo. Determining the gender in a very early stage,
before incubation, would be an ideal solution [8]. At this early
developmental stage, approximately 40,000–60,000 blasto-
derm cells are available in the germinal disc [9]. Female birds
are heterogametic with one Z and one W sex chromosome,
whereas male birds have two Z chromosomes [10]. By
taking small amounts of cells and performing a polymerase
chain reaction (PCR) analysis, the gender of a fertilized egg
can be reliably identified as male or female [11,12].
However, this approach is time-consuming, expensive, and
not yet available to scale up from laboratory to industrial
applications, for example in a hatchery [6]. In addition, PCR
requires relatively large amounts of blastoderm cells so that
the probability of the development of a normal chick will be
dramatically reduced. Another idea is based on the character-
ization of the egg morphology. For hundreds of years, poultry
breeders and scientists have speculated about a relationship
between the egg's size and the gender. A recent published
scientific study deals with the gender determination of
chicken eggs by a precise characterization of shape and size
of the egg [13]. However, this study clearly demonstrates
that no relationship between the morphology of the egg and
gender of the hatched chick exists. Although many efforts
have been made to determine the gender of unincubated
eggs, there is no potential solution on the horizon.
Following our successful study [14]inwhichwe
determined the gender of birds by Fourier transform
infrared (FT-IR) spectroscopic characterization of DNA-
rich cell material from feather pulps, we have extended the
methods on the investigation of blastoderm cells from
unincubated eggs. The discrimination is based on the
absorption bands of the phosphate groups from the nucleic
acids. The Z chromosome has approx. threefold higher
DNA content than the W chromosome [15]. All other
chromosomes are identical for males and females so that
the total DNA amount is approx. 2% higher in cells from
male (two Z chromosomes) than in cells from female birds
(one Z, one W chromosome) [16]. Although the extraction
of specific molecular information from an infrared spectrum
of cells is not possible, application of multivariate methods
for spectral feature selection has produced promising results
in such fields as tumor identification or classification of
different cell types. A reliable assessment of the blastoderm
cell samples requires a spectroscopic imaging technique
because the cells are in a solution of fatty acids and
proteins. FT-IR imaging spectroscopy combines the high
molecular sensitivity with a spatial resolution down to a
few micrometers so that spectra from blastoderm cells can
be selected for future data analysis. In contrast to common
(not spatially resolved) FT-IR spectroscopy, the imaging
approach allows the selection of only those spectra that
exhibit information about blastoderm cells. For example,
spectra that only represent yolk or albumen will be not
considered. The determination of the gender from a very small
amount of blastoderm cells, taken from an unincubated egg, is
the objective of the present study.
Experimental
Sample preparation
Chicken eggs of a white layer strain (Lohmann LSL) were
obtained from the Lohmann Tierzucht GmbH (Cuxhaven,
Germany). The freshly laid eggs were kept at approximately
4 °C until their use for preparation of the blastoderm cell
samples. For a better understanding of the preparation
procedure, Fig. 1a shows the general internal structure of a
bird egg. The egg consists of a yolk, which is covered by a
5- to 10-μm-thick yolk membrane surrounded by albumen
enclosed within the shell membrane and shell. The yolk is
rich in fats, vitamins, and minerals, whereas the albumin,
also known as “egg white,”is rich in protein. The chalazae
are twisted fibrous structures which extend outwards through
the albumen and hold the yolk in the center of the egg. After
rotating the egg, the yolk moves always upward. Once the
egg is fertilized, the germinal disc forms on top of the yolk
during the preoviposition, oviductal period [17]. In the
freshly laid egg, it can be identified on the surface of the
yolk as a 4- to 5-mm disc of a slightly different color. The
germinal disc contains about 40,000–60,000 blastoderm cells
[9]. The eggs were stored in a horizontal position for 1 day
to ensure a position of the germinal disc on the top of the egg
(as depicted in Fig. 1a). In a first step, the shell was carefully
opened by a cut of a hole (approximately 1 cm in diameter)
so that the yolk and the germinal disc were clearly visible.
Blastoderm cells were removed by a hook under eye control,
as illustrated in Fig. 1b. A small amount of the sample
(<1 μg) was placed onto an attenuated total reflection (ATR)
crystal surface and dried on air. The larger part of the
removed cells was taken for PCR analysis.
Polymerase chain reaction
DNA extraction was performed by alkaline lysis of the
samples as describe elsewhere [18]. For the extraction, the
germ discs were incubated in 60 μl 0.2 N NaOH for 20 min
at 75 °C. Subsequently, the samples were neutralized with
180 μl 0.04 M Tris–HCl (pH 7.5). After that, the samples
are centrifuged at 14,000 rpm and 4 °C for 10 min and the
supernatants transferred into a new reaction vessel.
Amplification of DNA from the germ disc sometimes
failed. Obviously, the adherent egg yolk rich in lipids
interferes with the amplification. This was avoided in most
instances by defatting of the supernatants. For that, 100 μl
2776 G. Steiner et al.
samples were mixed vigorously with 300 μl chloroform,
incubated for 2 min, and mixed again. The mixture was
centrifuged at 12,000 rpm at room temperature for 10 min
and the supernatant transferred into new vessels. Subse-
quently, the supernatant was taken to measure the content
of DNA by UV spectroscopy and used as a template for
PCR. The method used the primers as described by
Fridolfsson and Ellegren [19], which amplify a part of
the avian CHD-1 gene (chromohelicase DNA binding
protein) located on both sex chromosomes, but with a
different intron size on the Z and W chromosomes
(2550F: 5′-GTT ACT GAT TCG TCT ACG AGA-3′,
2718R: 5′-ATT GAA ATG ATC CAG TCG TTG-3′). The
DNA fragments were amplified in the PCR Cycler T Gradient
(Biometra GmbH, Göttingen, Germany). As a positive
control, DNA samples from known cocks and hens were
used; a sample with aqua bidest served as a negative control.
Tab le 1summarized the substances and their volume used for
PCR analysis.
The temperature profile consisted of an initial denaturing
step at 94 °C for 2 min, followed by a “touchdown”scheme
[20]. For 30 s of each cycle, the initial annealing temperature
is lowered from 60 °C through 0.5 °C to an end temperature
of 50 °C. The total number of the cycles was 40. When the
amount of DNA extracted from the germ discs was low, the
reaction was carried out as a two-stage “nested”PCR using
1μl of the first round for the second round.
The PCR products were separated by agarose gel electro-
phoresis, stained with ethidium bromide, and visualized by
transillumination. With the primers 2550F and 2718R, a part
of the avian CHD1-W and/or CHD1-Z gene is amplified.
Fragments of 600–650 bp from the Z chromosome and
fragments of 400–450 bp for the W chromosome are
expected.
Table 1 DNA amplification with the primer pair 2550F and 2718R [19]
Substances Volume (μl)
DNA 2.0
MgCl
2
(50 mM) 1.5
Reaction buffer (10×) 2.5
dNTPs (2 mM) 2.5
Primer 2550F (10 pmol/μl) 2.5
Primer 2718R (10 pmol/μl) 2.5
Taq-DNA polymerase (5 U/μl) 0.3
ddH
2
O25
Fig. 1 a Structure of an avian
egg. bIllustration of the sample
preparation for ATR FT-IR
spectroscopic imaging
IR gender determination of fertilized unincubated chicken 2777
FT-IR spectroscopic imaging
Infrared spectroscopic images were collected in ATR
mode using the Bruker spectrometer IFS60 (Bruker
Optik GmbH, Ettlingen, Germany) coupled to an ATR
macro imaging unit (Fast IR, Harrick). FT-IR images
were collected on a single-reflection ZnSe ATR crystal
with a 60° angle of incidence. The imaging detector was
a Santa Barbara focal plane MCT (mercury cadmium
telluride) 64×64 array detector. Since the ATR image
represents an area of approx. 4× 4 mm
2
, the samples,
however, covered an area on the crystal surface of approx.
12×5 mm
2
. Six individual FT-IR images from different
positions of the ATR crystal were recorded. This ATR
mapping was achieved by a slight moving of the ATR
crystal and a following adjustment of the reflected beam
on the FPA sensor chip. This approach ensures that all
ATR spectra are recorded under the same angle of
incidence. A reference spectroscopic image was captured
from the same ZnSe crystal without any samples. A total
number of 40 interferograms were co-added for each of
the 4,096 image pixels. The recorded data set comprised
6×4,096=24,576 spectra. The interferograms were
Fourier-transformed applying Happ–Genzel apodization
and zero filling factor of 1. Spectra at a resolution of 6 cm
−1
of the sample image were ratioed against the spectra of the
reference image and transferred to absorbance values. The
frame rate of the camera was 516 Hz, yielding a total
measurement time of approx. 4 min for each image. Finally,
the individual FT-IR spectroscopic images were assembled
into one large image. One pixel represents a sample area of
approx. 63×63 μm.
Data pre-processing
Evaluation of spectral data was performed using the Matlab
Package (version 7, Math Works Inc. Natick, MA, USA). A
main part of the data analysis is based on in-house written
programs, in particular for data pre-processing and image
processing. In order to minimize the data volume, only the
so-called fingerprint region between 950 and 1,800 cm
−1
was considered. Data pre-processing firstly involves the
removal of outliers. We define as outliers spectra that are
obviously not associated to the samples. An absorbance of
the amide I band of <0.02 was used as the criterion for
outlier detection. From the entire data set of 24,576 spectra,
22,238 spectra were detected as outliers. This relatively
large number results from the fact that samples only cover a
very small area on the ATR crystal surface. In a second
step, a linear two-point baseline correction was performed
for each spectrum. Absorbance values at 1,790 and
1,000 cm
−1
were used for the linear baseline correction.
Finally, each absorbance value of the spectrum is normal-
ized to the sum of all absorbance values.
Results and discussion
The results of the PCR analysis are shown in Fig. 2. The
female genome (labeled as “w”) exhibits two clear signals,
whereas the male genome (labeled as “m”) shows only one
signal. The labels “K33”to “K54”denote the sample
numbers. After a detailed inspection, samples K33, K34,
K37, K40, K43, K46, K48, K49, K50, and K51 could be
identified as female genome and samples K38, K39, K42,
Fig. 2 Separation of the PCR
products by gel electrophoresis
2778 G. Steiner et al.
K44, K45, K47, K52, K53, and K54 as male genome. No
assignment was possible for K35, K36, and K41.
The ZnSe ATR crystal covered with the samples and the
FT-IR ATR spectroscopic images of the blastoderm samples
are shown in Fig. 3a. Figure 3b shows the black and white
image of the ZnSe ATR crystal surface. Each spectroscopic
image (Fig. 3c) covers a sample area of 4× 4 mm. For
every pixel, the integral intensity across the spectral
range 950–1,800 cm
−1
is transformed to a rainbow scale.
Green to red indicates high absorption, whereas dark blue
pixels correspond to the pure ATR crystal surface. Figure 3c
was computed from integrated intensities; it does not reveal
structural properties of the samples in different spots of the
surface.
As a starting point for spectral classification, the infrared
spectra classes will be briefly discussed. Averaged spectra
and standard deviation (Fig. 4) were calculated from spectra
according to the PCR analysis of the corresponding sample.
At the first glance, the spectra appear quite similar to the other.
Both spectra are dominated by the amide I band at 1,650 cm
−1
and amide II band at 1,550 cm
−1
which arise from the C=O
stretching and N–H bending vibrations, respectively, of the
amide groups comprising the peptide linkages of proteins
[21]. The relatively strong band at 1,730 cm
−1
indicates a
high amount of esterified fatty acid [22]. Two more bands
that are significant appear around 1,380 and 1,464 cm
−1
which are assigned to C–H vibrations of lipids. Another
protein absorption includes the weak band at 1,308 cm
−1
[23]. The group between 1,000 and 1,250 cm
−1
is mainly
composed of absorption bands of C–OandPO
2
−
groups of
nucleic acids, phospholipids, and carbohydrates [24,25].
Although the spectra are generally very similar, a clear
difference can be observed at 1,088 cm
−1
, a spectral position
that mainly indicates symmetric stretching vibrations from
phosphate groups of nucleic acids [22].
The absorption profile between 1,000 and 1,100 cm
−1
has already been used to classify blastoderm cells from
feather samples of female and male specimens [14]. Despite
this obvious difference at 1,088 cm
−1
between the spectra
of female blastoderm cells and spectra of male blastoderm
cells, the key question addressed here is whether the
biochemical information latent in these spectra is distinct
enough to classify even the gender of blastoderm cells of an
unincubated egg.
Since gender-relevant spectral features are weak and
because the spectra may be affected by variations of the
cells as well as by material from the yolk or albumen, a
non-subjective classification was performed. Many studies
Fig. 3 a Photograph of the ZnSe ATR crystal with the 22 samples on
the surface. bEnlarged black and white image of the ZnSe ATR
crystal surface and assignment of the sample numbers. cComposition
of six individual FT-IR spectroscopic images showing the integral
absorbance in the spectral range from 950 to 1,800 cm.
−1
IR gender determination of fertilized unincubated chicken 2779
have demonstrated that supervised classification is superior
to non-supervised methods even when the spectral features
are small and overlapped [26,27].
Figure 5shows the flowchart of the spectral classification.
After pre-processing of the data set, spectra of samples K33,
K34, K37, K39, K42, and K44 were used as a training set.
According to the PCR results (see Fig. 2), the samples K33,
K34, and K37 represent female blastoderm cells. The other
three samples represent male blastoderm cells. The female
blastoderm cells encompass 525 spectra and the male
blastoderm cells 280 spectra. In the following step, 250
spectra from each class (gender) were chosen by the algorithm
and used as training set.
The spectral procedure for developing the classification
model employs two algorithms in tandem. The program
takes as input both the spectra in the training set and their
PCR-based assignment to either gender. With this informa-
tion, the algorithm identifies a set of spectral subregions that,
taken together, consists of a pattern which serves as a basis to
group spectra according to the gender of the blastoderm cells.
The first part of the algorithm is an optimal region selection
routine. Each spectrum, which originally comprised 442 data
points, was thereby re-expressed as a set of six intensity
values. The second step is classification by linear discriminant
analysis. This approach is similar to a routine describe
elsewhere [28]. The model that provided the best agreement
was used for the classification of the test set spectra. The
robustness of the classification was validated using the leave-
one-out cross-validation method. Finally, the spectra of the
test set were classified. The classify function of the Matlab
Package returns a matrix containing an estimation of the
probability that the male or female training set was the
source of the assignment. These probability values were
color-coded and reassembled into an image. Thus, the
classification results are represented by a blue–gray–red
color scale. Blue represents spectra classified as male and red
as female. Gray values indicate spectra that could not be
clearly classified as male or female. Figure 6displays the
averaged spectra of cells from the test set and the six regions
chosen for the classification.
Three selected spectral regions at 1,082, 1,100, and
1,164 cm
−1
encompass absorptions that are ascribed to the
phosphate groups of nucleic acids respectively to C–OH
groups from proteins [29,30]. The averaged spectra of male
Fig. 4 Averaged spectra (bold line) and standard deviation of all
spectra representing male or female blastoderm cells
Fig. 5 Flowchart of the data processing and classification approach
2780 G. Steiner et al.
cells exhibit a higher absorbance than the averaged spectra of
female cells. This result is not surprising because the male
genome is approx. 2% larger than the female genome.
Beside these more general observations, the algorithm
identified variance within spectral regions which arise from
lipids and proteins. The region at 1,392 cm
−1
is attributed to
deformation vibrations of methyl groups of proteins and to
aliphatic side groups of amino acids [31]. Finally, vibrations
from proteins at 1,528 cm
−1
(amide II band) and the
symmetric stretching vibration of fatty acid esters at
1,743 cm
−1
are also selected by the algorithm [32]. The
antisymmetric stretching vibration of ester groups appear at
1,240 cm
−1
. However, this band is overlaid by many other
vibrations like antisymmetric stretching of the PO
2
−
groups
and collagen. In addition, the antisymmetric stretching mode
of ester groups is weaker than the symmetric mode at
1,743 cm
−1
. This fact, together with the contribution of other
absorption bands, might explain why no clear differences
appear at 1,240 cm
−1
. The association of the selected regions
at 1,392, 1,528, and 1,743 cm
−1
to a gender-specific
biochemical composition of cells is not clear.
Figure 7shows the classifications predicted for samples
designated to female and male, respectively.
Red pixels indicate the female gender and blue pixels male
gender. Most of the samples are assigned unambiguously to
one of the gender. In comparison with the PCR analysis, no
misclassification is present. Samples K35 and K47 exhibit a
mixture of pixels assigned to both types of gender. However,
the majority of the spectra of these samples are assigned to
male, which is in accordance with the PCR results. It should
be noted that the classification may also be affected by
contributions of the rest of yolk and albumen. Furthermore,
variation in size and shape of cells can lead to differences in
the classification. The results might be even better when
individual cells are evaluated. Therefore, a wrong assignment
of individual spectra in Fig. 6may be due to several reasons.
Firstly, the spectroscopic-based method is also sensitive to
variations of the chemical composition or development stage
of the cell. Secondly, cell division may lead to a temporary
increase of the DNA/RNA concentration because young
cells are relatively small, and thus the content of nucleic
acids is relatively high compared to the rest of the cell.
Finally, absorption bands of yolk and albumen affect also the
spectral pattern and can lead to an incorrect classification.
All these slight variations of the molecular composition lead
Fig. 6 Averaged spectra of female (red) and male (blue) blastoderm
cells. The gray bars indicate the spectral regions selected by the
classification algorithm
Fig. 7 Classification results of the FT-IR spectra reassemble to the
sample arrangement. Samples K33, K34, K37, K39, K42, and K44
were used for the training set (green). Red indicates a high percentage
for the gender classification as female, blue for male. The PCR results
are denoted as “f”for female and “m”for male blastoderm cells (cf.
Fig. 2)
IR gender determination of fertilized unincubated chicken 2781
to small changes in the spectral fingerprint and will affect the
accuracy of the spectral classification.
The results of this pilot study clearly demonstrate that
the FT-IR spectroscopy has the potential to identify the
gender of fertilized but unincubated chicken eggs. The
spectroscopic-based classification of a sample was always in
agreement with the PCR analysis. This fact is important for
animal welfare because FT-IR spectroscopy is the only
method accurate and rapid enough to be applied in hatcheries
to select only female chicken eggs for the incubation and thus
avoiding the killing of millions of day-old male chicks. Future
developments are devoted to ATR spectroscopy of the
germinal disc under in ovo conditions by maintaining high
hatching rates [33].
Summary
The application of infrared spectroscopic imaging to
determine the gender of fertilized but unincubated eggs is
a novel approach. In this technique, the difference in DNA
between the Z and W sex chromosomes is a very sensitive
marker for gender. Cellular DNA content is approx. 2%
greater for males (ZZ) than for females (ZW). The results
obtained in this study show that FT-IR spectroscopic
imaging in conjunction with a supervised classification is
a powerful tool to determine the gender of fertilized
unincubated eggs. The method has also the potential for
in ovo measurements without removing blastoderm cells.
Using a needle ATR crystal, it should be possible to
determine the gender of the egg without removing
blastoderm cells. In this case, the tip of the ATR crystal is
placed onto the top of the germinal disc.
Acknowledgments This work was financially supported by the
German Federal Ministry of Food, Agriculture, and Consumer Protection
(BMELV) through the Federal Office for Agriculture and Food (BLE),
grant no. 511-06.01-28-1-33.010-07, and the Ministry of Environment,
Energy, Agriculture and Consumer Protection of Hessen. The authors
gratefully acknowledge the Lohmann Tierzucht GmbH (Cuxhaven,
Germany) for their financial support and for providing eggs. Also,
special thanks to Mrs. Andrea Büchner (Institute of Physiological
Chemistry, Faculty of Veterinary Medicine, University of Leipzig) for
her skillful technical assistance and to Mrs. Dr. Anke Förster (Lohmann
Tierzucht GmbH) for the insightful discussions. Last but not least, the
technical support of Bruker Optik GmbH (Leipzig, Germany) is also
acknowledged.
References
1. Klein S, Flock D, Ellendorff F (2003) World’s Poult Sci J 59:60–62
2. Seemann G (2003) World’s Poult Sci J 59:59–61
3. Flock DK, Preisinger R (1996) Dtsch Geflügelwirt Schweineprod
46:22–26
4. Phelps P, Bhutada A, Bryan S, Chalker A, Ferrell B, Neuman S,
Ricks C, Tran H, Butt T (2003) World’s Poult Sci J 64:32–36
5. Aerts S, Boonen R, Bruggeman V, De Tavernier J, Decuypere E
(2009) Culling of day-old chicks: opening the debates of Moria?
In: Millar K, Hobson West P, Nerlich B (eds) Ethical futures:
bioscience and food horizons. Wageningen Academic Publishers,
Wageningen, pp 117–122
6. Preisinger R (2003) World’s Poult Sci J 59:52–56
7. Turkyilmaz MK, Karagenc L, Fidan E (2001) Br Poult Sci
51:525–529
8. Kaleta EF, Redmann T (2008) World’s Poult Sci J 64:391–399
9. Ellendorff F, Klein S (2003) World’s Poult Sci J 59:5–6
10. Crawford RD (ed) (1990) Poultry breeding and genetics. Elsevier,
Amsterdam
11. Clinton M, Haines L, Belloir B, McBride D (2001) Sci Br Poult
42:134–138
12. Arnold KE, Orr KJ, Griffiths R (2003) Mol Ecol 12:3451–3458
13. Imholt D (2010) Morphometrische Studien an Eiern von Hybrid-
Rassehühnern mit Versuchen zur Detekion einer Beziehung
zwischen der Form von Eiern und dem Geschlecht der darin
befindlichen Küken. Eine oologische und mathematische Studie,
Disseration, Univ. Giessen, Germany
14. Steiner G, Bartels T, Krautwald-Junghans ME, Boos A, Koch E
(2010) Anal Bioanal Chem 396:465–470
15. Mendonça MAC, Carvalho CR, Clarindo WR (2001) J Hist Cyto
58:229–235
16. Tiersch TR (2003) World’s Poult Sci J 64:24–30
17. Klein S, Baulain U, Rokitta M, Marx G, Thielebein J, Ellendorff F
(2003) World’s Poult Sci J 64:38–44
18. Rudbeck L, Dissing J (1998) Biotechniques 25:588–592
19. Fridolfsson AK, Ellegren H (1999) J Avian Biol 30:116–121
20. Don RH, Cox PT, Wainwright BJ, Baker K, Mattick JS (1991)
Nucleic Acids Res 19:4008
21. Socrates G (2001) Infrared and Raman characteristic group
frequencies. Wiley, West Sussex
22. Fabian H, Jackson M, Murphy L, Watson PH, Fichtner I, Mantsch
HH (1995) Biospec 1:37–45
23. Richter T, Steiner G, Abu-Id MH, Salzer R, Bergmann R, Rodig
H, Johannsen B (2002) Vib Spec 28:103–110
24. Liu KZ, Man A, Dembinski TC, Shaw RA (2007) Anal Bioanal
Chem 387:1809–1814
25. Chirboga L, Yee H, Diem M (2000) Appl Spec 54:1–8
26. Salzer R, Steiner G, Mantsch HH, Mansfield J, Lewis NE (2000)
Fresenius J Anal Chem 366:712–726
27. Wang L, Mizaikoff B (2008) Anal Bioanal Chem 391:1641–1654
28. Steiner G, Küchler S, Hermann A, Koch E, Salzer R, Schackert G,
Kirsch M (2008) Cytometry 37A:1158–1164
29. Dovbeshko GI, Gridina NY, Kruglova EB, Pashchuk OP (1997)
Talanta 53:233–246
30. Yang Y, Sule-Suso J, Sockalingum GD, Kegelaer G, Manfait M,
El Haj AJ (2005) Biopolymer 78:311–317
31. Chiriboga L, Xie P, Yee H, Vigorita V, Zarou D, Zakim D, Diem
M (1998) Biospec 4:47–53
32. Mantsch HH, Chapman D (eds) (1996) Infrared spectroscopy of
biomolecules. Wiley, New York
33. Steiner G, Bartels T, Koch E, Krautwald-Junghans ME (2011)
Verfahren und Vorrichtung zur Bestimmung des Geschlechtes von
befruchteten und nicht bebrüteten Vogeleiern [Approaches and
devices for the determination of the gender of fertilized and
unincubated avain eggs]. Deutsches Patent-und Markenamt, DE 10
2010 006 161 B3, 13.01.2011
2782 G. Steiner et al.
