Discrimination between Cave and House-Farmed Edible Bird's Nest Based on Major Mineral Profiles

Abstract

The high priced cave edible bird's nest (EBN) has attracted unscrupulous EBN producers to adulterate EBN with lower priced house-farmed EBN due to the fact that both are almost indistinguishable by visual inspection. In the present study, major mineral contents such as calcium, sodium, magnesium and potassium of both EBN types were analysed using inductively coupled plasma-optical emission spectrometry (ICP-OES). Three pattern recognition techniques namely hierarchical cluster analysis (HCA), principal component analysis (PCA) and linear discriminant analysis (LDA) were employed to determine the influence of harvesting origins on mineral profiles. With the use of HCA and PCA, EBN samples have successfully been grouped into two distinct clusters. From the PCA score plot, principal component 1 (49.53 %) and principal component 2 (41.11%) accounted for 90.64% of the total variability. In addition, LDA presented excellent performance in discriminating and predicting membership of the two EBN sample types with classification rate of 100%.

Full text

Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 195 (2016)
ISSN: 1511-3701 © Universiti Putra Malaysia Press
TROPICAL AGRICULTURAL SCIENCE
Journal homepage: http://www.pertanika.upm.edu.my/
Article history:
Received: 29 January 2015
Accepted: 26 January 2016
ARTICLE INFO
E-mail addresses:
ekseow@hotmail.com (Seow, E. K.),
baharudin.ibrahim@usm.my (Ibrahim, B.),
syahidah.muhammad@usm.my (Muhammad, S. A.),
lamhong.lee@qiup.edu.my (Lee, L. H.),
japareng@usm.my (Lalung, J.),
lhcheng@usm.my (Cheng, L. H.)
* Corresponding author
Discrimination between Cave and House-Farmed Edible Bird’s
Nest Based on Major Mineral Proles
Seow, E. K.1, Ibrahim, B.2, Muhammad, S. A.3,4, Lee, L. H.5, Lalung, J.3 and
Cheng, L. H.1*
1Food Technology Division, School of Industrial Technology, Universiti Sains Malaysia, 11800 USM,
Penang, Malaysia
2Discipline of Clinical Pharmacy, School of Pharmaceutical Sciences, Universiti Sains Malaysia,
11800 USM, Penang, Malaysia
3Environmental Technology Division, School of Industrial Technology, Universiti Sains Malaysia,
11800 USM, Penang, Malaysia
4Doping Control Centre, Universiti Sains Malaysia, 11800 USM, Penang, Malaysia
5Faculty of Integrative Science & Technology, Quest International University Perak, No. 227,
Plaza Teh Teng Seng (Level 2), Jalan Raja Permaisuri Bainon, 30250 Ipoh, Perak, Malaysia
ABSTRACT
The high priced cave edible bird’s nest (EBN) has attracted unscrupulous EBN producers
to adulterate EBN with lower priced house-farmed EBN due to the fact that both are almost
indistinguishable by visual inspection. In the present study, major mineral contents such
as calcium, sodium, magnesium and potassium of both EBN types were analysed using
inductively coupled plasma-optical emission spectrometry (ICP-OES). Three pattern
recognition techniques namely hierarchical cluster analysis (HCA), principal component
analysis (PCA) and linear discriminant analysis (LDA) were employed to determine the
inuence of harvesting origins on mineral proles. With the use of HCA and PCA, EBN
samples have successfully been grouped into two distinct clusters. From the PCA score plot,
principal component 1 (49.53 %) and principal component 2 (41.11%) accounted for 90.64%
of the total variability. In addition, LDA presented excellent performance in discriminating
and predicting membership of the two EBN
sample types with classification rate of
100%.
Keywords: Edible bird’s nest, hierarchical cluster
analysis, inductively coupled plasma-optical emission
spectrometry, linear discriminant analysis, mineral
content, principal component analysis

Seow, E. K., Ibrahim, B., Muhammad, S. A., Lee, L. H., Lalung, J. and Cheng, L. H.
182 Pertanika J. Trop. Agric. Sci. 39 (2) 181 - 196 (2016)
INTRODUCTION
Edible bird’s nest (EBN) is highly consumed
by the Chinese community because they
uphold the belief handed down based on
anecdotal evidences that EBN is benecial
to relief respiratory ailments and enhance
body energy. The work by Kong et al.
(1987), which suggests the presence of
epidermal growth factor (EGF)-like
substance in EBN, has drawn the attention
of consumers as well as researchers. Since
then, extensive research activities have
been conducted to conrm the presence of
EGF-like substance in EBN and its potential
use in medical eld and cosmetic industry
for cell proliferative effect. This idea was
substantiated by positive results reported in
studies using human adipose-derived stem
cells (Roh et al., 2012), corneal keratocytes
(Zainal Abidin et al., 2011) and Caco-2 cells
(Aswir & Wan Nazaimoon, 2010). Apart
from that, EBN extract has been found
effective in curing erectile dysfunction (Ma
et al., 2012), improving bone strength and
dermal thickness (Matsukawa et al., 2011)
and inhibiting inuenza virus infection (Guo
et al., 2006).
Generally, EBN is built by gelatinous
strand of nest cement secreted by swiftlets,
namely, White nest swiftlet (Aerodramus
fuchipagus) and Black nest swiftlet
(Aerodramus maximus) during breeding
seasons (Koon & Cranbrook, 2002). These
swiftlets are found in the South-East Asia
region and inherently inhabit the caves
(Chantler & Driessen, 1999). Comparatively,
EBN produced by the White nest swiftlet is
of higher economic value as it is entirely
made of pure salivary nest cement with
only traces of impurities. On the other hand,
though the nest of Black nest swiftlet is full
with feathers and requires tedious cleaning
process, it is still heavily harvested as the
exploitation is worthwhile due to the fact
that the price of the nest is extremely high.
With the increasing demand for EBN,
the price of this product is expected to
increase as the stock available in the market
could not full the growing needs. A recent
survey reported by Manan and Othman
(2012) revealed that the raw pre-processed
EBN was sold at RM 3000/kg to RM 4500/
kg in the market in year 2010 to 2011. The
market price of EBN is always doubled
after the laborious and time consuming
cleaning process (Lim, 2006). Therefore,
many investors are lured by the lucrative
revenue and venture into EBN house-
farming. Efforts have been done by the
house farmers to ensure that only the pure
breed of White nest swiftlet, which could
produce EBN of high commercial value,
would inhabit and breed in the farm (Lim,
2006). Unfortunately, EBN harvested from
the house farm is much lower priced in the
market than those harvested from the cave.
Driven by the unscrupulous desire,
unethical EBN manufacturers tend to
adulterate cave EBN with lower priced
house EBN; some even make intentional
false claims by selling house nest as cave
nest. Besides, adulteration of EBN with
addition or substitution with less expensive
materials such as egg white, Tremella
fungus, gelatin, karaya gum, fried porcine
skin, starch, soybean and red seaweed

Major Minerals Composition Data
183Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 196 (2016)
(Marcone, 2005; Ma & Liu, 2012), is
commonplace.
Authentication methods at molecular
level using Taqman-based real-time
PCR (Guo et al., 2014), combination of
DNA based PCR and protein based two
dimensional gel electrophoresis methods
(Wu et al., 2010), DNA sequencing-based
method (Lin et al., 2009) and SDS-PAGE
electrophoresis (Marcone, 2005) have been
proposed. However, these techniques are
rather tedious, time-consuming and costly.
EBN was built by swiftlets inhabiting
in the caves and house farms and it was
hypothesised that the minerals prole of
EBN would be affected by the environments,
as well as the supporting materials it
attached to. The objective of this study
is to distinguish EBN samples harvested
from the cave and the house farm based
on simple minerals prole analysed using
inductively coupled plasma-optical emission
spectrometry (ICP-OES). Correlation of
mineral pairs within each group of sample
was analysed using Pearson correlation
analysis and pattern recognition techniques,
namely, hierarchical cluster analysis (HCA),
principal component analysis (PCA) and
linear discriminant analysis (LDA) were
employed to investigate the relationship
between elemental concentration and the
type of EBN samples studied.
MATERIALS AND METHODS
Materials
In this study, forty eight EBN samples were
analysed. Twenty four of these were house
nests harvested from different locations in
West Malaysia, namely, Alor Setar, Bukit
Mertajam, Kota Bharu, Segamat, Taiping
and Teluk Intan. The twenty four cave nests
were harvested from the caves located in
East Malaysia (Bau and Sandakan) and
Indonesia (Aceh and Medan). All EBN
samples used in this study were raw genuine
samples collected from different locations
(see Figure 1) with the assistance of reliable
suppliers and sponsors. All pre-processed
samples were cleaned and air-dried under
the same process. EBN samples were soaked
in water and the feathers and impurities were
removed using tweezers until the nests were
devoid of visible feathers and impurities
and followed by air-drying. Then, cleaned
nests were dipped into liquid nitrogen for 10
seconds prior to grinding them into powder
form. The samples were kept in air-tight
bottles and stored at room temperature until
further analysis.
Moisture Content
Moisture content of the samples was
determined by volumetric Karl Fischer
titration (784 KFP Titrino, Metrohm,
Switzerland) following AOAC Official
Method 2001.12.
Elemental Analysis
About 0.25 g of EBN powder was digested
in a mixture of 3 mL H2O + 2 mL HNO3
+ 1 mL H2O2 with a microwave digester
(MARSXpress, CEM Corporation,
Matthews, NC), following the method
described in Saengkrajang et al. (2013).
The digestion was carried out at 220ºC for
45 minutes until a clear transparent solution

Seow, E. K., Ibrahim, B., Muhammad, S. A., Lee, L. H., Lalung, J. and Cheng, L. H.
184 Pertanika J. Trop. Agric. Sci. 39 (2) 181 - 196 (2016)
was obtained. The digest was then made up
to 50 mL with 2% HNO3 solution and kept
chilled in plastic bottles prior to mineral
determination.
The concentrations of sodium
(Na), potassium (K), calcium (Ca) and
magnesium (Mg) were determined by
inductively coupled plasma-optical
emission spectrometry (ICP-OES), Perkin
Elmer optima 7000DV equipped with
S10 autosampler and WinLab32TM for ICP
V5.1 (Perkin Elmer, Waltham, MA). The
calibration was performed with standard
mixture from Perkin Elmer (Waltham MA)
and all elements were determined at axial
plasma view. The instrumental settings of
the ICP-OES were as follows: the source
equilibration delay was 15 seconds, plasma
parameters were set at plasma 15 L/min,
auxiliary 0.2 L/min, nebulizer 0.8 L/min
and power 1300 W. Flow rate of sample was
1.5 mL/min with Argon as carrier gas. There
was a washing step between the samples at
the rate of 1.5 mL/min for 30 seconds. The
wavelengths for each element were: Ca,
317.933 nm; Na, 589.592; Mg, 285.213 and
K, 766.490.
Method Verication
The raw data were pre-processed and
the concentration of each element was
expressed in unit of mg/100 g dry matter
basis to minimise data fluctuation.
Calibration curves for Ca, Na, Mg and K
were constructed using external standards
method. Coefcient of determination, r2 of
calibration curves for the elements were all
above 0.9900. Repeatability was determined
by intra- and inter-day variation studies,
while reproducibility was determined by
two different analysts that conduct the
same method. This method showed a
very good precision in repeatability and
reproducibility, with relative standard
deviation (RSD) of elements determined
ranged from 0.80 to 5.69%.
Statistical Analysis
Experimental data obtained were analysed
using the statistical package SPSS version
22 for Windows (SPSS Inc., Chicago, IL).
Independent samples t-test was conducted
to determine signicant difference between
mean values. Pearson correlation analysis
was used to study the direction (positive/
negative) and strength (weak/moderate/
strong) of the correlation between elements
within each type of nest samples. Three
pattern recognition techniques: hierarchical
cluster analysis (HCA), principal component
analysis (PCA) and linear discriminant
analysis (LDA) were used to observe the
possible pattern and trend in classication.
RESULTS AND DISCUSSION
Elemental Composition of the EBN
Samples
Calcium (Ca), sodium (Na), magnesium
(Mg) and potassium (K) composition of both
house EBN and cave EBN from different
locations and descriptive statistics of both
types of EBN are tabulated in Tables 1 and
2, respectively. Based on the independent
samples t-test result, it is evident that
Ca content in cave EBN is significantly
higher than house EBN but the Mg and

Major Minerals Composition Data
185Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 196 (2016)
Na contents are signicantly lower in cave
EBN. Nonetheless, there is no signicant
difference observed in the K content in both
types of EBN.
Since K is not signicantly different
for the two types of samples, mineral
composition could better or more accurately
be compared by its ratio after being
normalised to K content. Generally, the
average major minerals contents determined
in this study were arranged in the decreasing
order of Ca > Na > Mg > K, which is in
accordance with the research ndings of
Norhayati et al. (2010). For cave samples,
the ratio of Ca:Na:Mg: K is 101:13:6:1,
whereas for the house samples the ratio is
46:33:8:1. Obviously, calcium content in
the cave EBN samples was slightly more
than double of those found in the house
EBN samples, and the reverse is true for
Na content. The discrepancy in the element
contents of both samples could largely
be contributed by the inherent different
environmental conditions prevailing in the
cave and in the house farm (Sia & Tan,
2014).
Cave EBN is normally found as self-
supporting nests that attached to vertical or
Table 1
Major minerals prole of house nests and cave nests.
Location Type Sample size Ca Mg Na K
Alor Setar House Nest 5706±32 122±11 632±78 18±2
Bukit Mertajam House Nest 5780±62 123±6 625±66 12±3
Kota Bharu House Nest 5665±51 127±8 633±100 14±1
Segamat House Nest 3777±98 138±11 548±150 20±1
Teluk Intan House Nest 3787±13 130±7 358±13 18±0.4
Taiping House Nest 3750±19 112±7 228±22 18±1
Medan Cave Nest 41741±314 93±30 94±50 8±2
Aceh Cave Nest 41389±334 102±19 112±69 13±4
Sandakan Cave Nest 82263±207 130±23 294±166 33±10
Bau Cave Nest 81203±42 90±3 274±24 6±1
Values are mean±standard deviation reported in mg/100g dry matter.
Table 2
Descriptive statistics for house and cave edible bird’s nests.
Minerals content (mg/100g dry matter)
Element House nests (n=24) Cave nests (n=24)
Min. Max. Mean STDEV Min. Max. Mean STDEV
Ca* 586 891 737 67 1141 2542 1677 504
Na* 203 795 536 167 38 630 224 131
Mg* 106 149 125 10 60 159 106 25
K 8 21 16 3 5 51 17 14
*Mean values are signicantly different (P<0.05).

Seow, E. K., Ibrahim, B., Muhammad, S. A., Lee, L. H., Lalung, J. and Cheng, L. H.
186 Pertanika J. Trop. Agric. Sci. 39 (2) 181 - 196 (2016)
concave surface of a cave wall. Therefore,
it is easy to rationalise high Ca content
found in cave EBN. According to Northup
and Lavoie (2010), mineral dissolution
and precipitation processes in caves are
microbially mediated reactions. Cave
dissolution process involves iron-, sulfur-
and manganese- oxidising bacteria,
through which activities considerable
acidity is being generated and subsequently
used to dissolve cave wall that is rich in
calcium carbonate. Meanwhile, the mineral
precipitation process was reported to be
either passive where microbial cells acts as
nucleation sites or active, where bacterially
produced enzymes control mineralisation.
In passive mineralisation, dissolved metal
(Ca2+) was found to sorb onto amphoteric
functional groups (such as carboxyl,
phosphoryl and amino constituents) found
on negatively charged cell walls, sheaths
or capsules, following which carbonate
(HCO3-) precipitates and in turn serves
as nucleation site for calcium carbonate
precipitation (Lowenstem & Weiner, 1989;
Konhouser, 1997, 1998). It is believed that
similar mechanism could have occurred by
mineralisation on salivary strands (which is
high in proteins) of a cave EBN.
On the other hand, Na content in the
house EBN was found to be signicantly
higher than those cave EBN samples (Table
2). Interestingly, Na was also reported to
be the predominant element in processed
house-farmed EBN harvested from different
locations in Thailand (Saengkrajang et al.,
2013), pre-processed house-farmed EBN
in Penang, Malaysia and pre-processed
cave EBN in Sumatra Indonesia (Nurul
Huda et al., 2008). Our raw data showed
that Na content was extremely high in EBN
harvested from Alor Setar, Bukit Mertajam
and Kota Bharu house farms (Table 1 &
Figure 1), which are located at the coastal
locations facing the Malacca Straits. The Na
content recorded was 2-3 folds higher when
compared to the other samples harvested
from other locations. Based on the report of
Norhayati et al. (2013), this high Na content
could be attributed to the accumulation of Na
from marine aerosols through atmospheric
deposition into the EBN. It is believed that
sea salt concentration in the air could be
high at these locations as a result of the
persistent on-shore winds which generate
sea water droplets and marine aerosols (sea
sprays). The speculation was made based on
the unique drinking behaviour of swiftlets,
which capture the water droplets in the air.
Therefore, the Na content in marine aerosol
(swiftlet’s saliva) is assumed to contribute
to the nest Na content.
Besides the environmental factor,
swiftlet diets could contribute partly to
the difference in elemental prole of both
types of samples. According to Lourie
and Tompkins (2000), swiftlet’s diets vary
and are very much dependent on their
foraging regions and food availability. Apart
from this, White nest swiftlet’s diet was
discovered to be diverse and this species was
predicted to survive and adapt well in urban
areas (shop lot house farms). This could be
a factor that yields the different minerals
composition patterns in EBN harvested from
different origins.

Major Minerals Composition Data
187Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 196 (2016)
Figure 1. Edible bird’s nest sampling points.
Pearson Correlation Analysis
The correlation matrix between mineral
pairs of both types of EBN is presented
in Table 3. Ca was found to demonstrate
moderate positive correlation with Mg and
it was signicantly different at r = 0.450
(P < 0.05) for cave EBN samples. The
Na content correlated signicantly with K
content at moderate values with r = -0.477
and 0.505 (P < 0.05) for house EBN and
cave EBN, respectively. Interestingly, there
were strong positive correlations between
the mineral pairs in the cave EBN samples
such as Ca and K (r = 0.776, P < 0.01), Na
and Mg (r = 0.609, P < 0.01) and Mg and
K (r = 0.832, P < 0.01). The significant
relationship between the minerals leads us to
further analyse the inuence of macro- and
micro-environmental factors on minerals
composition of EBN.
Hierarchical Cluster Analysis (HCA)
Hierarchical cluster analysis (HCA) is
an unsupervised classification method
that discerns objects into groups based
on the level of similarity between them
based on the relative contribution of the
variables. The clustering method used
was the nearest neighbour (single linkage)
method, measured based on squared
Euclidean distance. A dendrogram was an
easy visualisation aid produced with the
samples of the same similarity level being
grouped together. The use of HCA has
successfully assigned the EBN samples
into two main clusters, i.e. house EBN (n
= 24) and cave EBN (n = 24), based on
the dendrogram cut at a distance of 17.5 as
presented in Figure 2. All the EBN samples
were accurately classied into their own
clusters which indicated that the elemental

Seow, E. K., Ibrahim, B., Muhammad, S. A., Lee, L. H., Lalung, J. and Cheng, L. H.
188 Pertanika J. Trop. Agric. Sci. 39 (2) 181 - 196 (2016)
composition could be appropriately used in
classication of the type of EBN sample.
Principal Component Analysis (PCA)
Principal component analysis (PCA) is
a chemometric tool used for dimension
reduction of data set through which the
most signicant and important data would
be extracted for further analysis (Abdi &
William, 2010). Basically, PCA demonstrates
primary evaluation and visualisation of
between-class similarity based on the
contributing variables variation direction in
a multivariate space. PCA was carried out
on a data matrix consists of EBN elemental
profiles. The principal component (PC)
scores and possible clustering results are
illustrated in Figure 3A. Only two PC were
extracted from the dataset to explain the
total variability up to 90.64%. Two clusters
Figure 2. Dendogram of hierarchical cluster analysis. Cluster 1: house nests; cluster 2: cave nests
Table 3
Pearson correlation of minerals content in house and cave edible bird’s nest.
House nests Cave nests
Element Ca Na Mg KElement Ca Na Mg K
Ca 1Ca 1
Na -0.069 1Na 0.151 1
Mg 0.338 0.275 1Mg 0.450*0.609** 1
K0.077 -0.477*0.208 1 K 0.776** 0.505*0.832** 1
** and * correspond to signicance of correlation at the 0.01 level and 0.05 level (2-tailed), respectively.

Major Minerals Composition Data
189Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 196 (2016)
Figure 3. Principal component analysis was applied to study possible clustering between (A) house nest and
cave nest, and their respective inuential variables loaded as shown in (B). A simple 3-D plot of Ca vs. Na vs.
Mg as illustrated in (C) gives a simple view of clustering potential
were identied and separated successfully
at the diagonal by PC1 (explaining 49.53%
of the variability) and PC2 (explaining
41.11%). Mg and Na, and Ca and K were
the highly loading variables in PC1 and
PC2, respectively, as shown in Figure 3B.
In particular, the loading scores for Mg,
Na, Ca and K were 0.941, 0.800, 0.911
and 0.751, respectively. A simple 3D-plot
of Ca vs. Na vs. Mg concentrations was
constructed to give a simple view of sample
distribution or clustering potential as shown
in Figure 3C, as these three variables
contribute most of the variance. This
3D-plot is in good agreement with the PCA
result (Figure 3A) that it provides a good
discrimination pattern whereby house nest
and cave nest are separated. As illustrated
in Figure 3A, house EBN was observed
to distribute more closely as compared
to cave EBN. Geographical origin with
different environmental conditions could
be the key factor that contributes to the
differences in EBN collected from different
locations. Recently, authenticity assessment
of commodities such as cabbages (Bong et

Seow, E. K., Ibrahim, B., Muhammad, S. A., Lee, L. H., Lalung, J. and Cheng, L. H.
190 Pertanika J. Trop. Agric. Sci. 39 (2) 181 - 196 (2016)
al., 2013), Croatian wines (Kruzlicova et
al., 2013), Spanish cherries (Matos-Reyes
et al., 2013) and Brazilian honey (Batista
et al., 2012) through determination of
mineral proles analysed by chemometric
analysis was found to be a valuable tool
in classication according to geographical
origins. Hence, PCA was further applied to
investigate the possible groupings within
class for both house and cave EBN.
As shown in Figure 4A, house EBN
samples collected from the northern region
(Alor Setar, Bukit Mertajam and Kota
Bharu), whereas the remaining samples
obtained from the central (Taiping and Teluk
Intan) and the southern region (Segamat)
in Peninsular Malaysia were separated
into two clusters by PC1 which accounts
for only 37.54% of the total variability.
Na and K were the variables highly loaded
in PC1, with the loading scores of -0.776
and 0.863, respectively, as illustrated in
Figure 4B. Likewise, PC1 with the highly
loading K, Mg and Ca variables, which
explained 67.82% of variability (Figure 4C)
categorised Sandakan cave EBN samples
as one cluster and the other samples from
Aceh, Bau and Medan as another cluster.
As shown in Figure 4D, K, Mg and Ca were
positively loaded in PC1 with the scores
at 0.963, 0.897 and 0.731, respectively.
The minerals prole of the cave EBN is
associated to the cave wall the nest adheres
to. The mineral prole of Sandakan cave
EBN, which differs from the other locations
may be attributed by the unique materials of
the cave wall in Sandakan. This is evidenced
by the geological survey that Sandakan
rocks, which consist predominantly of
mudstone, sandstone and siltstone with
minor coal seams and conglomerate (Lee,
1970), are different in composition from
the caves in Bau which are composed
of fossiliferous limestone (Wolfenden,
1965). The lithological variations between
different locations are due to facies changes
(Lee, 1970). The results are in good
agreement with the ndings discovered by
Saengkrajang et al. (2013), Norhayati et al.
(2010) and Nurul Huda et al. (2008) that
nutritional composition of EBN could be
distinguishable by breeding sites. However,
it could be observed that even for the
samples collected from the same location
and within the same breeding season, the
distributions were scattered. Hence, other
contributing factors should be taken into
considerations for future research studies.
A better distinct separation between the
samples harvested from different locations
could be achieved by increasing the sample
size.
Linear Discriminant Analysis (LDA)
Linear discriminant analysis (LDA) is a
supervised pattern recognition approach
which separates classes based on their
dissimilarities by maximising the variance
between classes and minimising the variance
within classes (Roggo et al., 2003; Liu et
al., 2012). A stepwise method was used
to investigate if cave and house EBN
could be differentiated by their elemental
composition. Cross-validation procedure
was carried out by employing the leave-
one-out technique to evaluate the robustness

Major Minerals Composition Data
191Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 196 (2016)
of the classication model. Each sample
was classied based on the discriminant
functions generated from the remaining
samples and the accuracy of the classication
was calculated as rate of cross-validation
(Lachenbruch, 2006). LDA is used to assess
the EBN samples with respect to the type
based on the elemental composition. Four
major elements (Ca, Na, Mg and K) were
evaluated through LDA and only one linear
discriminant function (DF) responsible in
elucidating the differences between cave
and house-farmed EBNs was derived. This
DF explained 100% of the total variability
between two types of EBN and the relative
contribution of each parameter identied is
as depicted in Eq. (1).
Z = 0.616 Ca - 0.489 Na - 0.290 Mg
– 0.012K [1]
Ca and Na exhibited strong contribution
in discriminating cave EBN from house
EBN, whereas Mg showed relatively lower
contribution in explaining the variation
between the cave EBN and house EBN.
Scores of DF for EBN samples of different
types correspond to the behaviour of the
parameters in the DF as depicted in Figure 5.
Figure 4. Principal component analysis was applied to study sample distribution within (A) house nest and (C)
cave nest for geographical origins, and their respective inuential variables loaded as shown in (B) and (D)

Seow, E. K., Ibrahim, B., Muhammad, S. A., Lee, L. H., Lalung, J. and Cheng, L. H.
192 Pertanika J. Trop. Agric. Sci. 39 (2) 181 - 196 (2016)
Overall, all the house EBN samples showed
negative contribution to the DF, whereas all
the cave EBN samples demonstrated positive
contribution. The sources of the samples
were correctly identied in accordance to
their own origins.
The membership of the EBN samples
was predicted by employing a stepwise
discriminant procedure, as shown in Table
4. A leave-one-out cross validation method
was used to evaluate the robustness of this
prediction model. Both overall classication
and cross-validation classification were
100%, which implied that all samples,
were correctly assigned to their own
cluster. The results exhibited that this
classication model was a very promising
tool in discrimination of the EBN samples
according to the types.
CONCLUSION
The use of elemental composition determined
by inductively coupled plasma-optical
emission spectrometry (ICP-OES) combined
with chemometric approach is veried to be
a powerful tool in discriminating edible
bird’s nest based on types. Ca and Na were
the elements which demonstrated strong
contributions in the differentiation of two
types of EBN samples. The robustness of
the classication model has been validated
and found to possess great predicting power
at a classication rate and cross-validation
rate of 100%.
ACKNOWLEDGEMENTS
This work was funded by Universiti Sains
Malaysia Short Term Research Grant
(Grant No: 304.PTEKIND.6313033). Seow
acknowledges the Fellowship awarded by
Universiti Sains Malaysia. The authors
would like to thank Mr. George Ng Aun
Figure 5. Scores for the discriminant function

Major Minerals Composition Data
193Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 196 (2016)
Heng, Dato Feasa, Mr. S.D. Hng, Mr. L.C.
Ling, Mr. Chaw Seow Peoh, Mr. Sia Meu
Seng, Mr. Tan Yoke Tian, Mr. Thomas Lee,
Mr. Tan Sooi Huat and Mr. Peter Lau for
sponsoring edible bird’s nest samples to
the project. The authors have declared no
conict of interest.
REFERENCES
Abdi, H., & William, L. J. (2010). Principal component
analysis. Wiley Interdisciplinary Reviews:
Computational Statistics, 2(4), 433-459.
Association of Ofcial Analytical Chemists (AOAC).
(2005). AOAC Ofcial Methods of Analysis (18th
ed.). Washington DC: Association of Ofcial
Analytical Chemists.
Aswir, A. R., & Wan Nazaimoon, W. M. (2010). Effect
of edible bird’s nest on Caco-2 cell proliferation.
Journal of Food Technology, 8(3), 126-130.
Batista, B. L., Da Silva, L. R. S., Rocha, B. A.,
Rodrigues, J. L., Berretta-Silva, A. A., Bonates,
T. O., ... & Barbosa, F. (2012). Multi-element
determination in Brazilian honey samples by
inductively coupled plasma mass spectrometry
and estimation of geographic origin with data
mining techniques. Food Research International,
49(1), 209-215.
Bong, Y. S., Song, B. Y., Gautam, M. K., Jang, C. S.,
An, H. J., & Lee, K. S. (2013). Discrimination of
the geographic origin of cabbages. Food Control,
30(2), 626-630.
Chantler, P., & Driessens, G. (1999). Swifts, a guide to
the swifts and treeswifts of the world. (2nd ed.).
UK: Pica Press.
Guo, C. T., Takahashi, T., Bukawa, W., Takahashi,
N., Yagi, H., Kato, K., ... & Suzuki, Y. (2006).
Edible bird’s nest extract inhibits inuenza virus
infection. Antiviral Research, 70(3), 140-146.
Guo, L., Wu, Y., Liu, M., Wang, B., Ge, Y., & Chen,
Y. (2014). Authentication of edible’s bird nest
by TaqMan-based real-time PCR. Food Control,
44, 220-226.
Kong, Y. C., Keung, W. M., Yip, T. T., Ko, K. M.,
Tsao, S. W. & Ng, M. H. (1987). Evidence that
epidermal growth factor is present in swiftlet’s
(Collocalia) nest. Comparative Biochemistry and
Physiology, 87(2), 221-226.
Konhauser, K. O. (1997). Bacterial iron mineralization
in nature. FEMS Microbiology Reviews, 20(3-4),
315-326.
Table 4
Classication of edible bird’s nest samples and percentage of classication according to types through
cross-validation method.
Type Predicted Group Membership Total
House nest Cave nest
Original Count House nest 24 024
Cave nest 024 24
%100 100 100b
Cross-validatedaCount House nest 24 024
Cave nest 024 24
% 100 100 100c
aCross validation is done only for those cases in the analysis. In cross validation, each case is classied
by the functions derived from all cases other than that case.
b100% of original grouped cases correctly classied.
c100% of cross-validated grouped cases correctly classied.

Seow, E. K., Ibrahim, B., Muhammad, S. A., Lee, L. H., Lalung, J. and Cheng, L. H.
194 Pertanika J. Trop. Agric. Sci. 39 (2) 181 - 196 (2016)
Konhauser, K. O. (1998). Diversity of bacterial iron
mineralization. Earth-Science Reviews, 43(3),
91-121.
Koon, L. C., & Cranbrook, E. (2002). Swiftlets of
Borneo: Builders of Edible Nest. Malaysia:
Natural History Publications (Borneo) (Chapter
1).
Kruzlicova, D., Fiket, Ž., & Kniewald, G. (2013).
Classication of Croatian wine varieties using
multivariate analysis of data obtained by high
resolution ICP-MS analysis. Food Research
International, 54(1), 621-626.
Lachenbruch, P. A. (2006). Discriminant analysis.
Encyclopedia of Statistical Sciences (p.3). New
Jersey: John Wiley and Sons.
Lee, D. T. C. (1970). Sandakan peninsula, eastern
Sabah East Malaysia. Malaysia: Government
Printing Ofce, Kuching, Sarawak, (Chapter III).
Lim, C. (2006). Make millions from swiftlet farming:
A denite guide. Malaysia: Truewealth (Chapters
1 & 6).
Lin, J. R., Zhou, H., Lai, X. P., Hou, Y., Xian, X. M.,
Chen, J. N., Wang, P. X., Zhou, L., & Dong, Y.
(2009). Genetic identication of edible bird’s
nest based on mitochondrial DNA sequences.
Food Research International, 42(8), 1053-1061.
Liu, X., Xue, C., Wang, Y., Li, Z., Xue, Y., & Xu,
J. (2012). The classication of sea cucumber
(Apostichopus japonicas) according to region of
origin using multi-element analysis and pattern
recognition techniques. Food Control, 23(2),
522-527.
Lourie, S. A., & Tompkins, D. M. (2000). The diets
of Malaysian swiftlets. Ibis, 142(4), 596-602.
Lowenstam, H. S. & Weiner, S. (1989). O n
biomineralization. (p.324). New York: Oxford
University Press.
Ma, F., & Liu, D. (2012). Sketch of the edible bird’s
nest and its important bioactivities. Food
Research International, 48(2), 559-567.
Ma, F. C., Liu, D. C., & Dai, M. X. (2012). The effects
of the edible bird’s nest on sexual function of
male castrated rats. African Journal of Pharmacy
and Pharmacology, 6(41), 2875-2879.
Malaysian Food Act 1983 (Act 281) and Regulations.
(2014). Malaysia: International Law Book
Services (Twelfth Schedule).
Manan, Z., & Othman, M. S. H. (2012). An economic
analysis of swiftlet edible bird’s nest in industry.
Malaysia: Forestry Department Peninsular
Malaysia (Chapter 4).
Marcone, M. F. (2005). Characterization of the
edible bird’s nest the “Caviar of the East”. Food
Research International, 38(10), 1125-1134.
Matos-Reyes, M. N., Simonot, J., López-Salazar,
O., Cervera, M. L., & de la Guardia, M. (2013).
Authentication of Alicante’s Mountain cherries
protected designation of origin by their mineral
prole. Food Chemistry, 141(3), 2191-2197.
Matsukawa, N., Matsumoto, M., Bukawa, W., Chiji,
H., Nakayama, K., Hara, H., & Tsukahara,
T. (2011). Improvement of bone strength and
dermal thickness due to dietary edible bird’s
nest extract in ovariectomized rats. Bioscience,
Biotechnology and Biochemistry, 75(3), 590-
592.
Norhayati, M. K., Azman, O., & Wan Nazaimoon, W.
M. (2010). Preliminary study of the nutritional
content of Malaysian edible bird’s nest.
Malaysian Journal of Nutrition, 16(3), 389-396.
Norhayati, M. T., Suhaimi, S., Foo, T. F., M. Suhaimi,
M., & M. Talib, L. (2013). Temporal distribution
and chemical characterization of atmospheric
particulate matter in the Estern Coast of
Peninsular Malaysia. Aerosol and Air Quality
Research, 13(2), 584-595.
Northup, D. E. & Lavoie, K. H. (2010).
Geomicrobiology of cave: A Review.
Geomicrobiology Journal, 18(3), 199-222.

Major Minerals Composition Data
195Pertanika J. Trop. Agric. Sci. 39 (2): 181 - 196 (2016)
Nurul Huda, M. Z., Zuki, A. B. Z., Azhar, K., Goh,
Y. M., Suhaimi, H., Awang Hazmi, A. J., &
Zairi, M. S. (2008). Proximate, elemental and
fatty acid analysis of pre-processed edible bird’s
nest (Aerodramus fuciphagus): A comparison
between regions and type of nest. Journal of
Food Technology, 6(1), 39-44.
Pettersen, R. C. (1984). The Chemical Composition
of Wood. In R. Rowell), The Chemistry of Solid
Wood (p. 119). American Chemical Society.
Roggo, Y., Duponchel, L., & Huvenne (2003).
Comparison of supervised pattern recognition
methods with McNemar’s statistical test
Application to qualitative analysis of sugar
beet by near-infrared spectroscopy. Analytica
Chimica Acta, 477(2), 187-200.
Roh, K. B., Lee, J., Kim, Y. S., Park, J., Kim, J. H.,
Lee, J., & Park D. (2012). Mechanism of edible
bird’s nest extract-induced proliferation of
human adipose-derived stem cells. Evidence-
based Complementary and Alternative Medicine,
2012.
Saengkrajang, W., Matan, N., & Matan, N. (2013).
Nutritional composition of the farmed edible
bird’s nest (Collocalia fuciphaga) in Thailand.
Journal of Food Composition and Analysis,
31(1), 41-45.
Sia, Y. H., & Tan, J. H. (2014). Swiftlets farming for
production of edible bird’s nests. Patents US
8651061 B2.
Wolfenden, E. B. (1965). Bau mining district, West
Sarawak, Malaysia Part I: Bau. Malaysia:
Government Printing Ofce, Kuching, Sarawak,
(Chapter II).
Wu, Y., Chen, Y., Wang, B., Bai, L., Wu, R. H., Ge,
Y., & Yuan, F. (2010). Application of SYBRgreen
PCR and 2DGE methods to authenticate edible
bird’s nest food. Food Research International,
43(8), 2020-2026.
Zainal Abidin, F., Chua, K. H., Ng, S. L., Mohd Ramli,
E. S., Lee, T. H., & Abd Ghafar, N. (2011).
Effects of edible bird’s nest (BN) on cultured
rabbit corneal keratocytes. BMC Complementary
and Alternative Medicine, 11(1), 94.