Journal of Agricultural Sciences
Vol. 60, No. 1, 2015
Pages 33-48
DOI: 10.2298/JAS1501033M
UDC: 633.15:575.22
Original scientific paper
CLUSTER AND PRINCIPLE COMPONENT ANALYSES OF MAIZE
ACCESSIONS UNDER NORMAL AND WATER STRESS CONDITIONS
Hafiz Saad Bin Mustafa1, Jehanzeb Farooq
2*, Ejaz-ul-Hasan
1,
Tahira Bibi1 and Tariq Mahmood
1
1Oilseeds Research Institute, AARI, Faisalabad, Pakistan 2Cotton Research Institute, AARI, Faisalabad, Pakistan
Abstract: In the current set of an experiment, forty maize genotypes were
assessed for drought associated traits. For evaluation of these traits, PC and
correlation analyses were employed to obtain suitable parents that can be further
exploited in future breeding programmes. Correlation analysis revealed some
important associations among the traits studied. Fresh root length had positive and
significant associations, but leaf temperature had a significant negative correlation
with root density at both 40% and 100% moisture levels while root density had
negative association at 100% and positive correlation at 40% moisture level with
chlorophyll content. The positive correlation among these yield contributing traits
suggested that these characters are important for direct selection of drought tolerant
high yielding genotypes. Principal component (PC) analysis showed first 4 PCs
having Eigen value >1 explaining 86.7% and 88.4% of the total variation at 40%
and 100% moisture levels respectively with different drought related traits. Cluster
analysis classified 40 accessions into four divergent groups. The members of
clusters 1 and 2 may be combined in future breeding programmes to obtain
genotypes/hybrids that can perform well under drought stress conditions. Members
of cluster 3 may be selected on the basis of root density, leaf temperature, dry root
weight and root shoot ratio by weight and can be combined with members of
cluster 4 due to higher leaf temperature and root shoot ratio by length. The results
showed that the germplasm having a wide genetic diversity can be thus utilized for
future breeding programme to obtain drought tolerant maize genotypes/ hybrids for
adaptation to water scarce areas.
Key words: Zea mays, cluster analysis, drought, genotypes, principle
component analysis.
*Corresponding author: e-mail: [email protected]
RETRACTED
Hafiz Saad Bin Mustafa et al.
34
Introduction
Maize is a very important crop of Pakistan and ranks third among cereal crops
after wheat and rice. In Pakistan, maize is grown in an area of 1,083 thousand
hectares with total production of 3,990 thousand tons (Anonymous, 2011 12). It is
consumed as food by humans and as feed for the livestock and poultry. It is also
used as raw material in different food, medicine and textile industries, which
finally manufacture corn oil, corn flakes, dextrose, and textile dyes etc. Maize has
high nutritive values as it contains 72 % starch, 10 % protein, 4.80 % oil, 8.50 %
fibre, 3.0 % sugar, and 1.70 % ash (Mustafa et al., 2014).
Among different abiotic stresses, drought stress or water deficit is an
inevitable and recurring feature of world agriculture. It has been estimated that
about one third of the world’s potentially arable land suffers water shortage and
crop yields are often reduced by drought (Khan et al., 2004). Drought can reduce
crop yield and pasture deterioration by increasing poisonous elements and causing
death of livestock. It strongly affects the production of cereals, and poses a serious
threat to the food security of households. World food security is dependent on
continuous crop improvement. The development of crops with increased tolerance
to abiotic stresses especially drought and salinity is the need of the hour throughout
the world (Denby and Gehring, 2005). Bänziger et al. (2000) observed that the full-
sib family selection scheme was the most extensively used at CIMMYT to improve
maize populations for drought and low-N tolerance. For high production in maize,
the precise selection of elite genotypes is very important for any area
(AshoftehBeiragi et al., 2010). Knowledge of the associations among genotypes
would help to identify a set of genotypes that have maximal diversity for the
analysis of the effects of genetic background.
To get precise information on nature and extent of genetic variation depends
upon the various techniques used for its estimation, like plant characterization based
on agronomical, morphological and physiological traits (Bajracharya et al., 2006).
Multivariate analysis based on Mahalonobis’s D2 statistics (MDS), principal
component analysis (PCA) and principal coordinate analysis (PCoA) are mostly used
to evaluate the magnitude of genetic diversity among the germplasm (Brown-
Guedira, 2000). Several authors suggested first principal component (PC) scores as
input variables for the clustering process (Mujaju and Chakuya, 2008). Hierarchical
cluster analysis has been suggested for classifying entries of germplasm collections
based on degree of similarity and dissimilarity (Van Hintum, 1995). Similarly, a
combination of cluster analysis and principal component analysis has been used to
classify maize (Zea mays L.) accessions (Crossa et al., 1995). Several studies have
reported comparisons of cluster analysis algorithms, including those used for
classification of germplasm collections (Peeters and Martinelli, 1989), and
classification of maize inbreds (Mummand and Dudley, 1994).
RETRACTED
Cluster and principle component analyses of maize accessions
35
Among these biometrical procedures, the main edge of principal component
analysis (PCA) is that each genotype can be assigned to only one group and it also
reflects the significance of the largest contributor to the total variability at each axis
of differentiation (Sharma, 1998). Genetic variation for morphological traits has
been estimated using principal component analysis, which has led to the
recognition of phenotypic variability in cotton (Sarvanan et al., 2006; Esmail et al.,
2008; Li et al., 2008).
The purpose of the present study was to evaluate the genetic diversity among
maize genotypes specifically for drought tolerance to select the best genotypes at
seedling stage by using PCA and cluster analysis that can be exploited in future
maize breeding programme.
Material and Methods
The investigation presented here was carried out in the glasshouse of the
Department of Plant Breeding and Genetics, University of Agriculture, Faisalabad,
Pakistan during 2010. The experimental material consisted of 40 accessions (inbred
lines) namely: EV-1098, EV-5098, EV-6098, SWL-2002, Agaiti-2002, Sadaf, Pak
Afgoyee, TLOZA, TLOOB-341, TLOOB-343, F-153, F-204, F-220, F-115, F-186,
F-135, F-202, F-206, F-209, F-211, F-210, F-192, F-114, B-54, AF-02B, F-109, F-
110, F-163, F-208, F-189, F-219, F-158, F-160, F-191, PR-98, 8288, 8441, 6525,
32B33 and 33H25. These accessions were seeded in polythene bags (18 × 9 cm)
filled with sandy loam soil (pH 7.8 and EC 1.7 dS m-1) using completely
randomized design with three replications. Field capacity of the soil was determined
before seeding. Two seeds per polythene bag were sown and thinned up to one
healthy seedling after emergence. All the recommended agronomic and cultural
practices were carried out. The moisture level was maintained by volume on alternate
days by using a moisture meter (ΔT-NH2, Cambridge, England). The average
temperature throughout the experimental period was 420C. At five-leaf stage, five
plants were selected randomly from each genotype and data was recorded for the leaf
temperature (LT), fresh root length (FRL), fresh shoot length (FSL), root-to-shoot
length ratio (RSR/L), dry root weight (DRW), dry shoot weight (DSW), root-to-
shoot weight ratio (RSR/W), root density (RD) and chlorophyll contents (CC) at
normal (100%) and 40% moisture levels of field capacity.
The data was subjected to basic statistics, correlation analysis, cluster analysis
and principal component analysis (PCA) using statistical software packages of
SPSS version 19 and Statistica version 5.0 (Sneath and Sokal, 1973). Cluster
analysis was performed using K-means clustering while tree diagram based on
eucladian distances was developed by Ward’s method. The first two principal
components were plotted against each other to find out the patterns of variability
among genotypes using SPSS version 19.
RETRACTED
Hafiz Saad Bin Mustafa et al.
36
Results and Discussion
Normal conditions (100% moisture level)
The basic statistics of various traits studied under normal conditions (100%
moisture level) have demonstrated considerable variability among 40 maize
genotypes (Table 1). Simple correlation coefficients revealed 10 significant
associations among traits (Table 2). FRL had a positive and significant association
with RSR/L, FSL, DSW. LT had a negative and significant association with RD.
FSL exhibited a positive correlation with DRW and DSW, but a negative and
significant association with RSR/L. DRW had a positive and significant association
with DSW and RSR/W. RD showed a positive and significant association with CC.
Table 1. Basic statistics for various traits of 40 maize genotypes under normal
conditions (100% moisture level).
Variable Mean Mean SE SD Variance CV Minimum Maximum Range
FRL 20.492 0.709 4.481 20.080 21.87 12.333 31.667 19.333
LT 31.167 0.165 1.046 1.094 3.36 28.500 33.667 5.167
FSL 51.26 1.45 9.14 83.55 17.83 31.33 69.00 37.67
DRW 0.6183 0.0292 0.1847 0.0341 29.88 0.2000 1.1667 0.9667
DSW 0.7075 0.0243 0.1539 0.0237 21.75 0.2333 1.0667 0.8333
RD 4.067 0.300 1.898 3.602 46.67 1.667 8.000 6.333
RSR L 0.4046 0.0132 0.0834 0.0070 20.62 0.2375 0.6640 0.4265
RSR W 0.8683 0.0204 0.1289 0.0166 14.85 0.6667 1.1852 0.5185
CC 0.3480 0.0719 0.4547 0.2068 130.69 0.0130 1.1240 1.1110
SE = Standard error, SD = Standard deviation, CV = Coefficient of variation, FRL = Fresh root
length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight, DSW = Dry shoot
weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot ratio weight,
CC = Chlorophyll content.
Table 2. Simple correlation coefficients of various physiological traits in maize
under normal conditions (100% moisture level).
Traits FRL LT FSL DRW DSW RD RSR/L RSR/W
LT -0.037
FSL 0.523** -0.118
DRW 0.379 -0.044 0.526**
DSW 0.456** -0.084 0.663** 0.847**
RD -0.074 -0.625** 0.065 -0.084 -0.035
RSR/L 0.623** 0.076 -0.333* -0.070 -0.114 -0.125
RSR/W 0.094 0.016 0.087 0.686** 0.207 -0.104 0.025
CC 0.041 -0.520 0.149 -0.036 0.028 0.877** -0.079 -0.098
FRL = Fresh root length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight,
DSW = Dry shoot weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot
ratio weight, CC = Chlorophyll content.
RETRACTED
Cluster and principle component analyses of maize accessions
37
In this study, out of total l9, four principal components (PCs) extracted had
Eigen value >1. These four PCs contributed 88.4% of the total variability amongst
the maize genotypes assessed for drought related traits (Table 3). However, the
remaining components contributed only 11.6% towards the total diversity for this
set of maize genotypes. The PC I contributed maximum towards the variability
(32.3%) followed by PC II (27.0%), PC III (16.9%) and PC IV (12.2%). The traits
like DRW, DSW, FSL and FRL showed considerable positive factor loadings on
PC I while LT and RSR/L showed negative loadings (Table 4).
Table 3. Principle component analysis of different physiological traits in maize
under normal conditions (100% moisture level).
PC 1 PC 2 PC 3 PC 4
Eigen value 2.9038 2.4278 1.5219 1.1008
% of total variance 32.3 27.0 16.9 12.2
Cumulative variance % 32.3 59.2 76.2 88.4
Table 4. Factor loadings by various traits (100% moisture level).
Variable PC I PC II PC III PC IV
FRL 0.370 0.075 -0.581 -0.193
LT -0.085 0.487 0.077 -0.153
FSL 0.457 -0.092 0.123 -0.456
DRW 0.537 0.086 0.143 0.273
DSW 0.523 0.013 0.088 -0.181
RD 0.006 -0.604 -0.067 0.123
RSR/L -0.012 0.162 -0.759 0.207
RSR/W 0.286 0.135 0.139 0.750
CC 0.053 -0.576 -0.117 0.060
FRL = Fresh root length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight,
DSW = Dry shoot weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot
ratio weight, CC = Chlorophyll content.
The 2nd PC was related to diversity among maize genotypes due to LT with
their positive loadings and RD and CC with negative loadings. The PC III was
explained by variation among genotypes due to DRW with their positive loadings
and negative loadings exhibited by RSR/L and FRL. The PC IV was elucidated by
diversity among the genotypes for FSL with negative loadings and RSR/W had a
positive value. A PC biplot in Figure1 showed that variables and genotypes are
super imposed on the plot as vectors. The distance of each variable with respect to
PC-1 and PC-2 showed the contribution of these variables in the variation of
genotypes used. The biplot showed that DRW, RD, CC and LT as a whole
contributed maximum towards variability in maize germplasms.
RETRACTED
Hafiz Saad Bin Mustafa et al.
38
Figure 1. Biplot between PCs 1 and 2 showing contribution of various traits in
variability of germplasms under normal conditions (100% moisture level).
Forty maize genotypes were grouped into 4 clusters based on various drought
related traits (Table 5). Cluster analysis showed that cluster 1 was comprised of 19
genotypes, cluster 2 of 5 while cluster 3 had 8 and cluster 4 contained 8 genotypes
(Table 6). The genotypes in cluster 1 showed higher values of FRL, RSR/L and
RSR/W. Similarly, the 2nd cluster was comprised of genotypes having the highest
value and reasonable values of LT. The members of the 3rd cluster were
characterized by higher value of RD. The cluster 4 is characterized by maximum
FSL, DRW, DSW and CC.
Table 5. Cluster analysis of various traits in maize under normal conditions (100%
moisture level).
Variable Cluster 1 Cluster 2 Cluster 3 Cluster 4
FRL 31.67 12.67 12.33 24.00
LT 30.80 33.10 31.33 31.00
FSL 53.33 53.33 31.33 69.00
DRW 0.57 0.43 0.20 0.67
DSW 0.63 0.53 0.23 0.87
RD 3.00 2.67 3.67 3.00
RSR/L 0.59 0.24 0.39 0.35
RSR/W 0.89 0.81 0.86 0.77
CC 0.03 0.04 0.04 0.08
FRL = Fresh root length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight,
DSW = Dry shoot weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot
ratio weight, CC = Chlorophyll content.
5.02.50.0-2.5-5.0
3
2
1
0
-1
-2
-3
-4
First Component
Se
co
nd
Co
mp
on
en
t
CC
RSR WRSR L
RD
DSWDRW
FSL
LT
FRL
Biplot of FRL, ..., CC
RETRACTED
Cluster and principle component analyses of maize accessions
39
Table 6. Cluster membership of various maize genotypes under normal conditions
(100% moisture level).
Cluster Name No. of genotypes Name of genotypes in each cluster
Cluster 1 19
EV-1098, EV-5098, EV-6098, Sadaf, Pak Afgoyee, TLOZA,
TLOOB-341, F-206, F-109, F-110, F-114, F-186, F-160,
F-211, B-54, AF-02B, 8288, 8441, 6525
Cluster 2 5 SWL-2002, Agaiti-2002, TLOOB-343, F-135, F-209,
Cluster 3 8 F-202, F-153, F-204, F-208, F-189, F-219, F-210, 33H25
Cluster 4 8 F-220, F-115,F-191, F-163, F-158, F-192, PR-98, 32B33
The tree diagram showed more or less similar results comprising two main
groups A and B each of which is further subdivided into two clusters (Figure 2).
Figure 2. Tree diagram of 40 maize genotypes based on different physiological
traits under normal conditions (100% moisture level).
RETRACTED
Hafiz Saad Bin Mustafa et al.
40
Drought stress conditions (40% moisture level)
The basic statistics of various traits studied under drought stress conditions
(40% moisture level) has demonstrated considerable variability among 40 maize
genotypes (Table 7). Simple correlation coefficients revealed 9 significant
associations among traits (Table 8). FRL had a positive and significant association
with FSL and RSR/L. LT had a negative and significant association with RD and
positive with CC. FSL exhibited a positive correlation with DRW and DSW, but a
negative and significant association with RSR/L. DRW had a positive association
with DSW and RSR/W, but DSW showed a negative correlation with RSR/W. RD
showed a positive association with CC.
Table 7. Basic statistics for various traits of 40 maize genotypes under drought
stress conditions (40% moisture level).
Variable Mean Mean SE SD Variance CV Minimum Maximum Range
FRL 20.275 0.686 4.341 18.843 21.41 12.333 28.000 15.667
LT 34.103 0.121 0.764 0.583 2.24 31.867 35.033 3.167
FSL 28.542 0.865 5.468 29.901 19.16 16.333 46.000 29.667
DRW 0.5123 0.0212 0.1343 0.0180 26.22 0.2000 0.9000 0.7000
DSW 0.3917 0.0151 0.0957 0.0092 24.44 0.2000 0.5667 0.3667
RD 3.258 0.188 1.190 1.416 36.52 1.667 5.667 4.000
RSR L 0.7231 0.0244 0.1544 0.0238 21.35 0.4262 1.0800 0.6538
RSR W 1.3368 0.0501 0.3169 0.1004 23.70 0.7692 2.2500 1.4808
CC 0.3382 0.0722 0.4565 0.2084 134.98 0.0113 1.1210 1.1097
SE = Standard error, SD = Standard deviation, CV = Coefficient of variation, FRL = Fresh root
length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight, DSW = Dry shoot
weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot ratio weight,
CC = Chlorophyll content.
Table 8. Simple correlation coefficients of various physiological traits in maize
under drought stress conditions (40% moisture level).
Traits FRL LT FSL DRW DSW RD RSR/L RSR/W
LT -0.110
FSL 0.355 0.087
DRW 0.044 0.145 0.351*
DSW 0.255 0.122 0.484** 0.595**
RD 0.301 -0.716** 0.093 0.195 0.197
RSR/L 0.662** -0.152 -0.442* -0.268 -0.189 0.191
RSR/W -0.221 0.070 -0.149 0.496** -0.381 -0.027 -0.078
CC 0.169 -0.606** 0.027 0.048 0.165 0.805** 0.143 -0.124
FRL = Fresh root length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight,
DSW = Dry shoot weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot
ratio weight, CC = Chlorophyll content.
RETRACTED
Cluster and principle component analyses of maize accessions
41
In this study, out of total 9, four principal components (PCs) extracted had
Eigen value >1. These four PCs contributed 86.7% of the total variability amongst
the maize genotypes assessed for drought related traits (Table 9). However, the
remaining components contributed only 13.3% towards the total diversity for this
set of maize genotypes. The PC I contributed maximum towards the variability
(29.9%) followed by PC II (24.9%), PC III (17.9%) and PC IV (14.1%). The traits
like RD and CC showed considerable positive factor loadings on PC I while LT
showed negative loadings (Table 10). The 2nd PC was related to diversity among
maize genotypes due to FSL, DRW and DSW with their positive loadings and
RSR/L with negative loadings. The PC III was explained by variation among
genotypes due to FRL with their positive loadings and negative loadings exhibited
by RSR/W. The PC IV was elucidated by diversity among the genotypes for FRL,
DRW, RSR/L and RSR/W with negative loadings. A PC biplot in Figure3 showed
that variables and genotypes are super imposed on the plot as vectors. The distance
of each variable with respect to PC-1 and PC-2 showed the contribution of these
variables in the variation of genotypes used. The biplot showed that as a whole CC,
RD and LT contributed maximum towards variability in maize germplasms.
Table 9. Principle component analysis of different physiological traits in maize
under drought stress conditions (40% moisture level).
PC I PC II PC III PC IV
Eigen value 2.6901 2.2366 1.6110 1.2663
% of total variance 29.9 24.9 17.9 14.1
Cumulative variance % 29.9 54.7 72.6 86.7
Table 10. Factor loadings by various traits (40% moisture level).
Variable PC I PC II PC III PC IV
FRL 0.340 0.040 0.501 -0.411
LT - 0.440 0.278 -0.191 0.213
FSL 0.111 0.513 0.168 0.120
DRW 0.060 0.513 -0.280 -0.418
DSW 0.196 0.525 0.210 0.070
RD 0.552 0.001 -0.246 -0.004
RSR L 0.225 -0.386 0.347 -0.491
RSR W -0.155 0.016 -0.536 -0.580
CC 0.507 -0.043 -0.230 0.150
FRL = Fresh root length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight,
DSW = Dry shoot weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot
ratio weight, CC = Chlorophyll content.
RETRACTED
Hafiz Saad Bin Mustafa et al.
42
Figure 3. Biplot between PCs 1 and 2 showing contribution of various traits in
variability of germplasms under drought stress conditions (40% moisture level).
Forty maize genotypes were grouped into 4 clusters based on various drought
related traits (Table 11). Cluster analysis showed that cluster 1 was comprised of 11
genotypes, cluster 2 of 16 while cluster 3 had 7 and cluster 4 contained 6 genotypes
(Table 12). The genotypes in cluster 1 showed higher values of FSL, DSW and RD.
Similarly, the 2nd cluster was comprised of genotypes having the highest value and
reasonable values of RD. The members of the 3rd cluster were characterized by higher
value of DRW and RD.
Table 11. Cluster analysis of various traits in maize under drought stress conditions
(40% moisture level).
Variable Cluster 1 Cluster 2 Cluster 3 Cluster 4
FRL 23.00 27.00 15.33 13.67
LT 34.20 34.27 34.53 35.03
FSL 46.00 25.00 29.33 16.33
DRW 0.73 0.37 0.90 0.60
DSW 0.53 0.27 0.43 0.27
RD 3.67 2.67 3.67 1.67
RSR L 0.50 1.08 0.52 0.84
RSR W 1.38 1.38 2.08 2.25
CC 0.06 0.07 0.23 0.02
FRL = Fresh root length, LT = Leaf temperature, FSL = Fresh shoot length, DRW = Dry root weight,
DSW = Dry shoot weight, RD = Root density, RSR/L = Root shoot ratio length, RSR/W = Root shoot
ratio weight, CC = Chlorophyll content.
43210-1-2-3
4
3
2
1
0
-1
-2
-3
-4
First Component
Se
co
nd
Co
mp
on
en
t
CCRSR W
RSR L
RD
DSWDRWFSL
LT
FRL
Biplot of FRL, ..., CC
RETRACTED
Cluster and principle component analyses of maize accessions
43
Table 12. Cluster membership of various genotypes under drought stress conditions
(40% moisture level).
Cluster Name No. of genotypes Name of genotypes in each cluster
Cluster 1 11 EV-1098, EV-6098, Pak Afgoyee, F-202, F-115, F-186,
F-158, F-210, F-192, F-211, B-54
Cluster 2 16 EV-5098, TLOZA, TLOOB-341, F-135 , F-209, F-153, F-220,
F-163, F-208, F-219, AF-02B, PR-98, 8288,8441, 6525, 32B33
Cluster 3 7 SWL-2002, Sadaf, TLOOB-343, F-114, F-204, F-160, 33H25
Cluster 4 6 Agaiti-2002, F-206, F-109, F-110, F-191, F-189
The cluster 4 is characterized by maximum LT and RSR/W. The tree diagram
showed more or less similar results comprising two main groups A and B each of
which is further subdivided into two clusters (Figure 4). Multivariate data analysis
assists a graphic display of the underlying latent factors and an interface between
individual samples and variables (Nielsen and Munck, 2003). Principal component
analysis (PCA) is usually used in plant sciences for the reduction of variables and
grouping of genotypes. Kamara et al. (2003) used PCA to categorize traits of maize
(Zea mays L.) that accounted for most of the variance in the data. Granati et al. (2003)
used PCA to scrutinize the relationship among Lathyrus accessions. Žáková and
Benková (2006) recognized traits that were the main sources of variation of genetic
diversity among 106 Slovakian barley accessions. Cartea et al. (2002) and Salih et al.
(2006) used PCA and cluster analysis to group kale populations and winter wheat
genotypes, respectively. The classification of diversity among the genotypes into
groups with similar traits can be used to design a collection strategy (Ariyo, 1993).
When dissimilarity between a pair of a variety is defined on a multivariate
criterion, it is useful to be able to determine the specific plant characters which cause
the dissimilarity and the relative contributions that the various characters make to the
total variability in the germplasm (Ariyo, 1993). The information regarding association
among various traits is an important part for the initiation of any breeding programme
as it provides an opportunity for the selection of genotypes having desirable traits
simultaneously (Ali et al., 2009). In the present set of the experiment, results of
correlation analysis revealed some important associations among the traits studied.
FRL had positive and significant associations at both 40% and 100% moisture levels,
but LT had a significant negative correlation with RD at both 40% and 100% moisture
levels while RD had a negative correlation at 100% and a positive correlation at 40%
moisture level with CC. FSL had positive significant associations with DSW and DRW
while negative with RSR/L both at 40% and 100% moisture levels.
DRW had a positive significant correlation with DSW and RSR/W both at 40%
and 100% moisture levels. DSW had showed a significant negative correlation with
RSR/W at 40% moisture level while its positive but non significant association was
found at 100% moisture level. RD had showed a positive and significant correlation
RETRACTED
Hafiz Saad Bin Mustafa et al.
44
with CC both at 40% and 100% moisture levels. The positive correlation among these
yield contributing traits suggested that these characters are important for direct
selection of high yielding genotypes. The maintenance and exploitation of genetic
resources could be made by partitioning the total variance into its components. It also
provides a chance for utilization of appropriate germplasm in crop improvement for
particular plant traits (Sneath and Sokal, 1973; Pecetti et al., 1996). The principal
component (PC) analysis divides the total variance into different factors. The Principal
Component Analysis is a powerful tool to obtain parental lines for a successful
breeding programme (Akter et al., 2009).
Figure 4. Tree diagram of 40 maize genotypes based on different physiological
traits under drought stress conditions (40% moisture level).
RETRACTED
Cluster and principle component analyses of maize accessions
45
In this experiment, the PC analysis partitioned the total variance into 4 PCs
contributing maximum to the total diversity among the genotypes due to the study
of various traits. Chozin (2007), Mujaju and Chakuya (2008), and Ali et al. (2011)
reported important contribution of the first PCs in total variability while studying
different traits. In our research, the 1st PC was mainly due to variations in CC, RD,
FRL and LT at 40% moisture level while variations were found for DRW, DSW
and FSL at 100% moisture level. These results are in agreement with the results of
association analysis which showed a positive association among these traits. The
2nd PC was related to diversity among cotton genotypes due to DSW, FSL and
DRW at 40% moisture level while variations were found for LT and RD, CC at
100% moisture level. The PC III was explained by variation among genotypes due
to FRL, RSR/L and RSR/W at 40% moisture level while diversity was found for
traits like FRL and RSR/L at 100% moisture level. Similarly, PC IV was explicated
by variation in RSR/W, RSR/L, DRW and FRL with their considerable negative
factor loadings at 40% moisture level while at 100% moisture level negative
estimates were found for traits like RSR/W and FSL. PC analysis ultimately
confirmed the amount of variation for the traits among the materials in hand which
could be utilized in designing a breeding programme aimed at improving drought
tolerance and ultimately grain yield as it is generally assumed that maximum
variation yields maximum heterotic effects.
While comparing the fresh root length and leaf temperature, the genotypes in
cluster 2 showed desirable results under stress and in cluster 1, the traits like FRL,
FSL, DSW, RD and RSR/W showed promising results. Thus, members of clusters 1
and 2 may be combined in future breeding programmes to obtain genotypes/hybrids
that can perform well under drought stress conditions. The members of cluster 3 may
be selected on the bases of RD, LT, DRW and RSR/W and they may be combined
with the members of cluster 4 due to higher leaf temperature and RSR/L. The
germplasms exploited to obtain maximum diversity in this experiment yield excellent
results. Amurrio et al. (1995) and Rabbani et al. (1998) reported a lack of relationship
between various clusters based on agronomic traits and origins of genotype in peas
(Pisumsativum) and mustard (Brassica juncea) respectively. The tree diagram was
comprised of group of genotypes showing considerable tolerance to drought stress.
The occurrence of this wide variation between the clusters is of great genetic value in
providing materials aimed at maize selection for adaptation to water scarce areas. A
similar kind of results related to germplasm grouping has been reported by Ayana
and Bekele (1999) and Grenier et al. (2001).
Conclusion
PC analysis, cluster analysis and correlation coefficient in this present set of
the experiment provided facilitation in the classification of genotypes and
identification of the subset of core genotypes having tolerance to drought stress.
RETRACTED
Hafiz Saad Bin Mustafa et al.
46
Various useful correlations and aforementioned information extracted from cluster
and PC analysis will be helpful in designing breeding programmes to obtain
drought tolerant maize genotypes/hybrids.
References
Akter, A., Hasan, M.J., Paul, A.K., Mutlib, M.M., Hossain, M.K. (2009): Selection of parent for
improvement of restorer line in rice (Oryzasativa L.). SAARC Journal of Agriculture 7:43-50.
Ali, M.A., Nawab, N.N., Abbas, A., Zulkiffal, M., Sajjad, M. (2009): Evaluation of selection criteria
in Cicerarietinum L. using correlation coefficients and path analysis. Australian Journal of Crop
Science 3:65-70.
Ali., M.A., Jabran, K., Awan, S.I., Abbas, A., Ehsanullah, Zulkiffal, M., Acet, T., Farooq, J.,
Rehman, A. (2011): Morpho-physiological diversity and its implications for improving drought
tolerance in grain sorghum at different growth stages. Australian Journal of Crop Science
5(3):311-320.
Amurrio, J.M., de Ron, A.M., Zeven, A.C. (1995): Numerical taxonomy of Iberian pea landraces
based on quantitative and qualitative characters. Euphytica 82:195-205.
Anonymous, 2011-2012. Economic Survey. Government of Pakistan, Finance Division, Economic
Advisor’s Wing, Islamabad.
Ariyo, O.J. (1993): Genetic diversity in West African okra (Abelmoschuscaillei L. (Chev.)Stevels-
Multivariate analysis of morphological and agronomical characteristics. Genetic Resources and
Crop Evolution. 40:25-32.
AshoftehBeiragi, M., SiahSar, B.A., KhavariKhrasani, S., Golbashy, M., Mahdi Nejad, N., Ali Zade, A.
(2010): Effects of genotype by environment interactions on morphological traits, yield and yield
components of new grain corn (Zea mays L.) varieties. Journal of Agroecology 2(1):136-145.
Ayana, A., Bekele, E. (1999): Multivariate analysis of sorghum (Sorghum bicolor (L.) Moench)
germplasm from Ethiopia and Eritrea. Genetic Resources and Crop Evolution 46:273-284.
Bajracharya, J., Steele, K.A., Jarvis, D.I., Sthapit, B.R., Witcombe, J.R. (2006): Rice landrace
diversity in Nepal: Variability of agro-morphological traits and SSR markers in landraces from
a high-altitude site. Field Crops Res. 95:327-335.
Bänziger, M., Edmeades, G.O., Beck, D., Bellon, M. (2000): Breeding for drought and nitrogen stress
tolerance in maize: From theory to practice, Mexico, D.F. CIMMYT 68p.
Brown-Guedira, G.L. (2000): Evaluation of genetic diversity of soybean introductions and North
American ancestors using RAPD and SSR markers. Crop Science 40:815-823.
Cartea, M.E., Picoagea, A., Soengas, P., Ordás, A. (2002): Morphological characterization of kale
populations from northwestern Spain. Euphytica 129:25-32
Chozin, M. (2007): Characterization of sorghum accessions and choice of parents for hybridization.
Journal of Akta Agriculture EdisiKhusus 2:227-232.
Crossa, J., Delacy, I.H., Taba, S. (1995): The use of multivariate methods in developing a core
collection. p. 77-92. In: Hodgkin, T., Brown, A.H.D., VanHintum, Th.J.L., Morales, E.A.V.
(Eds.) Core collections of plant genetic resources. John Wiley & Sons, Chichester, UK.
Denby, K., Gehring, C. (2005): Engineering drought and salinity tolerance in plants: lessons from
genome-wide expression profiling in Arabidopsis. Trends Biotech. 23:547-552.
Esmail, R.M., Zhang, J.F., Abdel-Hamid, A.M. (2008): Genetic diversity in elite cotton germplasm
lines using field performance and RAPD markers. World Journal of Agriculture Science
4:369-375.
Granati, E., Bisignano, V., Chiaretti, D., Crino, P., Polignano, B.G. (2003): Characterization of Italian
and exotic Lathyrus germplasm for quality traits. Genetic Resources and Crop Evolution
50:273-280.
RETRACTED
Cluster and principle component analyses of maize accessions
47
Grenier, C., Deu, M., Kresovich, S., Bramel, Cox, P.J. Hamon, P. (2001): Assessment of genetic
diversity in three subsets constituated from the ICRISAT sorgham collection using random vs.
non-random sampling procedures by using molecular markers. Theoratical and Applied
Genetics 101:197-202.
Kamara, A.Y., Kling, J.G., Menkir, A., Ibikunle, O. (2003): Agronomic performance of maize (Zea
mays L.) breeding lines derived from low nitrogen maize population. Journal of Agricultural
Science 141:221-230.
Khan, I.A., Habib, S., Sadaqat, H.A., Tahir, M.N.H. (2004): Selection criteria based on seedling
growth parameters in maize varies under normal and water stress conditions. International
Journal of Agriculture & Biology 6(2):252-256.
Li, Z., Wang, X., Yan, Z., Guiyin, Z., Wu, L., Jina, C., MA, Z. (2008): Assessment of genetic
diversity in glandless cotton germplasm resources by using agronomic traits and molecular
markers. Frontiers of Agriculture in China 2:245-252.
Mujaju, C., Chakuya, E. (2008): Morphological variation of sorghum landrace accessions on-farm in
Semi-arid areas of Zimbabwe. International Journal of Botany 4:376-382.
Mummand, R.H., Dudley, J.W. (1994): A classification of 148 U.S. maize inbreds: I. Cluster analysis
based on RFLPs. Crop Science 34:842-851.
Mustafa, H.S.B., Aslam, M., Hasan E.U. Hussain, F., Farooq, J. (2014): Genetic variability and path
coefficient in maize (Zea mays L.) genotypes. The Journal of Agricultural Sciences 9(1):37-43.
Nielsen, J.P., Munck, L. (2003): Evaluation of malting barley quality using exploratory data analysis.
I. Extraction of information from micromalting data of spring and winter barley. Journal of
Cereal Science 38:173-180.
Peeters, J.P., Martinelli, J.A. (1989): Hierarchical clustering analysis asa tool to manage variation in
germplasm collections. Theoretical and Applied Genetics 78: 42-48.
Pecetti, L., Annicchiario, P., Damania, A.B. (1996): Geographic variation in tetraploid wheat
(Triticumturgidum spp. Turgidumconvar. Durum) landraces from two provinces in Ethiopia.
Genetic Resources and Crop Evolution 43:395-407.
Rabbani, M.A., Iwabuchi, A., Murakami, Y., Suzuki, T., Takayanagi, K. (1998): Phenotypic variation
and the relationships among mustard (Brassica juncea L.) germplasm from Pakistan. Euphytica
101:357-366.
Salih, S., Grausgruber, H., Ruckenbauer, P. (2006): Agronomic and quality performance of
international winter wheat genotypes grown in Kosovo. Cereal Research Communication
34:957-964.
Saravanan, S., Arutchendhil, P., Raveendran, T.S., Koodalimgam, K. (2006): Assessment of genetic
divergence among introgressed culture of Gossypiumhirsutum L. through RAPD analysis.
Journal of Applied Science and Research 2:1212-1216.
Sharma, J.R. (1998): Statistical and biometrical techniques in plant breeding. New Age International
Limited Publishers, New Delhi, India.
Sneath, P.H.A, Sokal, R.R. (1973): Numerical Taxonomy: The Principles and practice of numerical
classification. Free-Man WF and Co, San Francisco, USA.
Van Hintum, T.J.L. (1995): Hierarchical approaches to the analysis of genetic diversity in crop plants.
In: Hodgkin, T., Brown, A.H.D., Van Hintum T.J.L., Morales, E.A.V. (Eds.) Core collection of
plant genetic resources. John Wiley and Sons, pp. 23-34.
Žáková, M., Benková, M. (2006): Characterization of spring barley accessions based on multivariate
analysis. Communication in Biometry and Crop Science 1:124-134.
Received: July 15, 2014
Accepted: October 10, 2014
RETRACTED
Hafiz Saad Bin Mustafa et al.
48
KLASTER ANALIZA I ANALIZA GLAVNIH KOMPONENTI GENOTIPOVA
KUKURUZA U NORMALNIM USLOVIMA I U USLOVIMA
VODNOG STRESA
Hafiz Saad Bin Mustafa1, Jehanzeb Farooq
2*, Ejaz-ul-Hasan
1,
Tahira Bibi1 i Tariq Mahmood
1
1Institut za seme uljarica, AARI, Fejsalabad, Pakistan
2Institut za pamuk, AARI, Fejsalabad, Pakistan
R e z i m e
U ovom ogledu, četrdeset genotipova kukuruza su procenjeni u pogledu
osobina povezanih sa sušom. Za evaluaciju ovih osobina korišćene su analiza
glavnih komponenti i korelaciona analiza, kako bi se preporučili odgovarajući
roditelji, koji se mogu koristiti u budućim programima oplemenjivanja.
Korelaciona analiza je pokazala da postoji značajna povezanost meĎu nekim
ispitivanim osobinama. Dužina svežeg korena je bila pozitivno i značajno
korelisana, a temperatura lista takoĎe značajno, ali negativno korelisana sa
gustinom korena kako pri 40% tako i pri 100% nivou vlažnosti. Nasuprot tome,
korelacija izmeĎu gustine korena i sadržaja hlorofila je bila negativna pri 100%
nivou vlažnosti, a pozitivna pri 40% nivou vlažnosti. Pozitivna korelacija meĎu
nekim osobinama koje doprinose prinosu ukazuje da su ove osobine važne za
direktnu selekciju visokoprinosnih genotipova tolerantnih na sušu. Analizom
glavnih komponenti izdvojene su četiri glavne komponente Eigen vrednosti veće
od 1 koje objašnjavaju 86,7% i 88,4% ukupne varijabilnosti pri 40% odnosno
100% nivou vlažnosti. Klaster analizom 40 genotipova kukuruza je klasifikovano u
četiri divergentne grupe. Članovi klastera 1 i 2 se mogu kombinovati u budućim
programima oplemenjivanja radi dobijanja genotipova/hibrida tolerantnih na
uslove stresa izazvane sušom. Članovi klastera 3 svrstani na osnovu gustine
korena, temperature lista, mase suvog korena i odnosa mase korena i izdanka mogu
se kombinovati sa članovima klastera 4, koje karakteriše viša temperatura lista i
odnos dužine korena i izdanka. Rezultati su pokazali da germplazma, koja ima
široku genetsku raznovrsnost, može biti iskorišćena u budućim programima
oplemenjivanja, radi dobijanja genotipova/hibrida kukuruza tolerantnh na sušu,
adaptiranih za područja koja oskudevaju vodom.
Ključne reči: Zea mays, klaster analiza, suša, genotipovi, analiza glavnih
komponenti.
Primljeno: 15. jula 2014.
Odobreno: 10. oktobra 2014.
*Autor za kontakt: e-mail: [email protected]
RETRACTED