P.D.Abeytilakarathna1,
H.M.Jayamenike2 and K.G.N.A.B.Wijethunga 3
1 Agricultural Research & Development Center, Department of Agriculture, Bandarawela, Sri Lanka.
2 Agricultural Research Station, Department of
Agriculture, Rahangala, Boralanda, Sri Lanka.
3 Fruit Crop Research & Development Center,
Gannoruwa, Peradeniya, Sri Lanka
ABSTRACT
A yield stability of seven bush bean lines were
evaluated with different cropping seasons in order to develop an efficient
method for selecting bush beans
promising varieties that secure high stable yield with climate changes in bean varietal
breeding programme. Seven bush beans lines were grown in randomised complete
block design with three replicates at the open field of agricultural research
station at Rahangala of department of agriculture during 2010 yala,
2011/12 maha and 2012 yala seasons.
Concurrently use of simple linear regression and
additive main effect and multiplicative interaction (AMMI) models were proposed
to analysis seasonal yield stability of bush beans. The mean yield of AMMI
model was determined by adding grand
mean, genotype effect, seasonal effect, residual and the product of Eigen value of single value interaction
principle component axes (IPCAs) with genotype and seasonal IPCAs scores. Varieties
that were showed above grand mean yield
and regression coefficients (ß1i) which were less than +5 % or
higher than -5 % of grand mean yield, coefficient of determination (r2)
and IPCAs were close to zero were considered as stable, high yielding promising
lines suitable for growing during both yala and maha seasons.
Bush bean variety “Wade” was showed the highest
yield stability over different seasons but its yield was low. Bush bean line,
BB 2904 was found the highest yield, but its yield was highly fluctuated with
different seasons. Crossing bush beans line, BB 2904 with Wade might be beneficial
to breed, high yielding and more yield stable promising variety to mitigate the
future climate change.
Keywords: AMMI model, Bush beans, , IPCAs, Linear
regression, Seasonal yield stability
INTRODUCTION
Bush bean (Phaseolus vulgaris L.) is an
important leguminous crop in the up country intermediate zone of Sri Lanka that
were grown both yala and maha seasons as short age crop which fit
to the existing cropping pattern. Beans are
growing mainly in Badulla, NuwaraEliya and Kandy districts with annual growing
extent of 7700 ha and average production of 5.5 mt/ha (AgStat, 2011).
According to CGIAR (2013) globally,
about 12 million metric tons of common beans are produced annually. In
addition, it provides protein and complex carbohydrates for more than 300
million people in the tropics. In many areas, common bean is the second most
important source of calories after maize. However,
the seasonal climate changes; especially wet seasons were more wetter, dry
seasons were more drier and unpredictability of drought could be effect to
production of bush beans. Farmers throughout the
country are experiencing changed seasons year on year for many the past four
years have been worst in term of aberrant weather. Paddy production in the yala season of 2012 was reduced by 4% due to drought. This also
cause to declining the GDP (20% in 2000 and 13 % in 2010) and it will further
reduced to 8% in 2020.
Climate projections and crop model for 2030 were
stated that South Asia and southern
Africa could suffer negative impacts on several important food security crops
(Lobell et al., 2008). When compare
with cowpea, mungbean like leguminous crops, beans are less adapted to extreme
environments such as very low rainfall, high temperatures, low fertility acid
soils. Nevertheless, gene pools and races within gene pools differ in
adaptation ranges (Beebe et al., 2011). Climate change will alter the
pest and disease incidence and intensity. Equatorial region will receive more
rain as climate changing while extreme rainfall will receive due to La Nina
phenomenon (Babee et al., 2011).
Common beans are grown well in area of mid-altitudes with
moderate temperatures, organic soils, and seasonally abundant rainfall (Beebe
et al. 2008).High yielding, promising varieties in plant breeding
programmes should be evaluated under different condition similar real condition
that they will experience when growing. New varieties should be high
performance for yield and agronomic traits over a wide range of environment
condition. Consequently, new lines are tested under multi-locations trials with
different condition of climate, soil fertility and different seasons of the
years. Locations and seasons are considered to be a single factor for
environmental condition (Acciaresi and Chidichimo, 1999; Becker and Leon,
1998). For this reason, growing season can be consider as the environment
factor in the primary yield trials with different seasons.
The additive
main effect and multiplicative interaction (AMMI) model is widely used for
analyzing genotype x environment interaction (GEI) . This model capture large
portion of GEI sum of squares ( Tarakanovas and Ruzgas, 2006). Genotypes with
above grand mean yield having the lowest mean deviation and single value of
interaction principal component axes (IPCAs) scores which are close to zero
were found to be more effective for selecting wide adaptability and stability
of lines in diverse environment (Abeytilakarathna, 2010).
Understanding
and measuring of the seasonal stability in varietal breeding programmes,
multi-location varietal trials and varietal adaptability trials will be
beneficial to produce promising bush bean varieties with seasonal stable
yields. This could be possible by breeding varieties for yield and yield
component stability ( Lal et al., 2010).
Consequently, there is great need
for technological development to reach the rural farmers. One of the aspects is
to breed new varieties of climate and pest resistant crop (Sunday Times, 2013).Therefore, this paper was aimed to
developed efficient statistical method to evaluate the bean varieties for
seasonal stability using AMMI model and IPCAs vs. mean yield bi-plot.
MATERIALS AND METHODS
Agronomic practices
Seven bush beans lines were
planted in the open field of agricultural research station at Rahangala of
department of agriculture in Boralanda. Beans lines were established on 1.5 x 2 m size plots in randomized complete
block design with three replicates. Planting space was 40 x 10 cm. Compost were
applied at the rate of 10 t/ha on 5 days before seed establishing. Urea, triple
super phosphate (TSP), and muriate of potash (MOP) were applied at the rate of
85, 165 and 65 kg/ha respectively as a basal fertilizer dressing. After three
weeks of planting, urea and MOP were used at the rate of 85 and 65 kg/ha
correspondingly. Pest and disease control and other cultural practices were
carried out according to the recommendation of department of agriculture of Sri
Lanka. Net plot yield were recorded by removing boarders. A combined AMMI
analysis and ANOVA were preceded using AssiStat
version 7.6 beta (2012) software.
AMMI model for
seasonal varietal yield stability testing
The
following model was proposed to measure the seasonal stability of a variety by
phenotypic performances of gth genotype in the sth
season.
Yij=µ
+ œg + ßs + Ø^g ðs + þgs
Where
Yij is the mean yield of the gth genotype in the sth
season; µ is the grand mean; œg is
the gth genotype effect; ßs is the sth seasonal
effect; Ø is the Eigen value of IPCAs; ^g and ðs are the gth genotype, sth
seasonal IPCAs scores; þgs is the residual.
Linear
regression model
The
normal liner regression models have several disadvantage due to environment
index is not independent of the response variable and biased estimators of
regression coefficients as independent variable is measured with errors (Becker
and Leon, 1998; Toller and Burrows, 1998; Storck and Vencovsky, 1994).
Consequently we proposed the following simple linear regression model without
using environment index for stability testing with AMMI model and IPCAs Scores.
Yij
= ß0i+ ß1iX+œij + êij
Where
Yij is the yield of ith
genotype in jth environment, ß0i is intercept, ß1i
is slope, X is season, œij is the deviation from regression and êij is error.
Varietal
selection method for seasonal yield stability
There
are several methods of measuring phenotypic stability of varieties by using
different models of GEI. A non-parametric method that was based on analysis of
variance component of each genotype over the environment was proposed by Shukala
(1972). Univariate parametric methods such as unisegmented and bi-segmented
linear regression are also used to explain the GEI interaction by regressing
the genotypes performance over the environmental yields (Eberhart and Russell,
1966; Finlay and Wilkinson, 1963, Toller, 1990; Toller and Burrows, 1998). AMMI
model use as a multivariate method to study phenotypic stability (Crossa, 1990;
Gauch, 1985; Gauch and Zobel, 1988). Both multivariate and univariate
techniques are useful to identify the stable and adopted genotype (Acciaresi
and Chidichimo, 1999).
A
high yielding promising, seasonal stable lines could be selected by simultaneously
using AMMI model with IPCAs vs. mean yield bi-plot and simple linear regression
especially in the primary yield trials. Selection criteria for selection of seasonal
stable lines was suggested as follows; selection could be done if (1) IPCAs was significant, IPCAs score was close to zero
and regression coefficient (ß1i) less than +5% or higher than -5% of
grand mean yield with coefficient of determination (r2) close to one
and yield was above grand mean, (2) IPCAs was not significant and yield was
above grand mean, (3) varieties recommended only for a specific season using
IPCAs vs. mean yield bi-plot, ß1i and r2 .
RESULTS AND DISCUSSION
The genetic and non genetic
interaction (GEI) on phenotypic expression contributes to non realization of
expected gain from selection (Comstock and Moll, 1963). Static
variability of a variety is best due to it secure constant yield in different
environments. Consequently, ideal varieties should have higher mean yields and
low degree of fluctuations (Tarakanovas and Ruzgas, 2006). Adaptability is define as the
ability of a crop variety to perform well over diverse environment ( Abeysiriwardena
et al., 1991). Therefore,
adoptability to cropping seasons can be express as a crop variety to perform
well over different cropping season under different weather condition.
Cropping seasons and genotypes were
found in highly significant (p=0.01). Genotype x season interaction was also
highly significant. Consequently, the genotypes respond in different way with
the different cropping seasons (Table 1 and 2).
The AMMI analysis is help to understanding the genotype environment
interaction, improving the accuracy of yield estimate, imputing missing data,
increase flexibility and efficiency of experimental designs ( Gauch, 1992;
Gauch and Zobel, 1996). The bi-plot of IPCAs explained how they achieved the
average yield (Abeytilakarathna, 2010).AMMI bi-plot ordinate with IPCAs
Score capture 100% of GEI (Abeytilakarathna, 2010).Abeysiriwardena (2001) used a
method to evaluate varieties for adaptability by estimating the average
varietal superiority by calculating mean deviation from maximum plot yield and
deviation across locations.
Table 1. Analysis
of variance of yield of 7 bush bean lines grown at Rahangala during 2010-2012
|
Source
|
DF
|
SS
|
MS
|
|
Replication
Cropping
seasons (S)
Genotypes
(G)
G x S
|
2
2
6
12
|
7816.67
1702939.95
1834454.94
959856.97
|
39108.34 ns
851469.98 **
305742.49 **
79988.08 **
|
**
Significant at 0.01 probability level, ns= not significant at 0.05 probability
level, DF= degree of freedom, SS= sum of squares, MS= mean
sum of squares
Table 2. Additive
main effect and multiplicative interactions analysis of variance for the yield
of 7 bush bean lines grown at Rahangala
during 2010-2012
|
Source
|
DF
|
SS
|
MS
|
|
Replication
Cropping
seasons (S)
Genotypes
(G)
IPCAs
Residual
|
6
2
6
12
0
|
118454
1702940
1834455
959857
0
|
19742 ns
851470 **
305742 **
79988 **
0
|
**
Significant at 0.01 probability level, ns= not significant at 0.05 probability
level
Table 3.
Mean yields of 7 bush beans lines grown at Rahangala during 2010-2012
|
Lines
|
Mean
yield (g/m2)**
|
IPCAs**
|
|
BB
2904
BB
21-2-3
BB
22-1-2
BB
22-2-1
BB
22-2-3
Cherokee
Wax
Wade
|
1013.33 a
525.74 c
540.93 c
516.48 c
745.00 b
641.85 bc
516.11 c
|
8.55
-0.56
-10.16
-1.20
-6.19
9.92
-0.06
|
|
CV%
|
24.18
|
|
**
Significant at 0.01 probability level
Means
followed by same letters in superscript are not significantly different at 5%
probability level of DMR test.
Table 4.
Mean yields and planting dates of 7 bush beans lines grown at Rahangala during
2010-2012
|
Cropping
Season
|
Date
of planting
|
Seasonal
mean yield (g/m2)**
|
IPCAs**
|
|
2012 yala
2011/12
maha
2010 yala
|
23.08.2013
09.01.2012
19.08.2010
|
875.16 a
519.84 b
533.33 b
|
-5.38
-9.04
14.42
|
|
CV%
|
|
24.18
|
|
**
Significant at 0.01 probability level
Means
followed by same letters in superscript are not significantly different at 5%
probability level of DMR test.
Table 5.
Regression coefficient (ß1i), coefficient of determination (r2)
and 5% level of grand mean yield 7 bush beans lines grown at Rahangala during
2010-2012
|
Lines
|
ß1i
|
r2
|
±5% level of grand mean yield
|
|
BB
2904
BB
21-2-3
BB
22-1-2
BB
22-2-1
BB
22-2-3
Cherokee
Wax
Wade
|
-24.44
-229.10
-337.50
-129.10
-196.90
-35.00
-263.00
|
0.019
0.590
0.998
0.910
0.793
0.019
0.837
|
32.36
32.36
32.36
32.36
32.36
32.36
32.36
|
When consider the seasonal mean
yield, higher yield than the above grand mean was seen in 2012 yala season while lower yield that below the mean were found in 2011/12 maha and 2010 yala seasons (Table 4).The highest mean yield was seen in
BB 2904 bush bean line (1013.33 gm-2). Then BB 22-2-3 was showed
higher yield than above grand mean (745 gm-2). Cherokee wax were also observed a significantly
higher yield (641 gm-2), but it was below the grand mean yield.
Wade, BB 21-2-3, BB 22-1-2 and BB 22-2-1 were found significantly lower yield
than below grand mean (516.11, 525.74, 540.93 and 516.48 gm-2
respectively) (Table 3).
Genotypes that IPCAs close to zero
are more stable. By increasing the IPCAs score cause to increase the unstable
state of genotypes (Abeytilakarathna, 2010). Even though the higher yield were
observed in BB 2904 line, it was unstable with cropping seasons due to higher
IPCAs score ( 8.55). Then the above
grand mean yield was seen in BB 22-2-3 line but it was also unstable with the
cropping seasons (IPCAs=-6.49). Wade was the most stable line for the cropping
seasons (IPCAs= -0.06), followed by BB 21-2-3 line (-0.56). But these two lines
were found below grand mean yield (Table 3 and figure 1). AMMI model have
limitation due to it do not reveal the GEI response pattern. To overcome this, regression model
with AMMI model should used simultaneously to estimate phenotypic stability. BB
2904 was showed the lowest ß1i (ß1i=32.36), but
association with seasons and yield was very poor (r2= 0.019).
Genotype BB 22-1-2, BB 22-2-1 and Wade were observed strong association with
yield and seasons (r2= 0.998). however, its yields were reduced
drastically with the seasons (Table 5).
Crossing wade that was showed more
stable and yield was below grand mean yield, with BB 2904 that was showed high
yielding and unstable, might be beneficial to breed high yielding, more stable
promising line in bush bean crop improvement programmmes. Seasonal stable lines
will be evaluated in multi-location trials to select the most promising
seasonal and location wise stable genotypes.
These selected genotypes might fluctuate little with the climatic
changed situation.
Fig. 1. IPCAs scores for mean yield of 7 bush bean lines grown
at Rahangala in 3 seasons
CONCLUSIONS
Even though, bush bean line, BB 2904 was showed the
highest yield, its yield was highly fluctuated with different seasons. The line
BB 22-2-3 was also observed a higher yield than above grand mean yield. But its
yield was also unstable with seasons. Variety Wade was showed high stable yield
but its lower yield was than grand mean yield like line BB 21-2-3. Lines BB
22-1-2, BB 22-2-3 and variety Cherokee wax and they were unstable with different
seasons. Crossing between Wade that was exhibited stable and lower yield than grand mean with
high yielding unstable line BB 2904 might be effective to breed high yielding
and seasonal yield stable promising line to secure economic yield irrespective
with the changing climatic condition.
REFERENCES
Abeysiriwardena, D.S. de Z, G.R. Buss
and P.F. Reese, Jr. (1991). Analysis of multi-environmental yield trials for
testing adaptability of crop genotype. Tropical Agriculturist. 147:85-97
Abeysiriwardena, D.S. de Z. (2001)
Statistical analysis of on farm yield trial for testing adaptability of rice.
Euphatica. 121(3):215-222.
Abeytilakarathna, P.D.
(2010). Statistical analysis for stability and adaptability testing of mungbean
(Vigna radiata (L.) Wilczek) genotypes. Electronic Jornal of
Plant Breeding. 1(3): 244-249
Acciaresi, H.A. and O.H.
Chidichimo (1999). Genotype environment interaction in Avena sativa L.:
employing AMMI and factorial correspondence models. Pesquisa Agropecuaria
Brasileira. 34(10): 1823-1830
AgStat (2011). Pocket book of
agricultural statistics, Socio economic and planning centre, Department of
agriculture. Peradeniya. pp 27-28
Becker, H.C. and V. Leon
(1998). Stability analysis in plant breeding. Plant Breeding. 101:1-23
Beebe
SE, I.M. Rao and C. Cajiao .(2008) Selection for drought resistance in common
bean also improves yield in phosphorus limited and favorable environments.
CropScience 48: 582–592.
Beebe, S., J. Ramirez, A. Jarvis, I. M. Rao, G.
Mosquera, J.M. Bueno, and M. W. Blair. (2011).
Genetic Improvement of Common Beans and the Challenges of Climate
Change. Crop Adaptation to Climate Change, S. S. Yadav, R. J. Redden, J. L.
Hatfield, H. Lotze-Campen and Anthony E. Hall (Eds.). Blackwell Publishing Ltd,
Oxford, UK pp 356-358
CGIAR (2013). Common bean.
Consultative Group on International Agricultural Research. Available at http://www.cgiar.org/our-research/crop-factsheets/beans/
Comstock, R.E and R.H. Moll (1963).
Genotype x environment interaction. Statistical genetic and plant breeding
NAS-NRG Publication 982:174-196
Crossa, J. (1990).
Statistical analysis of multilocations trials. Advances in Agronomy. 44: 55-85
Eberhart, S.A. and W.A.
Russell (1996). Stability parameters for comparing varieties. Crop Science. 6:
36-40
Finlay, K.W. and G.N.
Wilkinson (1963). The analysis of adaptation in a plant breeding programme.
Australian Journal of Agriculture Research. 14(6):742-754
Gauch, H.G. (1985).
Integrating additive and multiplicative models for analysis of yield trials
with assessment of predictive success. Department of Agronomy, Cornell
University, Mimeo.
Gauch, H.G. (1992). Statistical analysis
of regional yield trials. AMMI analysis of factorial designs. Amsterdam,
Elsevier.
Gauch, H.G. and R.W.Zobel (1998). Predictive and postdictive
success of statistical analysis of yield trials. Theoretical and Applied
Genetics. 76(1): 1-10
Gauch, H.G. and R.W. Zobel (1996). AMMI
analysis of yield. In: Genotype by environment interaction, M.S. Kang and H.G. Gauch,
(Eds.). CRC press. Boca Raton, Florida. pp 85B122
Lal, H.A., K. Muhumad, A.
Muhumad and A. Tariq (2010). Stability analysis for grain yield in mungbean (Vigna
radiata L. Wilczek) grown in different agro-climate regions. Emirate
Journal of Food and Agriculture. 22(6):490-497
Lobell, D.B, M.B. Burke and C. Tebaldi (2008) Prioritizing climate
change adaptation needs for food security in 2030. Science 319: 607–610.
Shukla, G.K. (1972). Some
statistical aspects of partitioning genotype environmental component of
variability. Heredity. 29: 237-245
Storck, L and R. Vencovsky
(1994). Stability analysis based on a bi-segmented discontinuous model with
measurement errors in the variable. Journal of Genetic. 17(1): 75-81
Sunday Times (2013). Can people tackle climate
change, when science remains uncertain? Environment, The Sunday Time Plus on
August 04, 2013, Colombo.
Tarakanovas, P. and V.
Ruzgas (2006). Additive main effect and multiplicative interaction analysis of
grain wheat varieties in Lithunia. Agronomy Research. 4(1):91-98
Toller, J.E. (1990). Pattern
of genotype performance over environmental arrays. Clemson University, Clemson.
p154
Toller, J.E. and P.M.
Burrows (1998). Genotype performance over environmental arrays: a non-linear
grouping protocol. Journal of Applied Statistics. 25(1):131-143
No comments:
Post a Comment