International Journal of Plant Breeding and Crop Science

Vol. 5(1), pp. 361-369, March, 2018. © www.premierpublishers.org. ISSN: 2167-0449

IJPBCS

Research Article

Genotype by Environment Interaction on Yield Components

and Stability Analysis of Elite Cassava Genotypes
*1Rotich D. C., 2Kiplagat O.K., 3Were V. W.
1
Department of Biotechnology University of Eldoret P.O Box 1125-30100, Eldoret, Kenya
2
Head of Department, Agriculture and Biotechnology, University of Eldoret, P.O. Box 1125-30100, Eldoret, Kenya
3
Senior Plant Breeder, Kalro Kakamega, P.O Box 169-50100 Kakameg, Kenya

Newly developed varieties can only contribute to increased productivity if high producing
varieties are released in production niches they are adapted to. In order to enhance adoption of
new improved cassava varieties in western Kenya, a study was conducted to evaluate the effects
of genotype by environment interaction (GEI) on agronomic and farmer preferred traits of cassava
and to asses yield stability of 16 cassava genotypes. The study was conducted in randomized
complete block design with three replications across five different environments of western
Kenya. AMMI analysis of variance identified highly significant (P= 0.001) GEI effects for plant
height, height at first branching, and fresh root yield. Generally, GEI effects accounted for 14.98%,
24.64% and 28.3% variability in PH, HB, and FRY respectively. GGE biplot analysis shows that
MM06/0138, MM96/9308, MM97/0293, MM98/3567, MM06/0074, MM96/4271 were high yielding and
stable genotypes. AMMI stability value revealed that genotype MM06/0143 combined high stability
for plant height, height at first branching, number of storage roots and fresh root yield. Genotypes
MM06/0138, MM98/3567, MM96/9308, MM97/0293, and MM06/0074 outperformed the check in
storage roots yield exhibited high yields in farmer preferred traits and were classified as stable
genotypes. Therefore, recommended for release to farmers.

Key Words: Elite cassava, Farmer preferred traits, Genotype X Environment interaction, AMMI analysis, GGE-
biplot analysis

INTRODUCTION

Cassava (Manihot esculenta Crantz) is an important staple
grown for its starchy tuberous roots. Its roots and leaves
are suitable for human consumption as well as animal
feed. The tuberous roots are an important source of
carbohydrates while the leaves are cheap valuable source
of proteins, minerals and vitamins A, B and C (Montagnac
et al., 2009). The storage roots are also used as industrial
raw materials like starch extractions for various industrial
uses, breweries, pharmaceutical, and biofuel among other
uses (Nweke, 2004; Jackson et al., 2014). Cassava is the
second most important food crop after maize in Western
and coastal regions of Kenya (Njeru & Munga, 2003).
However, production level in Kenya is 11 t/ha, below the
potential of 90 t/ha, which is attributed to low yield of
popular varieties, poor access to quality planting material,
lack of well adapted varieties, pests and diseases
(Mwango’mbe, et al., 2013). Cassava mosaic disease
(CMD) and cassava brown streak disease (CBSD) are the
leading yield limiting biotic constraints for cassava
production causing an estimated loss of more than US$ 14
million per annum in CMD (Alabi, et al., 2015). Cassava
genetic improvement has been difficult due to the biology
of the crop (Ceballos et al., 2004).
*Corresponding Author: Rotich C. Damaris, Department
of Biotechnology University of Eldoret P.O Box 1125-
30100, Eldoret Kenya. Email: (rotich.damaris@gmail.com
Tel: +254703276027. Co-Authors Email:
2
kiplagatoliver@yahoo.com, Tel: +254723967672,
3
vwoyengo@gmail.com, Tel: +254729981023

Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
Rotich et al. 362
Use of improved varieties is the current leading tool for
solving viral disease challanges in cassava (Alabi, et al.,
2015), hence continuous deployments of elite resistant
cultivars are necessary as CMDs are known to evolve
producing virulent strains while different strains of CBSD
are being reported in Kenya (Mware, 2009). However, low
adoption of the new improved varieties has been reported
in western Kenya (Odendo et al., 2010; Woyengo & Omari,
2014).
In the last three decades, cassava breeding has majorly
concentrated on increasing yields and resistance to pest
and disease (Ceballos et al., 2004). However, there has
been lack of focus on farmer preferred traits by breeding
programmes which has been observed to be the major
course of low adoption of improved varieties despite the
high yield and resistance to common pests and diseases
(Woyengo, 2011). Moreover, some of the improved
varieties fail to perform well in target production niches due
to lack of detailed stability studies of these traits. It’s
therefore a prerequisite that cassava varieties should not
only be released on the basis of average yield and reaction
to diseases pests but also on the presence of farmer
preferred traits and stability. The work of Achepong et al.
(2013), identified longevity and disease resistance as two
major attributes of cassava that influence adoption of
improved varieties in Ghana. Njukwe et al. (2013)
observed regional differences for farmer preference in
cassava attributes and cassava genotypes in Cameroon,
for instance farmers in Ebolowa and Bertoua preferred
leafy, sweet roots and early branching varieties while
those in Bamenda and Ngaoundere preferred tall, drought
tolerant and in some cases flowering varieties. In Kenya
the work of Were (2011), identified farmer preferred traits
that encourage adoption of improved cassava genotypes
by order of preference as high root yield, tall plants and
lower height of first branching. However, no study has
been reported on influence of environment on farmer
preferred traits and stability of new improved cassava
genotypes.
Genotype stability and adaptability are ultimate resources
for achieving food security which is an allusive goal for
Kenya and Sub-Saharan Africa at large (Muzari et al.,
2012). Lobell (2009) stated that agricultural adaptability
should be a priority in meeting food security presently and
in future in the face of sever climate change. This is
achieved by development of stable varieties of crops.
Stability of performance of quantitative traits is influenced
by genotype, environment and genotype by environment
interaction (GEI) effects. GEI is important in plant breeding
because it complicates demonstration of a superiority of a
variety. An effective method which has been used to
reduce GEI is stratification of environment such that the
sub-region in which the breeder is developing improved
varieties are somehow similar (specific adaptation).
However, this is not mineable to breeding since even with
the refinement of this technique the interaction of
genotypes within a location in a sub-region and with
environments encountered at the same location in different
Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
years frequently remains too large (Crossa 1990).
Moreover, Woyengo and Omari, (2014) clearly pointend
out that it is not feasiable to breed for specific adaptation
with current eratic climatic conditions and effect of
climimate change hence breeding for stable varieties
remains as the only viable option. Many statistical
procedures have been advocated for the basis of analysis
of GEI and stability of genotypes. Studies show that
stability of performance are expected to become more
relevant issues as greater emphasis is placed on
sustainability of agricultural systems (Kang et al., 2012).
The objective of this study therefore, was to evaluate GEI
effects on agronomic and farmer preferred traits of
cassava and to assess yield stability of 16 cassava
genotypes across five environments of western Kenya.
MATERIALS AND METHODS
The study was conducted across five environments
(Kakamega, Sang’alo, Alupe, Kibos and Migori), which
represent major cassava growing zones of western Kenya,
between 2014 and 2015. The experimental material
consisted of 15 elite cassava clones (G1=MH95/0183
G3=MM06/0013, G4=MM06/0046, G5=MM06/0074,
G6=MM06/0082, G7=MM06/0083, G8=MM06/0131
G9=MM06/0138, G10=MM06/0139, G11=MM0H6/0143,
G12=MM96/2480, G13=MM96/4271, G14=MM96/9308,
G15= MM97/0293, G16=MM98/3567) in advanced stage
of yield trials performance and one local check
(G2=migyera). Improved clones’ seeds were developed
and introduced from International Institute of Tropical
Agriculture (IITA). The clones were derived from half-sib
progenies of elite varieties. They have been tested by
Kenya agricultural livestock and research organization
(KALRO) Kakamega for resistance to CMD and CBSD.
The experiment was laid in randomised complete block
design with three replications and established under rain
fed conditions. No fertilizer nor pesticide were applied.
Each experimental plot had six rows and 30 plants spaced
of 1m by 1m between plants and rows.
Data Collection
Data was collected on agronomic traits (Dry matter content
and starch content) and traits preferred by farmers in
western Kenya as identified by Woyengo, (2011) which
include plant height, height at first branching, number of
storage roots per plant and fresh root yield. Data was
collected at twelve months after the date of planting and at
harvesting. Data were recorded on plant and plot basis.
Five plants per plot were sampled from the inner rows.
Data were recorded on plant height (PH), height at first
branching (HB), number of storage roots per plant
(NSR)and fresh root yield converted to tonnes per hectare
(FRY). Data on dry matter content (DMC) and starch
content was collected according to the methodology
described by Fukuda et al., (2010) converted to DM% and
Starch % as follows:
 Int. J. Plant Breed. Crop Sci. 363

Table 1: Agro-ecological description of experimental sites

Elevation Rainfall Temperature
Site Longitude Latitude Soil texture
(m.a.c.l) (mm) (Range 0C)
Kakamega 34047'E 00017’N 1554m 1191 (18.5-21.0) Red friable Nitosols
Sang'alo 3406'E 0005'N 1421m 1628 (20.9- 22.0) orthic ferralsols
Alupe 3407'E 00029’N 1173m 1627 (21.0-22.2) clay-loam Acrisols
Kibos 34048'E 00004’N 1690m 1912 (15.3-30.0) Black clay vertisols
Migori 34031'0E 00059'S 1423m 1396 (20.4-21.7) Mollic Nitosols

Root sample weighing 3-5kg was prepared, the weight of
the sample was measured in air (Wa) using a digital
weighing balance. The same sample was also measured
in water (Ww). Specific gravity(x) was computed at:
Ww/ (Wa-Ww)
%DMC was computed using the formula DMC = (158.3x-
142)100
%starch was computed using the formula starch = (112.1x-
106.4)100
Another important point was further explained by Yan et al.
(2007) that genotype and genotype-by-environment
effects must be considered simultaneously to make a
meaningful decision in selection. Significant genotype by
environment interaction was also analyzed by GGE biplot
which was also useful in ranking genotypes based on their
average performance and stability for farmer preferred
traits in cassava. The model for the GGE biplot based on
singular value decomposition (SVD) of first two principal
components is:
Yij    j  1i1 j1  2i2 j2 ij

Statistical Analysis Where: Yij = measured mean of genotype i in environment

Genotypic stability for each clone was computed using
GenStat software, 14th edition. The additive main effects
and multiplicative interactions (AMMI) statistical model
suggested by Gauch and Zobel (1996) was used to
analyze yield data to obtain (AMMI) analysis of variance
and (AMMI) mean estimates as follows as follows:
Yger = µ + αg +βe + ∑ʎn ygn δen + ρge + Eger
j, = grand mean, j = main effects of environment j,  + j
= the mean yield across all genotypes in environment j, 1
and 2= are the singular values (SV) for the first and
second principle components (PCA 1 and PCA 2)
respectively. i1 and i2 = are eigenvectors of genotype i for
PCA 1 and PCA 2 respectively; j1 and j2 = eigenvectors
for environment j for PCA 1 and PCA 2, respectively. ij =
residual associated with genotype i in environment j

Where: Yger = yield of genotype g in environment e for RESULTS AND DISCUSSION

replicate r, μ = grand mean, αg = genotype mean deviation
(genotype means minus grand mean), βe = environment
mean deviation, n = number of principal component
analysis (PCA) axes retained in the model, ʎn = singular
value for PCA axis n,ygn = genotype eigenvector values for
PCA axis n, δen = environment eigenvector values for PCA
axis n, ρge = residuals, Eger = error term.
The AMMI stability value (ASV) proposed by Purchase et
al. (2000) was used to quantify and rank genotypes
according to the yield stability. The ASV has been defined
as the distance from the coordinate point to the origin in a
two-dimensional scatterplot of first interaction principal
component axis (IPCA1) scores against the second
interaction principal component axis (IPCA2) (Farshadfar
et al., 2012). Since IPCA1 accounts for most of the GEI
variation, the IPCA1 scores are weighted by the ratio of
IPCA1SS (from AMMI ANOVA) to IPCA2 SS in the ASV
formula as follows:
2 to environments which necessitates the investigations of
𝑆𝑆𝐼𝑃𝐶𝐴1
𝐴𝑆𝑉 = √{ (𝐼𝑃𝐶𝐴1 𝑠𝑐𝑜𝑟𝑒)} + (𝐼𝑃𝐶𝐴2 𝑠𝑐𝑜𝑟𝑒)2
𝑆𝑆𝐼𝑃𝐶𝐴2
The lower the ASV, the more stable a genotype is.
The results from AMMI analysis of variance (Table 2)
reveal that environment gives the most effect (64.81 %) of
variability on plant height. Moreover, environmental
variation contributed to more than 47% of the total
variability in fresh root yield. Aina et al. (2007) in Nigeria
reported 88.9% of environment sum of squares (SS) when
evaluating for root yield stability in cassava. The large sum
of squares for environments indicated that the
environments were diverse, with differences among the
environmental means causing more than a third of the
variation in plant height, height at first branching and
number of storage roots. This might be probably due to the
differences in environmental conditions which has been
known to have impact on cassava yield (De Vries et al.,
2010). Besides, highly significant GEI interaction (P≤
0.001) was observed for PH, HB, and FRY. These results
are in agreement with the findings of Adjebeng-Danquah
et al. ( 2017) who observed highly significant GEI for these
trait.These suggests that genotypes responded differently
the nature of different response of the genotypes to
environments. GEI effects contributed with 14.98%,
24.64% and 28.3% for plant height, height at first

Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
Rotich et al. 364

Table 2: AMMI analysis of variance for16 cassava clones evaluated across five agroecological zones of Western Kenya
PH (cm) HB(cm) NSR FRY (t/ha) DMC (%) STARCH
Source Df MS %SS MS %SS MS %SS MS %SS MS %SS MS %SS
Total 239 2923 557 10.37 100.3 11.24 73.8
Treatments 79 8567*** 1584*** 28.72*** 288.9*** 20.84*** 162.5***
genotype 15 9119*** 20.20 4247*** 50.9 47.29*** 64.83 370.6*** 24.35 75.37*** 68.67 661.9*** 77.33
Environ 4 109651*** 64.81 7651*** 24.46 252.19*** 23.07 2701.7*** 47.35 19.76* 4.84 422.6*** 13.16
Block 10 474ns 43ns 1.05ns 7.1ns 7.08ns 19.7ns
GEI 60 1690*** 14.98 514*** 24.64 9.18ns 12.09 107.7*** 28.30 7.28ns 26.52 20.3ns 9.5
IPCA 1 18 3172*** 56.31 1132*** 66.13 15.18ns 49.59 178.7*** 49.79 15.25ns 30.00 37.5ns 55.46
IPCA 2 16 2104*** 33.2 524*** 27.20 9.49ns 27.55 104*** 25.72 5.22ns 26.67 18.4ns 24.2
IPCA 3 14 486*** 6.72 125** 5.69 6.64ns 16.86 84.8*** 18.38 4.68ns 23.33 11.7ns 13.45
IPCA 4 12 319** 3.77 25ns 0.98 2.74ns 6.00 32.7*** 6.11 1.10ns 20.00 7.1ns 6.97
Residuals 0 0 0
Error 150 114 50 1.33 7.1 6.46 30.6
*** Significant at (P≤0,001), ** (P≤ 0.01), *(P≤ 0.05) and ns = not significant respectively; PH = plant height, HB = plant height at first
branching, NSR = number of storage roots per plant, FRY =Fresh root yield, DMC = dry matter content, SS=% sum of squares,
IPCA=interaction principle component.

Table 3: Ranking of 16 cassava genotypes according to their AMMI stability value evaluated for PH, HB, SR and FRY
Genotype PH R HB R NSR R FRY R
MH95/0183 3.57 9 5.29 9 0.39 3 4.53 15
MIGHERA 9.32 14 7.59 13 3.1 16 4.01 14
MM06/0013 3.24 7 2.1 4 0.88 8 2.96 10
MM06/0046 1.07 1 1.5 2 1.04 9 2.84 9
MM06/0074 4.63 12 9.5 15 0.65 7 0.25 1
MM06/0082 9.83 15 3.35 6 1.6 11 3.75 13
MM06/0083 2.11 3 0.51 1 2.96 15 1.51 3
MM06/0131 13.4 16 6.86 12 0.5 5 5.5 16
MM06/0138 5.02 13 4.98 8 1.91 13 1.94 6
MM06/0139 2.6 5 10.5 16 1.65 12 3.61 12
MM06/0143 1.98 2 1.65 3 0.35 1 2.05 7
MM96/2480 4.1 11 3.73 7 1.36 10 2.49 8
MM96/4271 3.64 10 2.44 5 0.44 4 3.4 11
MM96/9308 3.3 8 6.8 11 0.62 6 1.55 4
MM97/0293 2.66 6 5.96 10 0.35 2 1.08 2
MM98/3567 2.41 4 7.76 14 2.32 14 1.83 5
Plant height (PH), height at first branching (HB), number of storage roots per plant (NSR) and fresh root yield (FRY)

branching and fresh root respectively, meaning more than
24% variability observed in height at first branching and
fresh root yield is due to GEI effects. In spite of this, the
magnitude of the genotype sum of squares for plant height
and height at first was branching was larger than that of
GEI (20.2% and 50.9%) respectively which indicates
presence of moderate control of genotype effects over
genotype by environment interaction effects for these
traits.
On the other hand, there was non-significant GEI effects
for number of storage roots, dry matter content and starch.
The phenomenon was also the same as reported by
Peprah et al. (2013) who observed non-significant GEI for
dry matter content. This finding also agrees with those of
Aina et al. (2007) who reported non-significant GEI effects
for number of storage roots and dry matter content.
Similarly, Benesi et al. (2004) reported non-significant GEI
for starch.
An obvious deduction from non-significant GEI effects on
number of storage roots, dry matter content and starch is
that, genotypes might have similar responses across the
Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
locations in which they were evaluated and that they can
consistently be evaluated under any of the locations used
for this study in impending performance trials. This view is
in conformity with the view of Peprah et al. (2013) who
found non-significant GEI effects for dry matter content
and reported that fewer environments may be needed to
distinguish clones with high and stable performance for
this trait. In other words, evaluating genotypes for these
traits concurrently in the various locations used for these
studies in consequent evaluation trials might not be
important. Thereby, offering an opportunity to manage
inadequate means available for testing programme (Tonk
et al., 2011).
Complementary to previous results, ASV was computed
for the traits so as to quantify and rank genotypes
according to their stability Table 3 shows the ranking of the
16 genotypes according to AMMI stability value.
Genotypes varied in ranking for stability across the studied
traits. However, some genotypes combined satisfactory
results for stability in various traits. For instance, genotype
MM06/0143 is very stable as it ranked 2 nd, 3rd, 1st, and 7th
for plant height, height at first branching, number of
 Int. J. Plant Breed. Crop Sci. 365

storage roots and fresh root yield. MM98/3567 combines
stability for plant height and fresh root yield while
MM97/0293 combines stability for plant height, number of
storage roots and fresh root yield. Genotypes MM06/0046,
MM06/0083, MM06/0143, and MM06/0074 were stable
genotypes for plant height, height at first branching,
number of storage roots per plant and fresh root yield
respectively. These suggest the possibility of identifying
clones exhibiting stable performance in both agronomic
and farmer preferred traits.
concentrated around zero for most of the clones
suggesting that the candidate clones were stable for this
trait (number of fresh root yield). Though clones G2
(migyera), G11 (MM0h6/0143) and G4 (MM06/0046)
revealed some displacement on the Y-axis from the origin,
it’s clear that the PC scores for these clones in this Y’-axis
is less than one (near zero) for the three clones hence it
was adjudged that all the candidate clones were stable for
this trait.

GGE Biplot analysis

According to Yan et al. (2007) genotype-by-environment

interaction effects must be considered simultaneously to
make a meaningful decision in selection. These requires
biplot analysis that considers genotype and GEI
simultaneously. Additionally, genotypes should be
evaluated based on combined performance of the mean
across environments with their stability which also
necessitates the use of a biplot analysis.

Analysis of GGE biplot further elucidated the yield

performance and stability of genotypes across the study
sites for fresh root yield and number of storage roots.
Analysis of yield stability of genotypes was evaluated
using GGE biplot by an average environment coordination
(AEC) method on fresh root yield and number of storage
roots. In this method, the average principle components
are used in all environments, as depicted in Figures 1 and
2. The AEC ordinate separate genotype with below
average means from those with above average means. A
line is then drawn through this average environment axis
and serves as the abscissa of the AEC.
The arrow points to a greater genotype main effects, the
AEC ordinate and either direction away from the biplot
origin indicates greater GEI effects and reduced stability.
Hence, the stability of a variety or environment was
determined by the length of the vector from genotype
marker to the average environment coordinate (AEC)
abscissa. The vector which was closer to the AEC
abscissa was considered to have less interaction effects
and hence regarded as stable. A clone located at the origin
is not influenced by environment in any way hence it would
rank the same in all the environments and therefore
considered as the most stable.
In Figure 1 the mean number of storage root per plant and
stability performance of cassava genotypes was depicted.
The genotypes were ranked along the average
environment co-ordinate (AEC) x-axis with an arrow
indicating the highest mean. Thus, results revealed that
G16 (MM98/3567) which was closer to the AEC had the
highest number of storage roots while G4 (MM06/0046)
was the lowest yielding genotype because they were
further away from the AEC axis. However, the lengths of
vectors from genotype marker to the AEC abscissa
Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
Figure 1: Mean performance and stability of 16 cassava
genotypes (G1 – G16) at five environments (K:Kakamega,
S:Sangalo, A:Alupe, B:Kibos, M:Migori) for number of
storage roots.
Ranking of genotypes along (AEC) for fresh root yield
revealed that candidate clones varied greatly in yield
performance and stability across the study sites.
Generally, G16 (MM98/3567) was the highest yielding
genotype because it was located closer to the (AEC) while
G4 (MM06/0046) was the lowest yielding genotypes
because it was located further away from (AEC). G9
(MM06/0138), G14 (MM96/9308), G15 (MM97/0293), G16
(MM98/3567) and G5 (MM06/0074) are high yielding and
stable depicted by shortest projection of genotype vectors
from the AEC axis and having less than 0.2 values along
the Y-axis. G12 (MM96/2480), G10 (MM06/0139) and G3
(MM06/0013) were also closer to zero-line value on the Y-
axis, had positive values above zero on X-axis and hence
were considered as high yielding with average stability.
G13 (MM96/4271) was stable but below average in yield.
G1 (MH95/0183), G8 (MM06/0131), G6 (MM06/0082) and
G4 (MM96/9308) are low yielding and unstable depicted
by longest projections of genotypes vectors from AEC axis
(Figure 2).
Rotich et al. 366
Figure 2: Mean performance and stability of 16 cassava
genotypes (G1-G16) at five environments (K:Kakamega,
S:Sangalo, A:Alupe, B:Kibos, M:Migori) for fresh root yield.
The polygon view of the GGE biplot explicitly displays the
’’which won where pattern’‘ and hence is a concise
summary of the GEI pattern (Figures 3 and 4) .The polygon
is formed by connecting the markers of the genotypes that
are further away from the biplot origin such that all the
genotypes are contained in the polygon (Yan et al., 2007;
Akinwale, 2011).Convex-hull are drawn from the biplot
origin which divides the biplot into sectors that demarcate
mega-environment the vertex genotypes in a sector of
environment are considered the most stable for that
environment. In Figure 3, the “which won where pattern” of
the GGE biplot on number of storage roots grouped all the
environments in one sector. Moreover, the genotypes
clustered around the origin of the biplot revealing that the
genotypes had the same response across the
environments. Generally, the GGE biplot on number of
storage roots accounted for 95.74% of the total GEI
variation due to GEI effects on number of storage roots
with PC1 and PC2 accounting for 87.52% and 8.22%,
respectively.
Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
Figure 3: “Which won where pattern” of GGE biplot for
sixteen clones (G1-G16) at five environments (K:Kakamega,
S:Sangalo, A:Alupe, B:Kibos, M:Migori) on number of
storage roots.
In Figure 4, the ‘‘which won where” pattern of the GGE
biplot on fresh root yield explained 77.15% of the total
variation due to GEI effects, PC1 accounted for 51.7%
while PC2 accounted for 25.45%. The biplot revealed the
best genotypes across environments and identified the
best clones with respect to site. The seven rays that divide
the biplot into seven sectors to which five environments fall
into two of them showed that (Alupe, Sanga’lo and Kibos)
environments fall into sector one and the vertex genotypes
for this sector was G16 (MM98/3567) Similarly, two
environments (Kakamega and Migori) fell into sector two
and the vertex genotypes for this sector was G9
(MM06/0131). No environment fell into sectors with G10
(MM06/0131), G4 (MM06/0046), G2 (migyera) and G8
(MM06/0131) as vertices indicating that these cultivars
were unstable in all the environments. Genotypes G5
(MM06/0074) and G13 (MM96/4271) were located at the
origin of the biplot revealing that they were highly stable
clones across the sites.

Figure 4: “Which won where” pattern of GGE biplot for 16
cassava genotypes (G1-G16) at five environments on fresh
root yield
Another important feature of GGE biplot analysis is its
ability to evaluate test environments for effective selection
of superior genotypes (Yan et al., 2007). In Figure 5 the
discriminatory power of the environments was detected by
the length of the vector from the origin of the GGE biplot to
the coordinate of the location. The length of the vectors
approximates the standard deviation within respective
environments which is a measure of the discriminating
ability of the environments (Yan, 2005). The longer the
vector, the more discriminatory power. Migori
environments was the most discriminating but the least
representative environment having the long vector length
from biplot origin with large absolute PC2 scores and large
PC1 scores. Contrarily, Kakamega environment was the
least discriminative and the most representative
environment based on short vector length from the origin
and having large absolute PC2 scores and small PC1
scores. However, Kibos environment was considered as
ideal environment for selection of superior clones based
on its discriminating ability and representativeness.
Noerwijati et al. (2013) identified Kediri environments as
ideal for selection of superior cassava genotypes based on
the discriminating and representative view of the GGE
biplot having small absolute PC2 scores and large PC1
scores. Likewise, Agyeman et al. (2015) identified (PK08)
as the ideal environment having a small angle
(representativeness) to the average environment axis and
a long vector length from the biplot (discriminating ability)
in a cassava study using GGE biplot analysis. Generally,
if financial limitations allow only few test environments
Kibos should be the first choice. Migori environments
cannot be used in selecting superior genotypes, but it is
useful in ‘culling’ unstable genotypes.
Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
Int. J. Plant Breed. Crop Sci. 367
Figure 5: Discriminating power of the five environments for
16 cassava genotypes (G1 to G16)
CONCLUSION
Genotype by environment interaction was significant for
fresh root yield, plant height, and height at first branching
indicating the need of assessing genotypes for stability
and adaptability before effective selection can be done. Six
genotypes (MM98/3567, MM06/0138, MM96/9308,
MM97/0293, MM06/007, and MM96/4271) were classified
as stable and outperformed the check cultivar in fresh root
yield across five environments of western Kenya. On the
other hand, number of storage roots, dry matter content
and starch are not influenced by GEI, this implies that
evaluation of genotypes for these traits can effectively be
done in a single location and variety selection can
effectively be done based on the mean performance of
genotypes.
ACKNOWLEDGEMENT
Exceptional appreciations to Alliance for Green Revolution
in Africa (AGRA) foundation for funding this study. The
assistance in field activities provided by the technical staff
of Kenya Agricultural Livestock and Research organization
Kakamega, Mr. Njaro, Mr. Otiya, James and Madam
Linnet is acknowledged and appreciated.
CONFLICT OF INTEREST
There is no conflict of interest for this paper
Rotich et al. 368
REFERENCES
Acheampong, P.P., Owusu, V. and Nurah G.K. (2013).
Farmers Preference for Cassava Variety Traits:
Empirical Evidence from Ghana. Contributor Paper
Prepared for Presentation of African Association of
Agricultural Economics. Hammet, Tunisia.
Adjebeng-Danquah, J., Manu-Aduening, J., Gracen, V. E.,
Asante, I. K., & Offei, S. K. (2017). AMMI stability
analysis and estimation of genetic parameters for
growth and yield components in cassava in the forest
and Guinea Savannah ecologies of Ghana.
International Journal of Agronomy, 10. doi:
https://doi.org/10.1155/2017/8075846
Agyeman, A., Parkes, E., & Peprah, B. (2015). AMMI and
GGE biplot analysis of root yield performance of
cassava genotypes in the forest and coastal ecologies.
International Journal of Agricultural Policy and
Research, 3(3), 222-232. doi: http:
//dx.doi.org/10.15739/IJAPR.034
Aina, O., Dixon, A., & Akinrinde, E. (2007). Additive main
effects and multiplicative interaction (AMMI) analysis
for yield of cassava in Nigeria. Journal of Biological
Sciences, 7(5), 796-800.
Akinwale, M. G., Akinyele, B. O., & Dixon, A. G. (2011).
Genotype X Environment Interaction and Yield
Performance of 43 Improved Cassava (Manihot
esculenta Crantz) Genotypes at Three Agro-climatic
Zones in Nigeria. British Biotechnology Journal, 1(3),
64-68.
Alabi, J., Legg, J., & Mulenga, R. (2015). Virus diseases of
tropical and subtropical crops. In P. Temnant, & G.
Fermin (Eds.). CAB International.
doi:10.1079/9781780644264.0042
Benesi, R. M., Labuschgne, M. T., & Dixon, A. G. (2004).
Genotype X Environment interaction effects on native
cassava starch use in the commercial sector. African
Crop Science Journal, 12(3), 205-216.
Ceballos, H., Igelesias, C. A., Pérez, J. C., & Dixon, A. G.
(2004). Cassava breeding: opportunities and
challenges. Plant Molecular Biology, 56, 503-516. doi:
10.1007/s11103-004-5010-5
Crossa, J. (1990). Statistical analyses of multilocation
trials. 44, 55-85. Advances in Agronomy, 44, 55-85.
De Vries, S. C., Van de ven, G. W., Van Ittersum, M. K., &
Giller, K. E. (2010). Resource use efficiency and
environmental performance of nine major biofuel crops,
processed by first-generation conversion techniques.
Biomass and Bioenergy, 34, 588-601.
Farshadfar, E., Mahamodi, N., & Yaghotipoor, A. (2012).
AMMI stability value and simultaneous estimation of
yield and yield stability in bread wheat (Triticum
aestivum L.). Australian Journal of Crop Science, 5,
1837-1844.
Gauch, H., & Zobel, R. W. (1996). AMMI analysis of yield
trials. In M. S. Kang, & H. Gauch (Eds.). USA: CRC
Press.
Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya
Jackson, B., Oladipo, N., & O., A. M. (2014). Cassava: A
potential crop for industrial raw material in Nigeria.
International Journal of Life Science, 3(3), 105-112.
Kang, M. S., Prabhakaran, V., & Mehra, R. B. (2012). Plant
Breeding: Mendelian to Molecular Approaches (Vol.
23). (H. K. Jain, & M. C. Kharkwal, Eds.) New Delhi,
India: Springer.
Lobell, D. (2009). Prioritizing climate change adaptation
needs for food security in 2030. Science, 319(5863),
607-610. doi:10.1126/science.1152339
Montagnac, A. J., Christopher, R. D., & Tanumihirdjo, S.
(2009). Nutritional value of cassava for use as a staple
food and recent advances for improvement.
Comprehensive Review in Food Science and Food
Safety, 8(3), 186-194. doi:10/1111/j.1541-4337.2009.
00077.x
Muzari, W., Gatsi, W., & Muvhunzi, S. (2012). The impacts
of technology adoption on smallholder agriculture: A
review. Journal of Sustainable Development, 5(8), 69-
77.
Mwango’mbe, A. W., Mbugua, S. K., Olubayo, F. O.,
Ngugi, E. K., Mwinga, R., Munga, T., & Muiru, W. M.
(2013). Challenges and Opportunities in cassava
production among the rural households in Kilifi County
in the coastal region of Kenya. Journal of Biology,
Agriculture and Healthcare, 3(10), 3033.
Mware, B. A. (2009). Transmission and distribution of
cassava brown streak virus disease in cassava growing
areas of Kenya. Journal of Applied Biosciences, 864-
870.
Mware, B. O., Narla, R., Amata, R., Olubayo, F., Songa,
J., Kyamanyua, S., & Ateka, E. M. (2009). Efficiency of
cassava brown streak virus transmission by two
whitefly species in coastal Kenya. Journal of general
and molecular Virology. 1(4), 40-45.
Njeru R. W. and T.L. Munga. (2003). Current status of
Cassava Brown Streak Disease in Kenya. In Legg J.P
and Hillocks, R.J. (eds) Cassava Brown Streak Virus
Disease: Past, Present and Future. Proceedings of an
International workshop, Mombasa Kenya, 27-30
October 2002. P12-13. Natural Resources
International, Aylesford.
Njukwe, E., Hanna, R., Kirscht, H., Araki, S. (2013).
Farmers Perception and criteria for cassava variety
preference in Cameroon. African Study Monographs,
34(4), 221-234.
Noerwijati, K., & Budiono, R. (2013). Yield and Yield
Components Evaluation of Cassava (Manihot
esculenta Crantz) Clones in Different Altitudes. Energy
Procedia, 65, 155-161.
Nweke, F. I. (2004). New challenges in the cassava
transformation in Nigeria and Ghana.International Food
Policy Research Institute. Washington, D.C
Odendo, M., Abele, S., Obiero, H., & Htawuruhanga, P.
(2010). Adoption and Impact of Improved Cassava
Varieties in Western Kenya. Retrived May 10, 2016
from:
 Int. J. Plant Breed. Crop Sci. 369

<http://www.kainet.co.ke/fileadmin/publications/10thpr Yan, W., & Kang, M. S. (2003). GGE biplot analysis: A

oceedings/Poster/Adop_Im act_Cassava.pdf> graphical tool for breeders, geneticists and
Peprah, B. B., Ofofi, K., Asante, I. K., & Parkes, E. (2013). agronomists. Boca Raton: FL CRC press.
Performance of nine cassava (Manihot esculenta Yan, W., Kang, M. S., Baoluo, M., Woods, S., & Baoluo,
Crantz) clones across three environments. Journal of M. G.-b.-e.-6. (2007). GGE biplot Vs. AMMI analysis of
Plant Breeding and Crop Science, 5(4), 48-53. genotype-by-environment data. Crop science, 641-653.
Purchase, J., Hatting, H., & Van Deventer, S. (2000). doi:10.2135/Cropsci2006.06.0374
Genotype x environment interaction of winter wheat
(Triticum aestivum L.) in South Africa: II. Stability
analysis of yield. South African Journal of Plant and
Soil, 17, 101-107.
Tonk, F., Ilker, E., & Tosun, M. (2011). Evaluation of
genotype x environment interaction in maize hybrids Accepted 3 March 2018
using GGE biplot analysis. Crop Breeding and Applied
Biotechnology, 11(1), 01-09. Citation: Rotich D.C., Kiplagat O.K., Were V. W. (2018).
Woyengo, V. W. (2011). A comparative analysis of Genotype by Environment Interaction on Yield
conventional and participatory cassava breeding Components and Stability Analysis of Elite Cassava
approaches in western Kenya. Doctorial dissertation, Genotypes. International Journal of Plant Breeding and
university of KwaZulu-Natal, South Africa. Crop Science 5(1): 308-316.
Woyengo, V. W., & Omari, O. M. (2014). Genotype X
environment interaction effect on farmer preferred traits
of cassava varieties adapted to the tropical climate
conditions of Western Kenya. American Journal of
Copyright: © 2018 Rotich et al. This is an open-access
Experimental Agriculture, 686-702.
article distributed under the terms of the Creative
doi:10.9734/AJEA/2014/6271
Commons Attribution License, which permits unrestricted
Yan, W (2005). Use of biplot analysis in crop breeding.
use, distribution, and reproduction in any medium,
Proceedings of the Eastern Wheat workers and
provided the original author and source are cited.
Southern Small Grain Conference (pp. 7-30). Greek:
Bowling.

Genotype by Environment Interaction on Cassava Components and Yield Stability Analysis of Elite Cassava Genotypes in Western Kenya