Molecular Plant Breeding 2016, Vol.7, No.11, 1
-
16
11
inferred to cluster 3. It is obvious that inferring the 58
genotypes from Afabet into three clusters indicates
that they are from three different ancestry. However,
the findings of Votava et al., (2005) can explain the
relationship among individuals in each group. In
Afabet 85% of the farmers prefer to save and use their
own seed and they grow pepper in small plots
(Maximum of 2 ha and minimum of 0.015) usually
adjacent to each other (Saleh et al., 2013). Thus the
genotypes that cluster together have either a relatively
narrow gene pool or they are derived from the same
sources, and now share sufficient genetic
characteristics to cluster together due to growing in
close proximity (Votava et al., 2005). The case of
NRSAF19 B, NRSAF11 A and NRSAF11 C is mostly
a migration situation.
Gonz
á
lez-P
é
rez et al. (2014) reported that grouping
of genotypes from Spain was according to fruit shape
and size which was not in agreement with the current
study. Red long, Red short, Gahteilay, Chocolate,
Group1A and Group1B are breeding lines from NARI.
The first four are of triangular shaped and for dry
consumption while the last two are with elongated
slim fruits and used for fresh consumption. Red long,
Gahteilay and chocolate are similar in fruit shape and
size but differ in fruit color. Due to the differences in
fruit characteristics, it was expected that individuals
for the dry consumption group would cluster together
and those of fresh consumption group cluster
separately or with other genotypes of similar
characters. However, one variant of Red long (Red
long A) and Group 1A clustered in the same
sub-cluster while, the rest two variants (Red long B &
C), Chocolate and Group 1B A clustered in another
sub-cluster, and Gahteilay and Red short grouped in a
separate sub-cluster. A similar condition was observed
with HAC genotypes. HD0083A &B and HD0023B
(Dry) grouped with HG0018A (Fresh) in a sub cluster
far away from all other HAC genotypes. This
indicates that fruit size, shape and color had no effect
in genotype clustering in the current study. The first
reason could be due to SSR markers measuring
genetic variation mainly in non-coding area, which
has a minor impact on phenotypic characters (Kwon et
al., 2005). Secondly, association between marker loci
and quantitative trait loci (QTLs) is necessary for
correlation between molecular and phenotypic
characters (Burstin and Charcosset 1997). Thus it is
possible that the 28 markers could not capture the
regions of these characters. However, since each of
the NARI and HAC breeding lines are a result of mass
selection, out crossing may had happened so that
genotypes of each group shared common gene pool to
cluster together.
Due to the seed system and selection methods
followed by farmers, high heterogeneity within
cultivars is common in Eritrea for major crops. Backes
et al., (2009) reported high heterogeneity within small
fields of barley and described it as striking results
compared to modern barley fields. Similarly
Asgedom et al., (2011) found non-uniformity within
tomato varieties maintained by farmers in Eritrea and
mentioned it unusual for true to type varieties. In
Eritrea pepper is mainly produced by small holding
farmers in land size as small as 0.015 ha who usually
keep their own seed by selecting the best plants in the
field (Saleh et al., 2013). In the current study three
individual plants were sampled from each seed sample
and each analyzed separately. The three individual
plants clustered in different or the same cluster but
none of them was identical to each other indicating
heteroginity within genotype. Since pepper is a
self-pollinating crop with high outcross, the high
heterogeneity observed within the genotypes is
expected. This was similarly found with Han
á
č
ek et
al. (2009) in pepper, QI-Lun et al., (2008) in maize
who analysed 3 and 12 samples of each cultivar or
landrace respectively.
Both plateau and delta K methods and the neighbour
joining clustering proved the existence of three
clusters. However, Pritchard et al. (2012) indicated
that the aim should be targeting the smallest value of
K that can capture the best structure. In the current
study raising the number of clusters to 4 moved all the
AVRDC and KALRO genotypes into a separate
cluster leaving the three clusters to be almost purely
for Eritrean genotypes. This indicates the richness of
genetic diversity and availability of considerable
amount of unique alleles in Eritrean pepper.
3 Conclusion
There is considerable amount of variation among
Eritrean pepper genotypes and the level of variation
