It is well established that the use of molecular markers together with phenotype
is more reliable than phenotype alone in crop improvement. However, lack of
access of molecular markers contributes to over reliance on phenotypes despite
their limitations especially in the developing world. Among molecular markers,
SNP genotyping is a promising platform for providing plant breeders with simplest,
useful and most cost-effective genotyping services for marker-assisted selection
(Lucas et al., 2012). An SNP is a single
base pair site in the genome that is different from one individual to another
(Acquaah, 2007). These SNPs are suspected to be more abundant
in plants than even those in the human genome (Gupta et
al., 2001; Zhu et al., 2008).
Cowpea is very important food crop in West Africa where most of the worlds
production, trade as well as consumption take place. In recent past, cowpea
has attracted considerable amount of research attention. Genetic linkage map
was developed by Muchero et al. (2009). Varshney
et al. (2007) assessed different types of molecular markers in diversity
studies and conservation of cowpea genetic resources. A number of genes have
been mapped on cowpea indicating the advance in cowpea research beyond traditional
phenotyping. However, the use of SNP in marker trait association in cowpea has
not been either reported or well known.
Efforts are being made to address problems of biotic stress which is of very
importance in cowpea production. There are pests that feed on practically every
part of the cowpea plant resulting in substantial economic loss if left uncontrolled
(Jackai and Daoust, 1986; Makoi
et al., 2010). Lucas et al. (2012)
reported markers linked with Thrips resistance in the crop. Different strategies
are being employed to reduce insect pests damage such as seed flavonoids
(Lattanzio et al., 2005; Makoi
et al., 2010); improved cropping system using integrated management
(Fatokun et al., 2000);
genetic transformation (Fatokun et al., 2000)
among others with the aim of host plant resistance improvement.
One of the important traits desired in cowpea is large seed size in West Afrpca
(Drabo et al., 1984; Langyintuo
et al., 2003; Tchiagam et al., 2011;
Egbadzor et al., 2013). However, much breeding
objectives have not been directly focused on seed size compare with such traits
as biotic and abiotic stress tolerance (Hall et al.,
1997; Orawu et al., 2013). This is not to
say that seed size has never been studied. Conflicting gene actions have been
reported to control seed size inheritance in cowpea through classical studies
(Drabo et al., 1984). This is indication of
complexity of the trait. Association mapping or linkage disequilibrium which
is an alternative to traditional QTL mapping is a powerful method in complex
traits studies (Abdurakhmonov and Abdukarimov, 2008) and
therefore can be used to better understand inheritance of seed size in cowpea.
Association mapping has advantage over QTL mapping in terms of time, allele
number and quality of resolution (Zhu et al., 2008).
SNPs present in the coding sequences, may determine the mutant phenotype and
will show 100% association with the trait and could therefore, be useful, for
MAS and gene isolation (Gupta et al., 2001).
This research, therefore, aimed at mapping seed size in the cowpea genome using
association population of mainly Ghanaian genotypes with SNP markers. Identification
of significant association between SNP markers and seed size could help in marker
assisted selection of the trait.
MATERIALS AND METHODS
A total of 78 cowpea genotypes were used in the association mapping. Sixty-nine of the genotypes were random samples of gene-bank materials from the Plant Genetic Resources Research Institute of the Council for Scientific and Industrial Research (CSIR-PGRRI), Bunso, collected from across Ghana. There were 5 improved varieties from Ghana and 4 from outside Ghana. Four of the five improved varieties were Tona, Nhyira, Asontem and Zaayura, all in cultivation in Ghana. The fifth variety, labeled Market was taken from the market. The four genotypes obtained from outside Ghana were UCR779, CB27, IT82E-18 and IT97K-556-6.
Genotypes were grown in single rows without replication at WACCI farms, University of Ghana in April 2011 under rainfall condition. Dry pods from five plants out of ten per row were bulk harvested and their seeds removed for data collection. Leaf tissues of one week old plants from each genotype were sampled to the KBioscience Laboratory in the United Kingdom for genotyping with SNP markers. Genotype and phenotype data were used for association mapping.
Data collection and analysis: Seeds were dried to approximately 12% moisture content in the sun after harvest and measurement taken on their sizes (mass, length, width and thickness). Seed length, width and thickness were measured with electronic digital calipers and mass of 100 seeds with scale. The facet of the seed representing length, width and thickness is clarified in Fig. 1.
|| Cowpea seed (W) Width, (L) Length and (T) Thickness
Phenotypic and genotypic data were recorded on excel sheet, the former was imported directly into GenStat for analysis, however, the latter was transported into notepad before importation into GenStat. Preliminary single environment analysis was done on the phenotypic data: vital information including F Probability and heritability estimates reported.
The procedure single trait single environment association mapping (VSN
International, 2012) was followed to identify SNP markers linked with seed
mass and thickness.
Preliminary single environment analysis was run twice: first with genotypes random in order to estimate heritability and variance parameters, the second time with Genotypes fixed to get unshrunken means for QTL analysis. Vital information from this analysis is presented in Table 1.
Correlations between the various attributes of seed size namely seed mass, seed length, seed width and seed thickness were estimated (Fig. 2).
The markers used for the mapping were fairly distributed across the cowpea
genome as shown in Fig. 3 and 4 show strength
of association between various markers and seed mass while Fig.
5 shows their positions in the cowpea genome. Figure 6
and 7 are similar to Fig. 4 and 5
but for seed thickness. Some of the markers having significant association with
the traits are shown in Table 2.
The insignificance of variability in seed length and width in the collection
of cowpea studied might mean that these traits might not be important in seed
size as compared with thickness and mass. Seed length and width have more to
do with seed shape than size. The length and width of cowpea seed are related
to shape such as globose, rhomboid and kidney shapes. These shapes are recognized
no matter the size of the seed and among other things, it is the length and
width of the seed that determine the shape. In this case, there could be small
or large kidney shaped cowpea seed. Disproportionate change in length and width
may alter the shape of the seed. For instance, increasing the width of a globose
seed without a change in the length may change the seed shape to crowder. Increasing
the length of globose seed may lead to ovoid if the width remains the same.
Variability in 100 seed mass and seed thickness of the genotypes studied were
significantly different from each other (0.006 and 0.016 F. probability, respectively)
as shown in Table 1. Although, differences between genotypes
with respect to seed width and length were not significant, their F probabilities
were quite small (0.08 and 0.06, respectively). Because the differences in the
seed width and length were not significant, only 100 seed mass and seed thickness
were used for association mapping with the SNP markers. Notwithstanding, heritability
estimate for all four seed size traits were very high: between 82 and 96%.
|| Results From preliminary single environment analysis on seed
|No. of observations (n) is 78 for all traits
|| Correlation between seed mass, length, thickness and width
It is however, important to note that the heritability estimate for seed width
and length (82 and 85%), respectively were lower than that of 100 seed mass
(96%) and seed thickness (94%). High heritability for seed size is known in
cowpea (Hall et al., 1997).
Significant variability was observed in the collection based on the measured traits (Table 1). With regard to 100 seed mass, the largest seed (20.27 g) was more than three times heavier than the smallest. The difference between the widest (7.53 mm) and the narrowest (4.94 mm) seeds, however was quite small. The longest seed on the other hand was about two times the length of the shortest seed. The seed with the highest value for thickness (5.82 mm) was two and a half times more than the least (2.3 mm).
There was positive correlation between all the traits studied. The highest
correlation (0.83) was between 100 seed mass and seed thickness: the traits
with significant variability. The smallest correlation value of 0.51 was between
seed length and width.
|| Markers mapped onto chromosomes
|| Markers association with seed mass at different p levels
|| Positions of markers with significant association with seed
mass [-log10 p≥5.0]
|| Markers association with seed thickness at different
|| Positions of markers with significant association with seed
thickness [-log10 (p)≥5.0]
|| Markers with significant association with traits [-log10
The strong positive correlation between seed mass and the other three traits
suggests that seed mass can be used to represent seed size. The thicker the
seed, the heavier it is expected, hence the high correlation between thickness
In addition to wide variability observed in seed mass between the genotypes, seed mass was easier to measure compared with the three. This also suggests that there could be much more error in measuring seed width, length and thickness than weighing to obtain the 100 seed mass.
Significant associated markers were observed for seed mass on all the eleven
chromosomes at the default level of p (2). There was just one associated marker
on chromosome 9. Chromosomes 2 and 7 had two associated markers each with seed
mass. There were 3 significant markers each for chromosomes 1 and 11. Marker
trait significance level was raised from the default of 2 to 5. The number of
significant SNPs associated with 100 seed mass also reduced drastically. No
marker on chromosomes 2, 3, 6, 8 and 9 were significantly associated with seed
mass. Chromosomes 1, 10 and 11 had one significant associated marker each. Chromosome
4 had 3 significant markers while 5 and 7 had two markers. This result does
not agree with Fatokun et al. (1992) where cowpea
seed weight genes were mapped on chromosomes 2 and 6. However, Fatokun
et al. (1992) used RFLP and not SNP. Kelly et
al. (2003) also mapped seed weight genes of cowpea on chromosomes 1,
3, 4, 6 and 7 showing some agreement with the current studies.
Trend in significance association between markers and trait for seed thickness was similar to that of 100 seed mass. The two traits shared a number of significant markers. This could be indication that seed thickness is responsible for seed mass and consequently seed size in cowpea. At [-log10 (p)≥5.0], there were 18 significant SNPs for the two traits. Seed mass had only 3 significant markers that were not significant with seed thickness. Seed thickness on the other hand had 7 significant markers that were not significant with seed mass. The significant SNPs shared with seed mass and seed thickness could be studied further as they might be the most important seed size markers in cowpea. In this regard, the single markers with significant association on chromosome 1, 10 and 11 are strongly suggested.
A key issue in the study of heredity and variations at the molecular level
is the detection of associations between DNA sequence variation and the heritable
phenotypes (Gupta et al., 2001). SNPs
are often linked to genes (Acquaah, 2007), suggesting
the association of these ones with seed size genes. The association between
SNP markers discovered in this experiment could be used in marker assisted breeding
of large seeded cowpeas which is of economic value in West Africa.
Results of the experiment showed seed mass and seed thickness as the major determinants of seed size in cowpea. Seed mass and thickness were significantly different within the population studied. It was also observed that these two traits were highly correlated (0.83). Seed width and length could be more of shape determinants than size. There were significant marker-trait association for both seed mass and thickness for 8 SNPs used in the experiment at [-log10 p≥5.0]. The markers with significant association with seed size are distributed throughout the cowpea genome from chromosomes 1 to 11 in the exception of 2, 3 and 9. It would be important to estimate the contribution of each of these loci to seed size differences. The distribution of the significant loci among other things confirmed quantitative nature of seed size. Developing on this knowledge can lead to marker assisted breeding of larger seeded cowpea varieties.
We appreciate the benevolence of CSIR-PGRRI for supplying us with the chunk of genotypes used for the studies. We are indebted to the quick response of Mr. I.F. Asare of the seed store section in supplying the germplasm and Mr. D. Ofori a National Service Person at CSIR-PGRRI for his assistance in phenotyping. We thank GCP for providing the financial support and KBioscience for the genotyping service: without them it might not be possible to do this work. We also wish to express our appreciation for support given by the following: Mr. E. Addo of the Biotech Centre, CACS and all staff of WACCI University of Ghana.