Cowpea (Vigna unguiculata): Genetics, genomics and breeding

Communicated by: C. Ojiewo Abstract Cowpea, Vigna unguiculata (L.), is an important grain legume grown in the tropics where it constitutes a valuable source of protein in the diets of millions of people. Some abiotic and biotic stresses adversely affect its productivity. A review of the genetics, genomics and breeding of cowpea is presented in this article. Cowpea breeding programmes have studied intensively qualitative and quantitative genetics of the crop to better enhance its improvement. A number of initiatives including Tropical Legumes projects have contributed to the development of cowpea genomic resources. Recent progress in the development of consensus genetic map containing 37,372 SNPs mapped to 3,280 bins will strengthen cowpea trait discovery pipeline. Several informative markers associated with quantitative trait loci (QTL) related to desirable attributes of cowpea were generated. Cowpea genetic improvement activities aim at the development of drought tolerant, phosphorus use efficient, bacterial blight and virus resistant lines through exploiting available genetic resources as well as deployment of modern breeding tools that will enhance genetic gain when grown by sub-Saharan Africa farmers.

mainly for its grains, which are rich in protein with most improved varieties containing between 20 and 25 per cent protein on dry weight basis. Fresh leaves are also used as pot herb especially in East Africa. Samireddypalle et al. (2017) reported that cowpea haulm (dried leaves, stems and pod walls) could be a good source of income for farmers in the dry savannah areas where the farmers also keep livestock. In an earlier study, Duke (1990) found that cowpea fodder could contain up to 18.6 g protein per 100 g dry weight.
Depending on the region, seed coat colour and texture could be very important to consumers. For example, in northern parts of Nigeria where cowpea is generally produced because of favourable climatic conditions white-coloured grains are preferred by consumers, whereas in the southern parts of the country, preference is for brown-seeded types. Consumers also have preference for varieties with short cooking time, as less fuel is needed when cooking such varieties. This article reviews key achievements in the genetics, genomics and breeding of cowpea which contribute to the enhancement of the productivity of this crop.

| Insect pests
The productivity of cowpea in typical SSA farmers' fields is abysmally low at less than 600 kg/ha when compared with a potential grain yield of over 2,000 kg/ha. A number of abiotic and biotic factors are responsible for these low yields. The adverse effects of these yield limiting constraints can be reduced to their barest minimum through genetic improvement. Cowpea seedlings can be attacked and even killed by aphids (Aphis craccivora) if not controlled with insecticides or planting of resistant variety. Aphids are more devastating when drought occurs shortly after seedling emergence in the field. The single dominant gene that conferred resistance to aphid and assigned the symbol Rac-1 by Bata, Singh, Singh, and Ladeinde (1987) has become ineffective against aphid. This is a typical example of a major resistance gene breaking down in cowpea. Ombakho, Tyagi, and Pathak (1987) and Pathak (1988) reported another dominant gene, derived from induced mutation and confers resistance to aphid, was given the gene symbol Rac-2. The flower bud thrips (Megalurothrips sjostedti) is yet another major insect pest of cowpea with devastating effects mainly on flower buds thereby inhibiting anthesis and pod formation. As no pods develop on plants attacked by this insect, they tend to remain vegetative for a longer period of time. The legume pod borer, Maruca vitrata, the most cosmopolitan of cowpea pests, is capable of causing extensive grain yield reduction (20%-80%) if not controlled (Singh, Jackai, Dos Santos, & Adalla, 1990). A number of pod sucking bugs (Clavigralla tomentosicollis, Riptortus dentipes and Nezara viridula among others) attack pods and suck sap from the seeds while still developing within the pods. Such seeds become seriously reduced in size and malformed thus making them unviable and unattractive to farmers and consumers. When cowpea seeds are placed in storage, weevils (Callosobruchus maculatus), which often accompany seeds from the field, cause extensive damage by the insect's larvae, that feed and develop inside with adults boring holes through which they emerge. Low levels of resistance to a few of these insect pests have been identified among some cowpea germplasm lines. Host plant resistance through incorporating desired genes in improved varieties would be the preferred approach to protect cowpea plants in the field and seeds in storage against these pests.
Effective insecticides are not readily available and could be unaffordable to majority of the farmers.

| Abiotic constraints
Although cowpea is known to be drought tolerant when compared to other crops, the productivity of cowpea could be hampered by erratic rainfall in the beginning and towards the end of the rainy season-a common phenomenon, in the semi-arid tropics where cowpea is mostly grown. With the current effects of climate change, the pattern of rainfall in the subregion, which either comes late or stops earlier than usual, requires that efforts be made to enhance the level of drought tolerance in currently existing improved crop varieties being grown by farmers (Fatokun, Boukar, & Muranaka, 2012). Cowpea yield can also be affected considerably by heat in sensitive varieties. When the night temperature reaches about 35°C, cowpea flowers abort due to poor pollen development, which can result in poor seed and pod set (Hall, 1993). Low soil fertility due to low organic matter and low phosphorus in the savannah soils is also a major constraint for cowpea production.

| GENETICS
Comprehensive reviews of cowpea genetics were published by Fery (1980,1985), Fery and Singh (1997) and Singh (2002). These reviews covered relevant literature on cytologic, qualitative and quantitative genetics. We will cover here some more recent literature. Padi (2003) studied the inheritance of leaf node pigmentation, flower (petal) colour, immature pod colour, seed coat colour, seed eye colour and seed eye colour pattern and reported that presence of pigment was dominant over the absence of pigment and the black seed eye was dominant over brown eye. He further reported partial dominance of the very small eye pattern to the Holstein eye type. Mustapha and Singh (2008) also reported that pod pigmentation is digenic while pigmentation in the pod tip followed two patterns of inheritance (monogenic and digenic), which agreed with Harland's (1920) findings. Cowpea seed coat colour is an important market trait of the crop in West Africa. Using six different biparental crosses, Egbadzor et al. (2014) could not classify segregating materials based on seed coat colour as chi-square goodness-of-fit test could not be conducted on them. They suspected that many genes might be involved in cowpea seed coat colour inheritance. Fery (1985) already mentioned that the complete mechanism of seed coat pigmentation was complex and not yet understood. Lachyan, Desai, and Dalvi (2016) studied also the inheritance of some traits in cowpea. The results showed monogenic inheritance for all four qualitative traits including growth habit, flower colour, seed coat colour and seed coat colour pattern. Joint segregation was observed between seed coat colour and seed coat colour pattern. Lopes, Gomes, and Filho (2003) reported that the number of genes that control 100-seed weight in cowpea is five and the high values for narrow sense heritability indicates that selection for seed size can be made in early generations. Pandey and Dhanasekar (2004) reported the presence of connate foliaceous stipules of primary leaves and their inheritance in cowpea genotype EC394736. They found the rudimentary stipules (RS) to be dominant over foliaceous stipules (FS). The F 2 segregation into 15 (RS): 1 (FS) indicated that duplicate recessive genes control the presence of the FS. Ishiyaku, Singh, and Craufurd (2005) showed that photoperiod in the field of 13.4 hr per day was long enough to delay flowering of photoperiod-sensitive cowpea genotypes and photoperiod sensitivity was found to be partially dominant. Additive and additive 9 dominance interactions were the most important gene actions conditioning days to flower. Narrow sense heritability of 86% was observed while at least seven major genes with an average delay of 6 days each control time to flowering in the cross.
Cowpea aphid-borne mosaic virus (CABMV) is a major virus disease that causes substantial cowpea yield loss. Orawu, Melis, Laing, and Derera (2013) have reported that CABMV resistance is conditioned by more than one recessive gene in eight populations, single recessive genes controlled resistance in other seven populations.
The continuous distribution of progeny for severity data observed in the F 2 populations suggests significance of quantitative inheritance for CABMV resistance. The general combining ability effects (59.8%) were more important than the specific combining ability effects (40.2%) in determining virus resistance in the tested cowpea varieties. In another study using genotypes KVx640 and KVx396-4-5-2D, Barro et al. (2016) found that resistance to CABMV is governed by two dominant genes, each variety contributing a resistant gene.
Three types of host reaction to bacterial pustule were observed by Patel (1981) during the screening of cowpea lines: brown hypersensitive resistant (BHR), non-hypersensitive resistant (R) and susceptible (S). Inheritance study of the BHR, R and S host reactions produced by three races of the bacterial pustule pathogen (Xanthomonas campestris pv. vignae unguiculatae) revealed that BHR reaction was dominant over R and S reactions, and R was recessive to S reaction (Patel, 1982). BHR reaction seemed to be controlled by two genes: one governing BHR reaction to race 1 and the other to races 1 and 2. Both of these genes were ineffective against race 3. The study showed that R reaction seemed to be controlled by one, two or three recessive genes that are effective against all the races.
Using six populations (Parent 1, Parent 2, F 1 , F 2 , BC 1 and BC 2 ) generated from each of four crosses involving four resistant and two susceptible cowpea varieties evaluated for resistance to Cercospora leaf spot (CLS), Booker and Umaharan (2008) found that mode of inheritance of resistance to Pseudocercospora cruenta can be oligogenic or polygenic depending upon the cross. This is the first report of polygenic inheritance of CLS resistance. In another study, the evaluation of CLS disease in F 2 plants and F 2:3 families derived from a cross between "CSR12906" (susceptible) and "IT90K-59-120" (resistant) revealed that the disease scores were continuously distributed, suggesting that the resistance in IT90K-59-120 is a quantitative trait (Duangsong, Kaewwongwal, Somta, Chankaew, & Srinives, 2016).
The genetics of flower thrips resistance was studied in crosses of four cowpea lines. Omo-Ikerodah, Fatokun, and Fawole (2009) found that resistance to thrips is quantitatively inherited with broad sense BOUKAR ET AL.
| 417 heritability ranging from 56% to 73% and maternal effect. Additive, dominance and epistatic gene effects contributed significantly to thrips resistance. These authors reported that resistance to flower thrips is oligogenic with different genes involved in the control of resistance in TVu1509 and Sanzi. ing by sequencing (GBS) was applied to discover SNPs in cowpea, which were used to estimate genetic diversity, population structure and phylogenetic relationships (Xiong et al., 2016).

| GENOMICS
Cowpea has a relatively small genome size estimated at 620 Mbp. Genome sequencing and analysis of the hypomethylated portion of the cowpea genome selectively cloned by methylation filtration (MF) technology were carried out by Timko et al. (2008). Over 250,000 genespace sequence reads (GSRs) with an average length of 610 bp were generated, yielding~160 Mb of sequence information.
About 74% of cowpea expressed sequence tags (ESTs) and 70% of all legume ESTs were represented in the GSR data set. Given that 12% of all GSRs contain an identifiable SSR, the data set is a power- QTLs for leaf shape, maturity time, grain weight (seed size), seed coat colour and patterns have also been identified. Detailed information on QTLs was provided in recent review  where precise parental genotypic information of parents is not available.
KBio converter is a conversion tool that allows the user to input a standard LGC Genomics SNP Viewer input file and output an equivalent file in which each SNP has been converted to a reference strand.

| Historical trends
Cowpea research has been underway in some African countries for  (Singh & Ntare, 1985).
Early breeding activities in sub-Saharan Africa involved germplasm collection, evaluation and maintenance, followed by screening for disease resistance. Efforts were directed to breeding for insect resistance, early maturity, improved plant types and desired seed quality. Pure line and mass selection were initially conducted to identify the lines with high yield potential, meanwhile followed by activities involving hybridization, population advance and selection. Seed quality, high grain yield potential, early maturity, insensitivity to day length, erect growth habit, lodging resistance and best fitting to intercropping were among the initial key desired traits. Some level of work on disease resistance was initiated in Tanzania with technical support from IITA. Mass selection, conventional bulk method and pedigree of bulk-progeny test were among the main breeding methods. Lines developed by IITA were evaluated and seeds of promising lines multiplied and released for general cultivation by national partners (Singh & Ntare, 1985). So far, more than 80 varieties from IITA breeding nursery have been released in over 60 countries.
The priorities for cowpea improvement at IITA were reviewed and modified regularly. In the 1970s, the focus was on diseases which resulted in the identification of lines with high potential: TVu 201(S), TVu 1190, TVu 1977 and TVu 4557 (Singh & Ntare, 1985). In the 1980s, focus shifted to cowpea improvement for insect and multiple disease resistances and white rough seed coat characteristics. In the 1990s to early 2000s, the IITA cowpea breeding programme embarked on the development of (1) extra-early-maturing (60-70 days) photoinsensitive grain type, mainly used in sole crop and short rainy seasons, (2) medium-maturing ( and TVx3236, respectively.  established. The diversity in the core collection was similar to that of the entire collection and correlated traits that may be linked were also preserved in the core collection (Mahalakshmi, Ng, Lawson, & Ortiz, 2007). A reference set, also called mini core, composed of 370 accessions representing the entirety of the genetic diversity of the core was constituted. The minicore is a critical resource for scientists to study new adaptive traits, conduct comparative genomics studies, and discover new favourable alleles and new lines for prebreeding activities.

| Genetic resources
In addition to the cowpea minicore, the first eight-parent cowpea multiparent advanced generation intercross (MAGIC) population was developed recently as an important genomic community resource for trait discovery and breeding. The eight founder parents were selected based on abiotic and biotic stress tolerance or resistance and agromorphological trait variability (Huynh et al., 2017). The eight parents were intercrossed using structured matings as described by Cavanagh, Morell, Mackay, and Powell (2008) with some modifications. Pairwise crosses (AxB; CxD; ExF; GxH) were performed followed by 600 double crosses producing F 1 seeds (300 ABxCD; 300 EFxGH). A total of 300 four-way random pair crosses were made (ABCD x EFGH). Three hundred and sixty-five (365) F 1 s were generated from the crosses and advanced by single seed descent (SSD) till F8 generation (Huynh et al., 2017). Prior to the advancement of the population through SSD, each of 365 eight-way individuals was genotyped using SNP genotyping with the cowpea Illumina 60K iSelect BeadArray to confirm each individual was derived from an eight-way cross.

| COWPEA BREEDING UNDER TLII
Cowpea is grown mostly in the dry northern guinea savannah, Soudan Savannah and Sahel agro-ecologies characterized by low annual rainfall. In recent times, the amount of rainfall received in these areas is declining and the distribution of the rains is irregular especially during early or late stages of the cropping seasons. These were released in Tanzania. During FPVS, we observed that the best lines in terms of grain yield in some locations were not necessarily the most preferred by farmers. Attributes such as striga resistance was a very important selection criterion coupled with grain and fodder yields by farmers in West Africa hence they selected IT97K-499-35 in Mali, Niger and Mali whereas farmers in Mozambique preferred IT00K-1263 because of its high grain yield ability (Fatokun, Boukar, Kamara, et al., 2012). More than 24 IITA breeding lines were released during the last 10 years in 13 different countries in sub-Saharan Africa (Table 2).

| Sustenance of the breeding pipeline
In order to identify sources of new genes to use in the development of better performing breeding lines, some germplasm and improved breeding lines available at IITA were screened for different desirable traits under the Tropical Legumes Project.

| Screening for drought tolerance
About 1,200 germplasm lines from the Genetic Resources Center at IITA were evaluated in the field for their responses to drought conditions. The evaluation was carried out under irrigation during the dry season. Drought stressed plants were exposed to terminal drought following withdrawal of irrigation at 5 weeks after sowing.
It was observed that drought depressed grain yield in all the germplasm lines tested. However, there were significant differences in the extent of grain yield reduction due to drought among the lines. IT99K-573-2-1, IT99K-573-1-1 Ghana was suspended. Some of these stay-green type had delayed flowering under drought (Fatokun, Boukar, & Muranaka, 2012). A total of 190 lines were found to show enhanced levels of drought tolerance as proportions of grain yield reduction due to drought were relatively lower among them. These were further evaluated, and the best 10 were selected for use in making crosses aimed at developing populations segregating for drought and from where selections have been carried out for breeding lines with superior drought tolerance.

| Screening for phosphorus use efficiency
The marginal soils in which cowpea is mostly grown in the Sahel additive as well as non-additive genetic variability (Inegbenose, 2016) in the development of BCMV resistant cowpea breeding lines.

| CONCLUSION
Although cowpea has for some time now been regarded as an orphan crop in view of the relatively low level of research attention given to the crop, modest progress has been made in the assemblage and conservation of its germplasm, generation of genomic tools for more effective breeding and development of improved varieties some of which are already available in SSA farmers' fields. The challenge of striga to cowpea production especially in the dry savannah agro-ecologies is being effectively contained with the development of varieties that show immunity to this parasitic weed. There is however room for the development of better performing varieties that will be characterized by high stable grain yield, good amount of protein, large seed size, white or brown seed coat colour, resistance to some insect pests which presently creating huge losses to the crop.
The over 15,000 and 2,000 accessions of cultivated and wild compatible wild relatives in the gene bank at IITA, respectively, need to be systematically screened for adaptive genes that control the abiotic and biotic stresses still limiting its productivity. In addition, cowpea breeding programmes at IITA and NARS are currently engaged in the implementation of modern breeding practices that will improve their efficiency. With the development of various platforms such as phenotyping, genotyping and data management coupled with necessary resources now being committed to cowpea research, more impacts in term of variety development and integrated crop management practices will be achieved.