Multiple shoot induction in zygotic embryos: a strategy for acceleration of banana breeding

The presence of residual female fertility in most of the parthenocarpic banana accessions encourages the banana breeder to develop new hybrids through conventional breeding. Desirable trait can be fixed in the first generation of hybrid progenies, but the evaluation of these hybrids in field is the time-consuming process owing to non-availability of uniform suckers/planting material. This can be overcome by developing multiple shoots from single embryo in a short period of time through embryo culture. A protocol for in vitro multiplication of plantlets from zygotic embryos was standardized in seeded accessions. Multiple shoots from zygotic embryos were achieved up to 55.2% and 64.1% in seeded accessions of Musa acuminata and M. velutina respectively in medium supplemented with 17.76 µM of BAP. The Single shoot derived (only germination) from zygotic embryos was decapitated and the apical meristem were disturbed for further multiple shoot formation in media supplemented with 17.76 µM of BAP. Present studies revealed that in total 75% and 91% of the M. acuminata and M.velutina embryos were able to produce multiple shoot from single embryo by manipulating the media composition and decortications technique. The above protocol was applied for zygotic embryos obtained from controlled pollination (18 cross combinations) and open pollination (nine accessions) of various genomic groups (ABB, AAB, AA). The multiple shoots derived from zygotic embryos and plantlet germinated from zygotic embryos was examined for genetic fidelity analysis by SSR markers. The protocol for multiple shoot formation from single zygotic embryo under in vitro culture developed in this study will accelerate the banana breeding program.


Introduction
Polyploidy and sterility are the major constraints in the improvement of banana and plantain through conventional breeding approach (Laliberté 2016). Although, cultivated banana is a vegetatively propagated crop, the seeded nature of its ancestors and frequent occurrence of mutations in the natural habitats led to the occurrence of broad banana diversity which is being exploited by hybridization program. Some of the commercial cultivars exhibit low residual fertility to some extent possibly due to seeded nature of the predecessors and transfer of this trait through few gametes having complete set of ancestral genome through balanced chromosomal segregation (Heslop-Harrison and Trude Schwarzacher 2007). Developing parthenocarpic varieties with desirable traits has been possible in banana due to occurrence of 2n gametes followed by hybridization (Ortiz 1997). Thus low rate of residual fertility and the occurrence of unreduced gametes offer the scope for the banana breeder to improve the commercial varieties (Ortiz Communicated Vuylsteke 1995). This has resulted in releasing many banana hybrids by FHIA, Honduras (Rowe and Rosales 1993) International Institute of Tropical Agriculture, Nigeria and Kenya (Crouch et al. 1999); EMBRAPA, Brazil (Ferreira et al. 2004) and ICAR-NRCB, Tamil Nadu Agricultural University, and KAU-Banana Research Station, India (Uma et al. 2015). Although residual female fertility is a desirable trait, the poor rate of seed set (Shepherd et al. 1987) and germination (Andrus et al. 1971) limit the chance of developing of large number of commercial hybrids. The complicated genetic system of banana with heterozygocity and polyploidy, causes the concurrence of abnormal meiotic recombination or coming togetherness of different sets of chromosomes which resulted in poor seed set and poor germination. It was also reported that an abnormal embryo-endosperm relationship and impermeability of seed coat, along with chemical barriers like growth inhibitor-induced dormancy etc. cause poor germination percentage of hybrid banana seeds (Simmonds 1962;Uma et al. 2011). The problem of seed set can be overcome by choosing the compatible male and female parents, optimal stage and time for pollination and by spraying of antoauxins on the flower bud (Backiyarani et al. 2016). The problem of seed germination can be overcome through embryo culture and embryo rescue (Vuyleseke et al. 1990), but the embryo maturity, culture medium (Johri and Rao 1984), positioning of embryo in the culture medium, allele combination of hybrid embryos are significantly affect the germination rate (Asif et al 2001). It has been reported that seed germination can be enhanced up to 10% and 30% by supplementing the media with plant growth regulators (Diro and Van Staden 2003;Uma et al. 2011) and combination of seed treatment with GA3 and supplementing with growth regulators (Arun et al. (2013) respectively, but it should be further enhanced for attaining better success in banana breeding (Batte et al. 2019). Apart from the seed germination, hardening of these embryo derived plantlets is another big task due to poor acclimation and adaptation of plantlets (Vasane and Kothari, 2006). Allam et al. (2000) reported that the successful acclimatization of in vitro produced banana plantlets largely depends not only on the post transfer growth conditions but also on the pre-transfer culture conditions. Therefore, interventions are required to improve the survival rate of hybrid progenies and their subsequent field establishment. The other important reasons for the limited success in banana improvement program are its longer duration and non availability of large number of uniform suckers for further evaluation(s) of each hybrid event for desirable trait(s). Breeding efforts should also be focused on minimizing the evaluation period. Hence the present study was undertaken to enhance the regeneration efficiency and to develop multiple plantlets from zygotic embryos through embryo culture.

Species/genomic groups used for the development of multiple shoots from single embryo
To standardize the media composition for enhancing the germination/regeneration efficiency and multiple shoot formation, the open pollinated seeds of two seeded accessions of Musa species namely M. acuminata and M. velutina were taken. The best media thus identified for multiple shoot development was tried for seeds derived through cross/ open pollinated commercial cultivars. Hybridization was taken up in the four accessions of ABB genome (Chinia, Karpuravalli, Kothia and Saba) and two accessions of AAB genome (Nendran and Poovan), one Nendran × Pisang Lilin (NPL 33) progeny, one Nendran open pollinated progeny (NOP46) and two accessions of AA genome (Matti, Cultivar Rose) using four male diploid accessions of A genome. Out of four male accessions two were parthenocarpic type (Pisang Lilin, Matti) and two were seeded type (Pisang Jajee, Calcutta 4) (Electronic Supplementary Information: Supplementary Table S1). The open pollinated bunches of triploid accession of ABB genome (Benkela, Boothibale, Chinia, Cuba, EnnaBenian, Saba, Vennuthumannan) and wild diploid accessions of AA genome (Calcutta 4 and Microcarpa) were also taken.

Hybridization among different genomic group
The whole inflorescence of the female parents was bagged before the opening of the first hand of female phase. Simultaneously the male phase of each male parent was bagged immediately after the completion of neutral phase. Each day the pollen grains were collected from the male parents during anthesis stage at 7.00 am. And the collected pollens were dusted over the female flowers which are about to open on that day (i.e., having receptive stigma) of the female parents during 7-10.00 am. Immediately after dusting, the female flower hands were carefully closed using its own bract and then the whole inflorescence was bagged. The next hand of the female flower of the same inflorescence was used for dusting on the following day. Hybridization was carried out in 18 different combinations and each pollinated inflorescence was labeled with tags containing relevant crossing information.

Seed extraction
The artificially and open pollinated bunches were harvested at full matured stage. The days taken for full matured stage (from flowering to harvest) of each female parent is given in supplementary table S2. The bunches were kept in the ripening chamber. The seeds were extracted from the ripened bunches and washed in running tap water. The extracted seeds were soaked in water to separate sunken and floating seeds. Since floating seeds are normally devoid of either endosperm or embryo, sunken seeds alone were subjected to surface sterilization by treating them with 4% sodium hypochlorite for 15 min, followed by dipping in 0.1% mercuric chloride for 15 min. Seeds were rinsed with sterile distilled water upon each treatment for about 2-3 min. In order to break the dormancy, the sterilized seeds were soaked overnight in water supplemented with 10 ppm GA 3 as described by Arun et al (2013). Subsequently debris was removed from the seeds by following the same protocol that was utilized for sterilization. Then seeds were transferred to sterile plate and subjected to embryo dissection.

Embryo initiation and multi shoot formation
To minimize the experimental error, a proper sample size of 100 seeds were taken from the cross combinations that recorded profuse seed setting while among the crosses where seed set was low, all the seeds were utilized for studies. The nature of seeds and embryos of controlled and open pollinated bunches of different accessions are given in the (Electronic Supplementary Information: Supplementary Table S2). A longitudinal fissure was made in GA3 soaked sunken seeds and the whitish, mushroom-shaped embryo, which consists of a haustorium and a meristematic stalk ( Fig. 1b) was excised. The haustorium region of the embryo was placed on the surface of medium. The excised embryos of seeded accessions M. acuminata and M. velutina were initiated in Murashige and Skoog (1962) medium supplemented with various concentration of BAP (4.44 µM, 8.88 µM, 13.32 µM 17.76 µM and 22.2 µM BAP). Initiated embryos were kept in dark condition approximately for a period of 7-10 days until germination took place. The tubes having germinating embryos were shifted to light and dark (16/8-h) condition for regeneration. The regenerated embryos were further subjected for sub culturing in the same medium supplemented with the same BAP concentration. The excised embryos of the seeds collected from artificially and open pollinated bunches of various accessions were placed in the best media composition (medium supplemented with the 17.76 µM BAP concentration) and cultured as per the protocol mentioned above.

Decortications of embryo derived single plantlets
Embryo culture derived single plants were further subjected to decortications to obtain the optimum width of basal shoots. To determine the optimum shoot size, three different sizes of the single plants were selected and the excised shoot tips with the size of 1, 2 and 3 cm 3 were decapitated and the apical meristem were disturbed and kept in the same media composition. After three subcultures the proliferated shoots were observed for multiple shoot formation.

Hardening of embryo derived plants
The shoots obtained through direct regeneration from single embryo were sub-cultured in rooting media containing auxin i.e. IBA-4.90 µM and NAA-10.74 µM. After the root formation, the single plantlets were primary hardened in the protray containing the mixture of coir pith with vermicompost. After 45 days, well grown plants were transferred for secondary hardening in a polybag containing a mixture of red soil, vermicompost and coir pith in equal proportion.

Statistical analysis
Experiment on standardization of media for multiple shoot formation was carried out in completely randomized design with three replications per treatment. The germination percentage data was transformed using arc sine transformation and the transformed data were analyzed using SPSS v. 11.09 (IBM Corporation, Armonk, NY, USA). The mean of three replicates of each parameter was expressed as mean ± SD.

Isolation of genomic DNA
Two gram of leaf tissues were collected from cigar leaf (recently emerged leaf which are tightly coiled, whitish, and particularly fragile and still rolled as a cylinder) from the secondary hardened plants (minimum of five plants) of each hybrid and for parents cigar leaf tissues were collected from field grown plants. The leaf tissues were freeze dried using liquid nitrogen and genomic DNA was isolated by cetyl-trimethyl ammonium bromide (CTAB) method (Gawel and Jarret, 1991) with a minor modifications. The isolated DNA was treated with 7 µl RNase (100 mg/ml), and quantified by spectrophotometer at 260 nm and 280 nm and the quality of DNA was checked with the help of agarose gel electrophoresis (0.8%).

PCR amplification
The genetic fidelity of multiple shoots obtained from single embryo was tested through PCR using thirty five in silico polymorphic SSR primers (Electronic Supplementary Information: Supplementary Table S3) retrieved from the Musa-transSSRDB-Musa transcriptome derived SSRs database (http:// bioin fnrcb. byeth ost7. com/ nrcbb io/) (Backiyarani et al. 2019). Polymerase chain reaction was carried out for a total volume of 10 µl containing 10X PCR buffer, 200 pmol of each forward and reverse primer, 2.5 mM of dNTPS, 0.5 unit of Taq polymerase, 100 ng of DNA plus nuclease free water to make up to the final volume. Amplification was performed on a programmable thermocycler (Applied Table 1 Efficiency of BAP concentration in seed germination and shoot formation of two wild banana species under embryo culture Data were analyzed using SPSS v. 11.09 (IBM Corporation, Armonk, NY, USA), and are presented as the mean of three replicates. Data were expressed as mean ± SD. Means of each column has the same letter/s are not significantly different at 0.05 level of probability according to Duncan's multiple range test  Biosystem) with a condition of initial denaturation at 94 °C for 5 min, followed by 35 cycles of denaturation at 94 °C for 1 min, annealing at the corresponding annealing temperature for the respective primers employed for 30 s and extension at 72 °C for 1 min, with a final step of final extension at 72 °C for 10 min. PCR products were loaded on 3% agarose gel and the gel was documented using gel doc system (Gelstan, medicare).

Results
To standardize the multiple shoot production in embryo culture, MS media with different concentration of BAP were tried in two seeded accessions of Musa To produce multiple shoots from the single plant developed from embryo, three different sizes ('A' type: 3 cm 3 girth; 'B' type: 2 cm 3 and 'C' type: 1 cm 3 ) of embryo cultured single plants were used as explants and subjected to initiation of shoot tip culture. Proliferation has been noticed within 30-35 days of sub culturing. The proliferated explants were again sub cultured and an average of seven plantlets were obtained (maximum of ten) among type A followed by 'B' type with an average of 5. Continuous sub culture resulted in enhancing the shoot formation and in the third sub culture, highest multiple shoots were obtained in 'A' type (10) followed by 'B' type (7) but no proliferation was obtained among 'C' type explants which only resulted in single plant development. This study revealed that in vitro plantlets with the stem girth of 3 cm 3 and 2 cm 3 are suitable explants sizes for the production of multiple shoots. It was observed that multiple shoots were induced in 52% and 75% of embryo derived single plantlets of M. acuminata and M. velutina from the decortications of 'A' type explants.
To understand the response to multiple shoot formation in the breeding material, sunken seeds obtained from 18 different cross combinations of different ploidy and genomic group such as AA × AA (2), AAB × AA (9) and ABB × AA (7) and open pollinated seeds of nine different genotypes such as ABB (7) and AA (2) were subjected to embryo culture in the media composition of MS with 17.76 µM BAP.
Genetic fidelity test of the multiple shoots derived from single embryo of each cross was carried out using in silico polymorphic primers. The parental polymorphic primers were detected by testing against the parents of the each cross combination using 35 primers. Of which 20 showed polymorphism among the parents used for controlled pollination for the respective cross combinations. All the primers showed monomorphic bands for the multiple shoots derived from each embryo irrespective of origin, whether control or open pollinated seeds.

Discussion
The success of any conventional breeding program depends on the production of vast amount of recombinant events among the compatible parents through sexual hybridization. Banana is a recalcitrant crop for sexual hybridization owing to its polyploidy, female and/or male sterility. In spite of these hurdles, owing to its residual female fertility natural seed set has been observed in some of the parthenocarpic genotypes paving way to improve the bananas through hybridization techniques (Uma et al. 2011;Ortiz et al. 1995). Still, the success rate of banana breeding is slow owing to low germination percentage of hybrid seeds (Batte et al. 2019). Direct sowing of hybrid seeds in potting mixture resulted in poor germination (1-1.4%) (Swennen et al. 1992;Talengera et al. 1996) with germination extending over a period of 12 months (Uma 2015). The slower and low germination percentage is due to water repellent substances present in the cell walls of the mesotesta (Guan 1991) and the presence of polyphenolic substances which hinder the diffusion of oxygen through the seed coat (Burgos-Hernández et al. 2014). The germination percentage can be improved either by extracting the seeds at the right maturity (Uma et al. 2011) and/or by using stratification and scarification methods (Wattanachaiyingcharoen 1990). But through these approaches, germination percentage could be enhanced only up to 4-16% in wild species (Burgos-Hernández et al. 2014). While Ahmed et al. (2006) and Dayarani et al. (2014) reported that it can be enhanced up to 40-60% through embryo culture of wild seeds under in vitro conditions. Asif and Othman (2001) reported that apart from enhancing the seed germination, the germination period could also be reduced from 50 days to one week through in vitro embryo culture over direct seed germination.
Survival of plantlets derived from hybrid embryo is also an important factor for the success of banana breeding program. The risk of mortality of the embryo derived plants can be reduced by developing multiple plantlets from a single embryo. This indirectly helps to maintain all the regenerated hybrid events and facilitate to evaluate large number of heterogeneous population in one go which is essential for the selection of best recombinant variants. Uma et al. (2012) developed multiple shoot formation protocol from immature embryos of banana through embryo rescue via callus induction. Similarly, many researchers evidenced that plantlets could be produced through indirect regeneration from the mature zygotic embryos by media augumentation with high 2, 4, D and IAA in various Musa species such as M. ornata (Dayarani et al. 2014) and M. acuminata (Uma et al. 2011) respectively. Sivanesan (2007) and Marco A. Ramírez and Lourdes (2015) reported that plantlets regenerated through indirect organogenesis, showed high genetic and epigenetic variations (Grafi and Barak 2014). Varshney et al. (2001) also reported that the indirect regeneration involving a callus phase is more vulnerable to DNA damage during micro-propagation which led to higher frequency of somaclonal variation (Peschke and Phillips 1992). The occurrence of somaclonal variation could be reduced by avoiding the long-term culture and using axillary shoot induction systems. Zapata et al. (1999) evidenced that direct regeneration is a faster and a time saving approach for obtaining whole plants without the callus interphase. Among the purine type of cytokinin, BAP is preferred to for its stable nature (Klem et al. 2004) and effect in stimulation of multiple shoots (Kadota and Niimi 2003).

Multiple shoot formation from single embryo
Variation was observed among the two seeded accessions for days taken for germination, M. acuminata showed early response (3-7 days) than compared with that of M. velutina (3-10 days). Though no specific trend was observed for the days taken for germination with different concentrations of BAP, enhanced germination was noticed with the increased BAP concentrations, in both the species. Similar result was also reported by in Aconitum hererophyllum (Pandey et al. 2000) and Amaranthus sp. (Tiryaki et al. 2009). The supplementation of BAP in the media composition enhances the germination percentage as it is involved metabolically in the stimulation of germination (Dissanayaka et al. 2015) and reduces the chances of mortality as it provides a higher competitive ability (Zhang & Maun 1990). Franklin et al. (2000) reported that, incubation of embryo in dark followed by light conditions is a prerequisite for the induction of morphogenic potential of cells and plantlets formation (regeneration) respectively. Thus, the germinated culture was shifted to light and dark (16/8-h) condition and shoot elongation was noticed from third day on wards. Proliferation of shoots was observed earlier in the media composition of MS with BAP 17.76 µM (15-20 days) compared to other concentrations. In the present investigation, the normal shoot elongation was observed in all the BAP concentrations except in 22.2 µM where the elongation rate was reduced. Chang et al. (1992) also reported that higher BAP concentration stimulated higher shoot formation but inhibiting its further development.
Interestingly number of multiple shoot increased with increased concentration of BAP up to MS + 17.76 µM and beyond which it is reduced. In this, multiple shoots emerged directly from embryo without any intermediate callus tissue. Similar trend of multiple shoot formation as in the present study were also observed with embryos of barley (Ganeshan et al. 2003), rice (Zhang et al. 1996) groundnut (Palanivelet al.2002 and papaya (Bhattacharya et al. 2003). Many reports reported that meristematic cells of germinated embryos have the ability to produce multiple shoots under the higher concentration of cytokinins (Sobhakumari and Lalithakumari 2003;Rao et al. 2006), but rosette of distorted leaves with up-normal growth of shoots was observed in higher concentration of BAP (22.2 µM BAP) in M. acuminata and M. velutina. Maximum of nine shoots with an average of seven shoots from single embryo was obtained in M.accuminta and seven shoots with an average of six shoot were obtained in M. veluntina in the media supplemented with 17.76 µM BAP.
Two types of multiple shoot formation were observed in this study. In the first type, primary apical meristem shoot was developed from the haustorium region of embryo and from these started the initiation of the first leaf. Then 3-4 secondary shoots arose simultaneously around the primary meristem from the same haustorium region (Fig. 1e). In the second type, embryos produce shoots simultaneously and produce very short shoots with big leaves which resulted in a rosette appearance. And normal shoot production was observed by continuous sub culturing in the same media composition. This variation in pattern of multiple shoot formation might be due to the genotype effect i.e. variation in the level of endogenous hormone (Venkatachalam et al. 1999).
Interestingly, except in MS media supplemented with 22.2 µM BAP, root initiation was also observed in all other concentrations. The elongated shoots were excised and subcultured onto MS medium supplemented with the auxin IBA-4.90 µM and NAA-10.74 µM. Overall the present experiment suggested that 17.76 µM BAP is the optimum concentration not only for enhancing the germination and regeneration efficiency but for multiple shoot production too in the short span of time. This phenomenon has been applied for the production of multiple shoots/embryo in NRCB breeding program.

Induction of multiple shoots from embryo derived single plantlets
It was observed that through media manipulation technique, 44.8% and 35.9% of the embryos produced single plantlets in M. acuminata and M. velutina respectively. Thus, to develop multiple shoots from a single shoot, bud manipulation technique, decapitation of the apical meristem, was performed to remove apical dominance. Suman (2017) reported that the success of shoot tip culture mainly depends on the final size of the explants. But the size of the single plantlets which is ready for primary hardening after removing the sheath is very small (1 cm 3 ) which is not suitable for decortications. Hence to increase the size of the explants, embryo derived single plantlets were subjected to three sub cultures in the same media composition of MS + 17.76 µM BAP. Three different sizes of explants were subjected to decortications followed by sub cultures for the production of shoots and the results revealed that stem girth of diameter 3 cm 3 and 2 cm 3 width in vitro derived plantlets are suitable size explants for the production of multiple shoots. It was observed that multiple shoots were induced in 52% and 75% of embryo derived single plantlets of M. acuminata and M. velutina respectively from decortications of A type explants (Data not shown). Both these studies revealed that in total 75% and 91% of the M. acuminata and M. velutina embryos were able to produce multiple shoot from single embryo by manipulating the media composition and decortications technique (Fig. 2).
Normally for evaluation of hybrid progenies for a specific trait, it needs a minimum of five uniform suckers which could be obtained only after 18-24 months of field planting. Uniform size plantlets of hybrid progenies will be available, if the multiple shoots are regenerated through direct regeneration from embryo culture. Within 6-7 months the hybrid progenies will be ready for field evaluation under replication trial whereas it will take 10-11 months if decortications also involved. These results imply that this technique will reduce the duration of breeding programme by 9-12 months.

Understanding the response of various genomic groups for seed germination and multiple shoot formation
The embryos of open and controlled pollinated seeds were initiated and bulging was observed in 56% of embryos, which confirmed their viable status while 44% failed to show any growth induction in terms of size, colour (green) and callus development and turn into black color. Similar result has also been recorded in cucumber (Ali et al. 1991). The non-germination of embryos i.e. lethality of hybrid seed has been attributed due to interploidy crosses (Scott et al. 1998;Köhler et al. 2003). The low germination percentage in the hybrids seeds might be due to different type of genetic imbalance owing to chromosome mismatch of male and female gametes (Simmonds (1962). This may be occurred due to variation in the chromosome structural heterozygosity, ploidy among the intra specific and inter specific genomic groups (González de León and Fauré 1993). In general, irrespective of the genomic group, higher embryo germination and regeneration percentage was observed in the control pollinated seeds than the open pollinated ones. This could mean that the chances of union of compatible gametes are more in case of controlled pollination as the male parents are highly polleniferous and diploid in nature. The occurrence of variation in the regeneration ability of embryos derived from the same and various genomic groups used for crossing with diploids revealed that reproductive and regeneration potential are not only the genome specific but also genotypic/cultivar dependent. This could be due to specific requirement of plant growth regulators which varies from genotype to genotype (Mamidala and Nanna 2011) and regeneration efficiency could be enhanced by manipulating media using growth regulators (Mensuali-Sodi et al. 1995).
In general, the present investigation revealed that the regeneration ability of germinated embryos having cultivar rose as the male parent was comparatively less with interspecific triploid female parent than the intraspecific AA diploid one. In spite of low rate of embryo germination, 100% regeneration was observed in most of the embryos obtained under controlled pollination and all the open pollinated accessions except Ennabernian. The reason for having low regeneration percentage in some of the cross combinations might be that these embryos could have developed by the union of non-functioning of gametes with more unbalanced chromosome numbers which could have occurred as the result of chromosome assortment encountered in the meiotic process of the triploid (Punyasingh 1947).
In general, there is no specific trend noticed in the multiple shoot formation and genomic composition. Irrespective of the male parents used, all the Nendran cross combinations responded for multiple shoot formation whereas no trend was observed for all other cross combinations. In general, this study revealed that multiple shoot induction through both the techniques, influence of female parent is higher than the male parent in all the cross combinations except in Karpuravalli cross combinations. The multiple shoot formation from all the regenerated embryos of the Nendran based cross combinations revealed that they are highly amenable for multiple shoot induction through media manipulation and shoot tip culture.

Genetic fidelity testing
The SSR marker study was carried out to determine the hybridity of the progenies and genetic integrity of the multiple shoots derived from single embryo. In case of controlled pollinated progenies visualization of both the parental bands in the respective hybrid confirmed their hybrid nature and monomorphic banding pattern among the multiple shoot derived plants of each embryo confirmed their genetic fidelity (Fig. 3). Whereas in open pollinated derived plantlets nearly 80% of the primers showed monomorphism with that of female parent. But multiple shoots developed from each embryo did not show any polymorphic bands which proved that multiple shoot formation through direct organogenesis/ decortications from embryo is considered to be more effective, easy and time saving approach compared to indirect organogenesis and/or normal conventional germination of hybrid seeds in the banana genetic improvement. This results is in concordance with the findings of Venkataramana et al. (2015), Rotchanapreeda et al. (2014); RaviShankar et al. (2013) and Manchanda Pooja and Gosal (2011) who used the SSR markers to confirm the genetic fidelity of tissue culture derived plants of various Musa species and cultivars having various genomic composition developed through direct regeneration protocols.

Conclusion
This study proved that media augmented with 17.76 µM BAP concentration enhanced the regeneration efficiency of the pollinated embryos and produced multiple shoots from the regenerated embryos with a range of 36.36-100%.
Decortications of in vitro derived plantlets produced multiple shoots with a range of 50-100% in the remaining embryo derived single shoots. Combination of these two techniques could lead to multiple shoot induction up to 66.7-100% in various genomic accessions of both controlled and open pollinated seeds. Production of multiple shoots from single embryo will reduce the risk of mortality of each hybrid event and reduces the duration of the breeding program by 9-12 months. This multiple shoot formation from single embryo technique can be extended in banana improvement program either through chromosome and/or gene manipulation techniques by doubling the chromosome and developing parthenocarpic trait in seeded accessions through antimitotic agents and genetic transformation respectively.
Author contributions SBR and SU conceived the study, designed and managed the experiments. SS, SE and RK performed in vitro trials and collected data at laboratory. PD and VS performed hybridization and collected data at field. SBR and SK completed statistical analyses of phenotypic data. SBR wrote the manuscript and SU participated in correcting the manuscript. All authors contributed to writing the manuscript. Fig. 3 Confirmation of hybrids and genetic fidelity of multiple shoots derived from shoot tipculture A Cultivar Rose × Calcutta 4, B Nendran × Cultivar Rose, C Nendran × Pisang Jajee, (Lane 1-Female parent, Lane 2-Male parent, 3-7 Multiple shoots from single embryo) D Microcarpa open pollinated (Lane 1 collected from field, Lane 2 to 7-Multiple shoots from single embryo)