Effect of F₂ Inter se Mating on Quantitative Traits Mean, Range, Variance and Heritability in Dolichos Bean (Lablab purpureus L. Sweet var. Lignosus)

Dolichos bean var. lignosus is one of the important grain legume crops extensively grown in southern India. It is predominantly a self-pollinated crop (Kukade and Tidke, 2014) with 2n=22 chromosomes (She and Jiang, 2015). It is predominantly grown as a rainfed crop for fresh beans for use as a vegetable. While fresh pods are the harvestable and marketable products, fresh beans are consumable products in dolichos bean (Ramesh and Byregowda, 2016). Pedigree breeding is the most commonly used method for developing improved cultivars in dolichos bean (Ramesh and Byregowda, 2016), as is practiced in other legumes and inbreeding species. However, pedigree method of breeding enable exploiting only fixable genetic effects as a result of rapid increase in homozygosis (Clegg et al., 1972), which further limits opportunities for recombination retaining and/or favoring accumulation of haplotypes harboring genes controlling desirable and undesirable traits (Hanson, 1959; Pederson, 1974; Bos, 1977; Stam, 1977). One or a few cycles of inter-mating in F₂ generation results in break-up of linkage blocks and help create populations with high frequency of desired recombinants which otherwise cannot be realized in later generations (Hanson, 1959; Bos, 1977; Stam, 1977; Yunus and Paroda, 1983). Further, complete association of ‘+’ genes (which enhance phenotypic expression) is extremely rare and dispersion (which causes reduction in phenotypic expression) is normal in genotypes particularly those that are from the source material for breeding programmes (Kearsey and Pooni, 1996). The probability of genes being in dispersion phase could be minimized by random mating F₂ for few generations before deriving best recombinant inbred lines (RILs) (Stam, 1977). Under these premises, the present investigation was carried out to assess impact of F₂ inter se [Bi-parental (BIP)] mating on quantitative traits means and variances, additive genetic variance and heritability.

The basic material for present investigation consisted of F₂ population derived from HA 4 × GL 37. The HA 4 is a high yielding photoperiod in-sensitive determinate type released variety, while GL 37 is a photoperiod sensitive in-determinate type local collection being maintained at Zonal Agricultural Research Station (ZARS), University of Agricultural Sciences (UAS), Bengaluru.
 
Development of experimental material
 
To raise F₂ population,150 seeds harvested from F₁ (HA 4 × GL 37) plants were sown with inter-row spacing of 0.6 m and of 0.3 m between plants within a row at the experimental plots of Department of Genetics and Plant Breeding, UAS, Bengaluru during 2012-13 rabi-summer season. Due to non-germination of a few seeds and mortality of a few seedlings, the F₂ population consisted of 132 plants at flowering. Randomly selected 20 pairs of F₂ plants with flowering synchrony were inter se mated. Seeds from 20 paired crosses (here after designated as BIPF₃ progenies) and selfed seed (F₃) from their 40 parents were collected, cleaned, dried and preserved. Twenty BIP F₃ and 40 F₃ progenies constituted the experimental material for the study.
 
Evaluation of experimental material
 
Seeds of 20 BIP F₃ and 40 F₃ progenies along with the two parents and three checks (HA 3, HA 4 and Kadle Avare) were sown in Augmented design (Federer, 1961) in five compact blocks at the experimental plots of UAS, Bangalore during 2015 rainy season. Each block consisted of 12 BIP F₃ and/or F₃ progenies, three checks and two border entries. Each entry in each block was sown in a single row of four meters length with a row spacing of 0.45 m and 0.3 m between plants within a row. Each of BIP F₃ and F₃ progenies consisted of 18 to 20 plants. Recommended agronomic and plant protection practices were followed to raise a healthy crop.
 
Sampling of plants and data collection
 
The data on eight quantitative traits (QTs) namely, days to flowering, primary branches plant^-1, racemes plant^-1, nodes raceme^-1, fresh pods node^-1, fresh pods plant^-1, fresh pod yield plant^-1 and fresh bean yield plant^-1 were recorded on 15 plants (avoiding border plants) of BIP F₃ and F₃ progenies using the protocol described by Byregowda et al., (2015).
 
Statistical analysis
 
ANOVA was performed following Augmented design (Federer, 1961) using WINDOSTAT 8.5 version. Adjusted trait values of each of the individuals in BIP F₃ and F₃ progenies were estimated (Federer, 1961). Adjusted means were used for estimating eight QTs means, absolute range (SR), standardized range (SR) (highest trait value-lowest trait value/trait mean), variance, additive genetic variance (σ²_A) and narrow-sense heritability (NS-h²). Separate ANOVA were performed to partition total variation in BIP F₃ and F₃ progenies into sources attributable to ‘between’ and ‘within’ BIP F₃ (Kearsey, 1965) and ‘between’ and ‘within’ F₃ progeny (van Ooijen, 1989) variances, respectively. The σ²_A and NS-h² of eight QTs in BIP F₃ and F₃ progenies were estimated using the information from respective ANOVA. The significance of differences in QTs means and variances between BIP F₃ and F₃ progenies were tested using two-sample ‘t’ test and Levene’s  test (Levene 1960), respectively.
 
Impact assessment of F2 BIP mating on QTs dynamics
 
BIP F₃ progenies were compared with those of F₃ progenies with respect to per se performance (mean), AR/SR, variance, σ²_A and NS-h² of nine QTs.

Analysis of variance
 
Significant differences among the experimental progenies (BIP F₃ + F₃ progenies) for all the traits, except primary branches plant^-1, nodes raceme^-1 and fresh bean yield plant^-1 (Table 1) supported analysis of variance of BIP F₃ and F₃ progenies separately. Significant mean squares due to ‘between BIP F₃ progenies’ and ‘between F₃ progenies’ (Table 2) suggested sufficient variability in F₂ population attributable to segregation of genes for which the parents (HA 4 and GL 37) differed. Significant variability between BIP F₃ and F₃ progenies adequately provided statistical and genetical validity for comparative assessment of BIP F₃ and F₃ progenies for mean, standardized range, variance, σ²_A and NS-h² of eight QTs.
 

 

 
Impact of BIP mating on QTs mean
 
In F₂ population derived from crosses between diverse parents, genes controlling the QTs for which the parents differ would be in significant linkage disequilibrium (LD) (Hanson and Hayman, 1963). Despite the presence of LD between the genes controlling the QTs, means of BIP F₃ and F₃ progenies is not expected to alter (Kearsey and Pooni, 1996). Also, per se performance of progenies derived from F₂ BIP mating compared to those derived from inbreeding is not expected to change unless there exists non-additive gene effects associated with change in LD (Hanson and Hayman, 1963). In the present study, significantly delayed flowering with higher primary branching and raceme producing ability of BIP F₃ progenies compared to F₃ progenies (Table 3) could be attributed to involvement of non-additive action of genes controlling these traits. Higher means of BIP F₃ progenies for days to flowering, number of primary branches and number of racemes compared to those of BIP F₃ progenies could also be attributed to breakage of undesirable linkages which otherwise conceal genetic variation accumulation of favorable genes controlling these traits (Hanson, 1959; Bos, 1977; Stam, 1977). The results of present investigation are in conformity with those of Sharma and Kalia (2003) in garden pea and Chand (2000) in blackgram and Kampli et al., (2002), Singh (2004) and Jahagirdar et al., (2005) in chickpea. Contrastingly, F₃ progenies manifested higher nodes raceme^-1 compared to BIP F₃ progenies. Stam (1977) through his extensive theoretical investigation opined that environment variance might also tends to provide selfed progenies, a relative advantage over random mating-derived progenies especially for traits directly related to fitness (such as nodes raceme^-1 in the present study) of the individuals. However, BIP F₃ progenies were comparable to F₃ progenies for fresh pods node^-1, fresh pods plant^-1, fresh pod yield plant^-1and fresh bean yield plant^-1. These results could possibly due to inadequacy of one cycle of inter-mating to release genetic variability and/or linkage equilibrium (LE) of genes controlling these traits in F₁’s. Thus, per se performance of progenies derived from F₂ BIP mating and inbreeding varied with traits in dolichos bean depending on LD/LE of genes controlling traits.
 

 
Impact of BIP mating on QTs variance
 
Inter-mating F₂ individuals derived from diverse parents is expected to decrease and increase genetic variability for traits which are controlled by genes that exist in coupling phase and repulsion phase, respectively (Hanson and Hayman, 1963). While the decrease in genetic variability could be attributed to appearance of individuals with intermediate phenotypes, increase in genetic variability could be attributed to appearance of individuals with extreme phenotypes. In the present investigation, higher genetic variance in BIP F₃ than that in F₃ progenies (Table 3) suggested predominance of repulsion phase of linkage of genes controlling racemes plant^-1 and fresh pods node^-1.
       
Increase in QTs variances following F₂ inter-mating were also reported by Kampli et al., (2002), Singh (2004), Jahagirdar et al., (2005) in chickpea, Koli and Punia (2012) in aromatic rice and Gauthami et al., (2016) in maize. Hanson (1959) found that with two generations of inbreeding, about 65 per cent of chromosome lengths were left intact (no crossing over), whereas, only about 42 per cent was intact with inter-mating. Thus, F₂ BIP mating is expected to offer better opportunities than selfing to prevent early fixation of undesirable genes in homozygous condition in dolichos bean as demonstrated by Meredith and Bridge (1971) in cotton.
       
Variances for traits such as days to flowering, primary branches plant^-1, nodes raceme^-1, fresh pods plant^-1, fresh pod yield plant^-1 and fresh bean yield plant^-1 were comparable between F₃ and BIP F₃ progenies. While detection of increase/decrease in variance following F₂ BIP mating is an indication of repulsion/coupling phase of linkage between genes controlling QTs, failure to detect the same does not rule out the presence of linked loci (Hanson and Hayman 1963; Meredith and Bridge, 1971). Fairly conclusive evidence for the presence of linked loci controlling the inheritance of these QTs could be elicited by comparing the variances between (F₁ × P₁) and (F₂ × P₁), (F₁ × P₂) and (F₂ × P₂) and between (F₁ × F₁) and (F₂ × F₁) (Perkins and Jinks, 1970).
 
Impact of BIP mating on QTs range
 
It is expected that random mating would result in the distribution of genotypic values different from that expected in F₂ population. The distribution is less concentrated about the parental and F₁ values and thus, higher QTs range is expected (Hanson, 1959). In the present study, a wider standardized range of traits mean values of BIP F₃ progenies compared to F₃ progenies (Table 4) could be attributed to appearance of extreme phenotypes presumably due to disruption of repulsion phase of linkage (in F₁’s) between genes controlling racemes plant^-1, fresh pods node^-1 and fresh pod yield plant^-1. Dutta et al., (1987) have also reported enhanced range in QTs means following F₂ inter-mating in groundnut and they attributed it to recombination of genes from upper and lower ends of genotypic values. On the contrary, wider standardized range of QTs means of F₃ progenies  compared to BIP F₃ progenies could be attributed to appearance of intermediate phenotypes presumably due to breakage of coupling phase of linkage (in F₁’s) between genes controlling days to flowering, primary branches plant^-1, nodes raceme^-1, fresh pods plant^-1 and fresh bean yield plant^-1.
 

 
Impact of BIP mating on σ²_A and NS-h² of QTs
 
Additive genetic variances are generally inflated by coupling phase and deflated by repulsion phase consequent to F₂ BIP mating (Kearsey and Pooni, 1996). As the QTs are controlled by a large number of genes, very rarely all the ‘plus’ genes will be in complete association (which enhance phenotypic expression) or ‘plus’ and ‘minus’ genes will be in complete dispersion (which causes reduction in phenotypic expression) particularly those that are from the source material for breeding programmes (Kearsey and Pooni, 1996) and thus display coupling/repulsion phase of linkage. However, dispersion does not automatically lead to predominance of repulsion phase as a few gene pairs will be in repulsion phase while others will be in coupling phase. Nevertheless, dispersion of genes between the parents always reduces σ²_A. The extent to which reduction in σ²_A occurs is more pronounced when ‘plus’ and ‘minus’ genes alternate in the sequence in a haplotype (Kearsey and Pooni, 1996). Higher magnitudes of σ²_A and hence NS-h² in BIP F₃ than those in F₃ progenies in the present study (Table 4) suggested predominance of coupling phase of linkages of genes (in F₁’s) controlling all the traits, except days to flowering and primary branches plant^-1 for which they were comparable between BIP F₃ and F₃ progenies. Kampli et al., (2002), Singh (2004) and Jahagirdar et al., (2005) in chickpea have also reported higher magnitudes of σ²_A in BIP F₃ than those in F₃ progenies. In self-pollinated crops like dolichos bean, selection is generally among pure-lines and hence exploits σ²_A and σ²_AA epistasis.

Random mating in F₂ population was effective in increasing the means, variances, σ²_A and NS-h² of racemes plant^-1, fresh pods node^-1, fresh pods plant^-1 and fresh pod yield plant^-1 in dolichos bean. Considering that random mating followed by selfing is a method of genetic improvement for future breeding rather than for immediate use (Hallauer 1984), our results suggest long-term genetic gain and better prospects of deriving superior pure-lines with desired traits/combination of traits from inter-mated population in dolichos bean.