Genetic Variability, Correlation and Path Coefficient Analysis in Advanced Matromorphic Generations of Garden Pea (Pisum sativum L.)

Garden pea [Pisum sativum (L.) var. hortense], is the widely cultivated leguminous  crop worldwide. It has been a valuable model crop for genetic research since the time of G.J. Mendel (Bar and Ori, 2015). The presence of genetic variability for the desirable traits in garden pea is considered crucial in the crop improvement programme for the development of commercial cultivars (Tiwari and Lavanya, 2012). Garden pea is highly self pollinated crop hence, the production and maintenance of homozygous lines constitute the raw material in pea improvement program. Matromorphy is a novel technique that facilitates in obtaining instant homozygous lines from heterozygous base population. In order to formulate an optimal breeding program targeted at harnessing the natural variability of the original population, variability indices such as GCV and PCV, heritability (h²) and genetic advance (GA) are crucial (Raj et al., 2020). The phenotypic variability varies with environmental factors, whereas genetic variability remains constant and advantageous in breeding programme for selection or hybridization. Heritability (h²), which represents the efficiency of genotype selection based on phenotypic performance has speculative aspect in crop breeding (Songsri et al., 2008). The yield and quality traits can be selected to best possible extent with the available information about correlations between such traits followed by detailed knowledge of the association between the relevant characters obtained by path coefficient analysis (Devi et al., 2017). An appreciable yield enhancement in garden pea can be achieved by understanding the interactions between yield and contributing traits at genotypic and phenotypic levels and the selection of highly heritable traits. Therefore, the current investigation was undertaken with the aim of estimating genetic variability, heritability and correlation coefficients among different horticultural traits in garden pea to identify superior genotypes for their utilization in breeding programme with the aim to develop new varieties with higher yield.

The present investigation involving 22 garden pea genotypes (12 matromorphic progenies (Mat3/4) and 10 parental lines) was conducted at the Experimental Farm-I, Division of Vegetable Science and Floriculture, SKUAST- Jammu during Rabi season of 2020-21 and 2021-22. The plants were grown in 3 × 2.5 m² plot with 45 cm × 10 cm (R × P) with three replications in randomized complete block design (RCBD) following all recommendations (Anonymous, 2020). The data on various horticultural traits viz., days to 50% flowering (DFF), node at which first flower appears (NFFA), plant height (cm) (PH), number of primary branches per plant (NPB), internodal distance (cm) (ID), number of pods per node (NPPN), pod length (cm) (PL), pod width (cm) (PW), number of seeds per pod (NSPP), number of pods per plant (NPP), days to first picking (DFP), average pod weight (g) (PW), pod yield per plant (g) (PYPP) and shelling percentage (SP) was recorded on five plants of each genotype which were chosen randomly. The average mean for each trait over three replications was computed for each genotype and analysed statistically to determine analysis of variance (ANOVA) for all the traits as recommended by Goulden (1959). The phenotypic coefficient of variation (PCV) and genotypic coefficient of variation (GCV) was estimated using Burton’s approach (1952). Broad sense heritability (h²bs) was assessed as per the procedure given by Lush (1949). Genetic advance (GA) and GA as percent of mean (GAM%) were calculated using the formula given by Johnson et al., (1955) and Genotypic (r_g) and phenotypic correlation coefficient (r_p) were computed using procedure recommended by Miller et al., (1983) and Johnson et al., (1955). Path coefficient analysis was done according to Dewey and Lu’s method (1959).

The ANOVA (Table 1) showed significant substantial differences between the tested genotypes for three traits viz., DFF, NFFA and PH under investigation. This variation in genotypes for all the traits may be a result of genotypic diversity, environmental effect and their interaction (G × E). Present results concur with the findings of (Barcchiya et al., 2018; Singh et al., 2019; Thapa et al., 2020).
 

 
Genetic Variability, heritability and genetic advance
 
For a successful selection programme of any crop, assessments of variability involving genetic characteristics such as genotypic variance, broad sense heritability (h²) and genetic advance are essential (Thapa et al., 2020). The higher PCV values with greater magnitude were recorded for all the studied traits as compared to GCV values indicating that the genetic material used was highly variable due to both genotypic effect and environmental influence (Table 2). However, there wasn’t much of a difference between GCV and PCV values, suggesting that character expression was more influenced by genetic factors than environmental ones. As a result, the selection might be made based on phenotypic performance, which would allow for crop improvement. The different PCV and GCV values of the traits due to environmental effect was also recorded by Bijalwan et al., (2018) and Azam et al., (2020).
 

       
The maximum PCV and GCV value of >20 per cent were found for PYPP (37.84 and 37.72%), NPP (31.23 and 31.07%), NPPN (29.95 and 29.84%), PH (24.21 and 24.07%) and PW (20.92 and 20.83%), respectively. As has been observed in the present study maximum PCV and GCV values of these traits was also reported by various workers (Katoch et al., 2016; Singh et al., 2019 and Thapa et al., 2020).
       
Furthermore, it might be feasible to ascertain the extent to which a character is passed on from parent to progeny by assessing heritability (Hanson et al.,1956). In the present investigation, broad sense heritability ranged from 55.88 per cent (SP) to 99.41 per cent (PYPP) and high heritability (> 60%) was exhibited by thirteen traits except for shelling percentage. Kassaye (2006) reported that high broad sense heritability (h²bs) along with GA would be a useful method for the selection of superior genotypes. High heritability (h²) and GAM (>20%) was noticed for all the traits except PL and SP showing the preponderance of additive gene effect. The maximum value of GAM (77.48) was shown by PYPP followed by NPP (63.65), NSP (61.26), PH (49.31), APW (42.72), DFF (33.87), NFFA (30.51), ID (29.41), DFP (28.69), NSPP (26.28), NPBP (25.07), PW (24.52), PL (15.89) and SP (14.35) except NPBP which showed maximum heritability (95.20%)  but low GA (0.53) (Table 2). Out of all the traits under study, five traits viz., PYPP, NPPN, PW, NPP and PH recorded maximum values for heritability (h²bs), GCV and GAM%  thus, depicting the effect of additive gene action on these traits and therefore, may be helpful for efficient selection. The findings of numerous studies are congruent with the current results (Georgieva et al., 2016; Singh et al., 2019; Thapa et al., 2020).
 
Correlation coefficients
 
The correlation coefficient, which offers a balanced assessment of the extent of relationship among two traits, aids in determining the type and amount of association between yield and its components (Singh et al., 2018). The genotypic correlation coefficients between distinct traits were similar to the phenotypic correlation coefficients in the experiment in terms of sign and nature. However, genotypic correlations were larger than phenotypic correlations.
       
PYPP had highly significant and positive correlation with NFFA (r_g= 0.424, r_p= 0.399), PH (r_g= 0.523, r^p= 0.520), ID (r_g= 0.565, r_p= 0.549), NPPN (r_g= 0.439, r_p= 0.434), PL (r_g= 0.529, r_p= 0.445), NSPP (r_g= 0.662, r_p= 0.646), NPP (r_g= 0.968, r_p= 0.958), DFF (r_g= 0.468, r_p= 0.458), PW (r_g= 0.710, r_p= 0.706) and SP (r_g= 0.672, r_p= 0.495) both at genotypic and  phenotypic levels (Table 3). DFF showed significant and positive correlation with NFFA, PH, NPBP, ID, PL and DFP. These traits were identified as most important component traits and were linked positively with PYPP implying that pod yield would increase by simultaneous selection for these traits. Earlier reports  have also confirmed existence of strong positive correlation of  PYPP with PH, NPPN, PL, NSPP and NPP (Kumar et al., 2019), days to first picking and shelling percentage (Rahman et al., 2019), average pod weight (Tiwari et al., 2020).
 

 
Path co-efficient analysis
 
Path coefficient analysis is an efficient approach to separate correlation coefficients into direct and indirect component effects since it assesses the direct impact of one variable on the other. This strategy is used to investigate the source and effect of a relationship between variables. Correlation studies give a greater understanding of the causes and effects of relationships between different pairs of component traits and the main trait when paired with path coefficient analysis (Verma et al., 2021). In the present study, PYPP was taken as a resultant (dependent) variable while 13 other traits were independent variables. The cause and effect relationship of PYPP and its related traits have been given in Table 4 (Fig 1). The results revealed that ID had maximum positive (1.713) and direct effect on PYPP followed by NSPP (1.124), APW (1.022), NFFA (0.682), DFF (0.483), NPPP (0.368), NPPN (0.291) and PW (0.234) suggesting that they are the major contributors to pod yield per plant and that if other characters remain constant, an increase or decrease in each of these characters will reflect in increase or decrease in pod yield. Positive and direct effect of NSPP, NPP, PW and DFF on PYPP was reported earlier also in field and garden pea (Tofiq et al., 2015; Kumar et al., 2019; Verma et al., 2021). PH had highest indirect effect (1.654) via ID on PYPP. The residual effect was 0.041 which inferred that the character under study contributed 95.90% to the pod yield.
 

In the present study, it was reported that among all the traits studied, node at which first flower appears, internodal distance, number of pods per node, no. of seeds per pod, no. of pods per plant and days to first picking revealed significant positive correlation and had direct impact  on pod yield per plant. This indicated that yield improvement is associated with these traits and selection would result in yield improvement in pea.

None.