Correlation Coefficient and Path Analyses of Yield and Yield Related Traits of Korean Double Haploid Rice for Germplasm Improvement in Nigeria

The success of plant breeding depends on the availability of genetic variation, knowledge about desired traits, and efficient selection strategies that make it possible to exploit existing genetic resource. This study was carried out at Africa Rice Center, International Institute of Tropical Agriculture (IITA) Ibadan, Nigeria, and 239 double haploid lines from Republic of Korea were established along with one of the rice mega varieties (FARO 44) as check. The experiment was conducted using Alpha lattice design with four blocks each planted with 60 entries in two replications. Analysis of variance revealed highly significant differences (P ≤ 0.001) among the genotypes indicating existence of variation among the genotypes. Correlation coefficient of the yield and its association traits revealed significantly positive correlation of grain yield with number of tillers, plant height, days to 50% flowering, panicle length, effective tillers, leaf area, leaf area index, fertility, biomass, panicle weight, number of grains per panicle and grain yield per plant, hence, selection for these traits can improve yield. Path coefficient analysis revealed that days to 50% flowering, leaf area index, fertility, biomass, panicle weight, number of grains per panicle and grain yield per plant exhibited positive direct effect on grain yield. Among all traits examined, panicle weight had the highest significant positive correlation and high positive direct effect. Stepwise regression showed that characters such as panicle weight, grain yield per plant, flag leaf, days to 50% flowering, effective tillers and 1000 grain weight contributed more to the total grain yield. Therefore, selections for the aforementioned characters will assist breeders in making good improvement in rice grain yield.


Introduction
Rice (Oryza sativa L.), one of the most important crops and second most widely cultivated cereal in the world, is a staple food for more than half of the world's population [1] The world population is expected to reach 8 billion by 2030 and rice production must be increased to meet the growing demand. [2]. With the ever-increasing rate of population, the demand for rice is increasing rapidly [3]. Hence, increase in rice production is very important in the food security and poverty alleviation [4]. The economic trait of rice is the grain yield which is influenced by complex genes and many yield contributing factors and is also highly influenced by the environmental conditions [5]. Breeding strategy in rice mainly depends upon the degree of association among characters as well as its magnitude and nature of variation. Therefore, the knowledge of association of yield components is of great importance to plant breeders as it helps in selection of crop for better quality [6].
Correlation and path coefficient analysis will give a better insight of association among the characters governing the grain yield. Correlation analysis provides information on the nature and extent of association between pairs of metric characters and helps in selection for crop improvement. While path coefficient analysis, allows separation of direct effects and indirect effects through other attributes by partitioning the correlations [3]. Studies on path analyses therefore provide useful information regarding the direct and indirect effects of yield component characters on grain yield and thus aid in the identification of better selection criteria for yield improvement [7]. Path coefficient analysis assists plant breeders in identifying traits that should be considered in improving a complex trait [8] Rice grain yield, being a complex trait, depends on the various yield contributing traits like plant height, days to 50% flowering, number of productive tillers per plant, panicle length, panicle weight, number of grains per panicle and weight of 1000 grains [1]. Information on association of characters, direct and indirect effects contributed by each character towards yield will be an added advantage in aiding the selection process. Correlation and path analysis established the extent of association between yield and its components and also bring out relative importance of their direct and indirect effects, thus giving an obvious understanding of their association with grain yield and this also assist the breeders in their selection strategies to improve grain yield. [9]. Therefore, the objectives of this study were to determine the nature and degree of association between yield and its component characters and their direct and indirect effects on grain yield in rice.

Experimental Location
The experiment was conducted at experimental field of African Rice Centre at International Institute of Tropical Agriculture (IITA), Old Oyo road, Ibadan, Oyo state, Nigeria. During 2020 cropping season. IITA is located at longitude 7°30'8''N, latitude 3°54'37''E and at elevation 243m above sea level [10] Experimental materials and design The experimental materials for the study consist of 239 double haploid lines from Republic of Korea were established along with one of the rice mega varieties (FARO 44) as check.
The experiment was conducted using Alpha lattice design with four blocks each planted with 60 entries in two replications. Seeds of the double haploid lines were sown in the wet prepared nursery bed to raise seedlings needed for the trail, after 25 days the seedlings were transplanted to the field. A single row of size 0.2m × 3 m was used as plot at spacing of 20 cm ×20 cm within and between rows and chemical fertilizer NPK (15:15:15) was applied as a basal application of 200 kg/ ha (N 2 P 2 O 5 and K 2 O). Urea was applied in two splits at the rate of 65 kg/ha at tillering stage and the second application at the rate of 35 kg/ha at the beginning of panicle initiation stage. Weed control was done twice using selective herbicide Vespanil Plus (250 ml/ 20 liters of water) at early stage of crop development and before flowering.
Data collection Data was collected at appropriate stage of the crop development. The agronomic characters were measured by randomly selected plants from each experimental unit (row). The Standard Evaluation System (SES) by International Rice Research Institute (IRRI) for Rice reference manual was used for all trait measurements except were stated otherwise [11]. The data collected were plant height, number of tillers per plant, days to 50% flowering, leaf area, leaf area index, flag leaf length, panicle fertility (%), productive tillers per plant, panicle length, panicle weight, biomass, number of grains per panicle, 1000 grain weight, grain yield per plant and grain yield per plot (kg/ha). Leaf area (LA) was determined using a leaf area meter (li-3100, Lincoln, NE USA). And leaf area index was calculated in meter square as described by [12] Data Analysis Data analysis was done using Analysis of Variance technique (ANOVA). Differences were declared statistically significant when (P ≤ 0.05). Correlation coefficient among all the traits as well as Stepwise regression and path coefficient of direct and indirect contribution of traits on grain yield were analyzed using Statistical Analysis System [13]

Agronomic Trait Evaluation of the Genotypes
Analysis of variance (ANOVA) results revealed significant differences among genotypes, block within replication and replication for all traits under consideration. Genotypes were highly significant for all traits under consideration. Significant variation was observed in block within replications for most of the traits (Table 1). These results indicate the existence of variation among the genotypes for yield and yield components under studied.

Correlation Coefficient Analysis
The correlation coefficients between grain yield and its component characters and the inter correlation among different traits that play an important role in selection of desirable genotypes ( Table 2). All the traits except flag leaf (-0.018) and 1000 grain weight (-0.039) showed different level of significant positive correlation with grain yield. Grain yield, grain yield per plant and panicle weight correlated with all the traits under consideration and number of tillers had negative but significant correlation with most of the traits under consideration ( Table 2).
Days to 50% flowering and panicle length recorded positive and significant correlation with other traits except positive and nonsignificant correlation registered between days to 50% flowering and 1000 grain weight (0.062) and between panicle length and effective tillers (0.033). Flag leaf showed positive and significant correlation with leaf area (0.479), leaf area index (0.479), biomass (0.314), panicle weight (0.119), number of grains per panicle (0.094) and 1000 grain weight (0.328). Positive and non-significant observed with fertility (0.052) and grain yield per plant (0.079) whereas negative and significant correlation noticed between flag leaf and effective tillers (-0.111). (Table 2).
Effective tillers are tillers harvested at maturity which contributed to the total grain yield had positive and significant association with fertility (0.224), biomass (0.217), panicle weight (0.468), number of grains per panicle (0.195) and grain yield per plant (0.506). Effective tillers showed negative and significant correlation with 1000 grain weight (-0.090) ( Table 2).
Fertility showed positive and significant correlation with biomass (0.123), panicle weight (0.358), number of grains per panicle (0.411) and grain yield per plant (0.442). Biomass showed positive and significant correlation with all other grain yield components except number of tillers ( Table 2).

Path Coefficient Analysis
The direct and indirect effects of different yield components as partitioned by correlation coefficient using path analysis presented in (Table 3).

Direct Effect
The result of direct effect as shown in (

Stepwise Regression Analysis
Stepwise regression analysis for variables (Table 4) had highly significant (P < 0.001) for panicle weight, grain yield per plant, flag leaf and days to 50% flowering and significant (P < 0.05) for effective tillers and 1000 grain weight.

Correlation Coefficient Analysis
Good knowledge of germplasm diversity and genetic relationships among breeding materials could be a valuable tool for crop improvement. Genetic diversity determines the inherent potential of heterosis and frequency of desirable recombinants that could be present in advanced generations [14]. A large number of anther-culture genotypes and the improved variety used in this study showed high genetic diversity and genetic variation among the genotypes. The presence of high genetic variability could be due to diverse materials used in this study as well as environmental influence, this corroborate the report and similar genetic variations have been reported who observed significant genetic differences for all characters studied in different rice genotypes [15,16,17,18].
The observed positive and significant correlation of grain yield with rice yield components have also been reported [8,12]. Effective tiller is the harvested tillers at maturity of the rice crop, the higher the number the higher the total grain yield. The high positive correlation of grain yield and effective tillers indicates an effective index in increasing the total grain yield and genotypes exhibiting this trait are potential for germplasm improvement similar reports on copper and iron compounds on rice [19,20]. Panicle weight showed positive and significant correlation with number of grains per panicle similar results were reported [21,22]. Therefore, priority should also be given to those traits that had high positive correlation with grain yield such as grain yield per plant and panicle weight while making selection for yield improvement. Potential double haploid lines have been identified for rice improvement in Nigeria.

Path Coefficient Analysis Direct Effect
Path coefficient analysis provides an effective means of determining the magnitude of each contributing trait to the grain yield either directly or indirectly. Breeding new variety with yield potential is one of the main objectives of plant breeders and there are several yield components controlling the yield potential directly or indirectly. Leaf area index is responsible for photosynthesis efficiency and it plays an important role in grain yield production [23]. High positive direct effect of leaf area index was recorded similar report this indicates that leaf area index is an important criterion for increasing rice grain yield [6].
Panicle weight is one of the yield components, panicles has to be long to accommodate more grains, which translate to higher panicle weight. Panicle weight had high positive direct effect on grain yield similar reports [24,25]. Therefore, genotypes with high panicle weight could possess high yielding ability and good for crop improvement. Grain yield per plant is one of the determinants of high grain yield. Results showed that grain yield per plant had high positive direct effect on yield.
Apart from genetic influence, other factors also affect days to 50% flowering such as plant spacing, irrigation and weeding induce early flowering. Days to 50% flowering influences maturity period and showed positive direct effect on yield, the result corroborate the reports [26,27]. Biomass accumulation is important for grain yield formation and improvement of rice yield could results in the increase of biomass. Biomass showed positive direct effect on yield [28].
Number of grains per panicle is important yield contributing trait that require long panicle to accommodate more grains. Number of grains per panicle found to be direct effect and positive on yield this result corroborate [29,30,31] Some traits expressed negative direct effects on yield such as leaf area, flag leaf, effective tillers, 1000 grain weight, number of tillers, plant height and panicle length (Table 3) suggesting that these traits are not reliable in the selection process for grain yield. Direct negative effects of plant height on grain yield were also reported [5, 25, 30, 31 32] for plant height, but positive direct effect of plant height on grain yield which is contradictory with the present study was reported [26,33] Indirect Effect Path coefficient also indicates traits that have indirect effects with grain yield. It could be suggested that breeders should also pay more emphasis on those traits with positive indirect effects on grain yield. Plant height showed positive indirect effect through fertility where the result in conformity with [34] and as well, plant height expressed negative indirect effect through panicle length corroborate the report [35]. The results showed that days to 50% flowering had indirect positive effect on grain yield through panicle weight, panicle weight and number of grains per panicle similar results were reported by [25,36]. Panicle length showed positive indirect effect on grain yield through days to 50% flowering, panicle weight and number of grains per panicle.
Effective tillers expressed positive indirect effect through days to 50% flowering, panicle weight, number of grains per panicle similar results were reported [36]. Panicle weight exhibited negative indirect effects through plant height and indirect effect of number of grains per panicle on grain yield were negative through panicle length and effective tillers, the negative indirect effects may not be very beneficial in selection for crop improvement and similar reports [3, 35 37]. The high residual effect (0.605) in this study indicated that beside the characters studied, there are some other attributes for grain yield. This result is in concurrence with findings reported [38,39]. Hence, considering the nature and magnitude of character association and their direct and indirect effects, it can be deduced that simultaneous improvement of grain yield is possible through manifestation of leaf area index, panicle weight, grain yield per plant and days to 50% flowering. High direct effects of these traits therefore, appeared to be the main factor for their strong association with grain yield. Hence, these traits should be considered as important selection criteria in all rice improvement programme and direct selection for these traits is recommended for yield improvement.
Stepwise Regression Analysis Stepwise regression is a semi-automated process of building a model by successively adding or removing variables based on the t-statistics of their estimated coefficients [40]. Based on the stepwise regression analysis for variables, indicates highly significant (P < 0.001) for panicle weight, grain yield per plant, flag leaf and days to 50% flowering and significant (P < 0.05) for effective tillers and 1000 grain weight. This showed that these characters had contributed highly to grain yield and could be used as criteria for selection in rice breeding programme. These results are also in agreement with the findings [41,42]. Panicle weight had the highest R-Square and explained 51% of the total variations relative to grain yield. Also, this trait had strong positive direct effect (0.457) and the highest correlation coefficient (0.715) with grain yield. These results showed that regression analysis agreed with path analysis in this study.

Conclusion
The study showed the existence of a considerable level of diversity among the rice double haploid genotypes, which could be used for rice improvement in Nigeria. Correlation coefficient analysis revealed that grain yield had high positive and significant correlation (P ≤ 0.001) with panicle weight, grain yield per plant, number of grains per panicle and biomass. Path coefficients analysis revealed the importance of panicle weight, grain yield per plant, number of grains per panicle and biomass as selection criteria for effective grain yield improvement. Path analysis showed that leaf area index, panicle weight, grain yield per plant and days to 50% flowering had high direct effects on grain yield. Among these characters, panicle weight had very significant positive correlation and high positive direct effects. Hence, selection for this character could accelerate improvement in grain yield and yield components. Stepwise regression showed the magnitude of the characters panicle weight, grain yield per plant, flag leaf, days to 50% flowering, effective tillers and 1000 grain weight as more contributed factors to grain yield. Therefore, selection base on these characters will be more effective in rice breeding and for germplasm improvement in Nigeria.