Multi-variate analysis for yield evaluation in rice (Oryza sativa L.) genotypes

Field experiment was carried out at the experimental field of the Department of Plant Science and Biotechnology beside the screen house, to study yield evaluation in rice genotypes. The plot was laid out in a randomized complete block design (RCBD) and the treatments replicated three times. Seven rice genotypes sourced from the department crop improvement unit were sowed in a plot size of 5×1m. The study was conducted in 2016/2017 planting seasons. Data were collected on yield component traits i.e. number of tillers, plant height, number of days to panicle initiation, number of days to maturity, panicle length, panicle weight, number of spikelet per panicle, number of grain per panicle, 1000 grain weight, grain yield per hill and flag leaf length. The result of factor analysis indicated that the first factor was positively loaded for yield component trait however; the first six principal components jointly accounted for 98.99% of the total variation among the genotype studied. High levels of variability expressed among the varieties suggested that further improvement in the varieties is possible.


Introduction
Rice, Oryza species is a grass belonging to the family of Poaceae. As an annual plant the height ranges between 36 -150 cm. It is the most consumed cereal grain after wheat globally (FAO, 2004) and are cultivated in most countries of the world. It has about 20 different species, of which the cultivated varieties are Oryza sativa belonging to Asia and Oryza glaberrima Africa (Vaughan and Morishima, 2003). The major resource of plant breeders is the genetic variability in gene pool accessible to the crop of interest without which improvement may not be possible . The success of crop improvement programs is highly dependent on the efficient manipulation of the germplasm that results in sufficient genetic variability. Morphological markers have played a major role in crop improvement since the beginning of modern breeding programs (Mignouna et al., 1996). When more sophisticated attributes are required, isozyme, restriction fragment length polymorphism (RFRP) and other molecular level analysis are used (Ng and Padulosi, 1992;Flavel, 1991). Information from the use of these morphological makers is used by breeders and taxonomists to study genetic diversity with the aim of identifying variations and correlations among living organisms (Ng and Padulosi, 1992). The evaluation of germplasm frequently includes recording traits of agronomic interests, such as resistance to pests and diseases and tolerance to physiological stresses that are influenced by environment. These data are most sought after by plant breeders according to (Ng and Padulosi (1992); Peters and Williams, 1984). Continuous evaluation of germplasm contributed to variation in the rice populations. Therefore this study is conducted to identify traits that contribute to variability and for their possible exploitation in breeding programs.

Material and methods
The location of the experiment was Akungba Akoko, longitude 50 501E and latitude 9° 051 N, at the experimental fields of the Department of Plant Science and Biotechnology, to assess genetic variations that exist among 7 improved rice varieties. The experiment was conducted during the rainy seasons of 2017 in an upland agro-ecology. The improved varieties were obtained from the Crop improvement section, of the Department of Plant Science and Biotechnology, Adekunle Ajasin University, Akungba Akoko. The plot was laid out in a randomized complete block design (RCBD) and replicated 3 times. Each entry was planted in a plot size of 5 × 1 m and the spacing adopted was 0.2 m × 0.2 m. Cultural operations were carried out such as weeding, and NPK 15:15:15 fertilizer was applied. Data were collected on yield component traits i.e. number of tillers per hill, plant height, number of days to panicle initiation, number of days to maturity, panicle length, panicle weight, number of spikelets per panicle, number of grains per panicle, 1000 grain weight, grain yield per hill, flag leaf length. The data was analyzed using Genstat version 2004.

Results
The analysis of variance (ANOVA) for all the traits is shown in Table 1. Significant effects of genotypes were observed for all the studied traits. Number of tillers per hill, plant height, number of days to maturity, number of days to flowering, panicle length, panicle weight, number of spikelets per panicle, number of grains per panicle, 1000 gain weight, yield per hill and flag leaf length were significantly different at p≤ 0.05.    Heritability estimates the relative contributions of differences in heritable and non-heritable factors to the total phenotypic variance in a population. It is an important concept in quantitative genetics, particularly in selective breeding. Heritability in broad sense (h2) varied from 0.42% to 99.96% for flag leaf length and1000 grain weight respectively. All the traits studied had high heritability (greater than60%) exceptfor panicle weight, flag leaf length and number of tillers. The high value of genetic advance (GA) was recorded in number of grains per panicle (71.17) and low (0.12) for flag leaf length. Genetic advance as per cent mean (GAM) range from 0.32% for flag leaf length and132.51%for yield per hill.  Results of simple correlation analysis among eleven quantitative traits were presented in Table 4

Discussion
The study of genetic variance and other genetic parameters had help in no small measure in formulating a suitable breeding program for improvement of the crop especially in rice. It is a prerequisite for successful crop improvement programs. A genetic diversity study is an important component in characterization of breeding materials which is very key in crop improvement. More variability observed in the base population is a veritable effort in the choice of parent for hybridization in the improvement of crops.
Based on above understanding, variance analysis revealed high significant differences among the genotypes for all the traits indicating huge genetic variability that existed among the genotypes, this appears that further improvement through selection for these traits may be effective, this results are in agreement with those found by Ogunbayo et al., The most important function of the heritability in the genetic study of quantitative traits is its predictive role in the selection process and as a guide to breeding value (Falconer and (2015).Correlation analysis measures the extent of association between two traits or the extent to which they vary together; it is an important tool that can be used to determine the direction of selection in a breeding program. Based on the above understanding significant and positive correlation between two traits suggests that the traits can be improved simultaneously in a selection program, the present study showed that number of tillers exhibits a highly significant positive correlation between number of days to maturity, number of days to flowering and yield per hill indicating that simultaneous selection for these traits would result in the improvement of yield. Similar findings were earlier reported by Gulzar  Factor analysis and principal component analysis identified some similar traits as the most important for grouping the variation that existed among the studied rice genotypes and these includes, plant height, number of grains per panicle, one thousand grains weight, flag leaf length, number of days to flowering, panicle length, number of days to maturity and yield per hill. The similarity between the two techniques had been reported earlier in okra by Ariyo (1993), rice by Nassir and Ariyo (2007). Meanwhile, the two techniques produced similar result, but their underlying principles are different from each other. Principal component analysis does not rely on any statistical model and assumptions but factor analysis does. It is also important to note that factor analyses suffer from some disadvantages such as absence of error term and the dependence upon scale used to measure the variables (Bartual et al., 1985). A combination of the identified traits will give a good description of the variability and hence discriminate among the genotypes.

Conclusion
Principal component analysis was utilized to examine the variation and to estimate the relative contribution of various traits for total variability. The PCI showed 53.94%, while, PC2 and PC3 exhibited 29.90% and 10.72% variability. It can be concluded that principal component analysis highlights the characters with maximum variability. The varieties that are higher yielding and stable are selected for farmers in the location for enhance productivity. In the experiment carried out, NERICA 8 can be recommended to farmers because it has a greater yield, has faster date of maturity and more grain per panicle than the other genotype.