Regression Analysis, Heritability and Inter-Generation Correlation in Wheat (Triticum aestivum L.)

In the present study F5 and F6 generation of crosses between WH 711 and WH 542 of bread wheat (Triticum aestivum L.) was evaluated for yield and its contributing traits under timely and late sown condition. Parent progeny regression analysis was undertaken to determine the genetic potential transferred from one generation to next generation. The significant and positive intergeneration correlation and regression was observed for character like grain yield per meter, grain length, plant height, harvest index, grain weight per spike, grain breadth, number of tillers per meter and biological yield per meter. It indicates the chances of selecting high yielding segregants at early generations. The narrow sense heritability for different characters based on parent offspring regression coefficient was moderate for grain yield per meter (78.80%), grain weight per spike (68.60%), harvest index (63.20%) and number of tillers per meter (54.00%) under both timely and late sown conditions. Higher the narrow sense heritability, higher is the chances of improving the characters through selection.


INTRODUCTION
Wheat is the world's most important crop that excels all other cereal crops both in area and production, thereby providing about 20.0 per cent of total food calories for the people of the world (Nukasani, Approximately 90% of the total wheat production of India is contributed by the five states i.e., Uttar Pradesh, Punjab, Haryana, Madhya Pradesh and Rajasthan (Ramdas, Singh, & Sharma, 2012). In Haryana, wheat was grown over an area of 25.10 Lha, with a production of 11.65 mt and productivity of 4.64 t/ha during 2018-2019 (Anonymous, 2019). In view of increasing population and global demand there is an imperative need to increase 40-60% of wheat production, to meet out the requirement of developing world in the coming 40 years (Goutam, Kukreja, Yadav, Salaria, Thakur, & Goyal, 2015). Indirect selection has been shown to be more efficient, less efficient, or equally efficient compared to direct selection when selection was practiced to improve a trait in one environment by selecting the trait in another environment (Sinebo, Gretzmacher, & Edelbauer, 2002). Efficiency of indirect selection compared to direct selection depends on the heritability of the trait used as selection criterion and on correlation between this trait and the indirectly targeted trait (Falconer & Mackey, 1996, Hannachi, Fellahi, Rabti, Guendouz, & Bouzerzour, 2017. Regression analysis is the better way to make crop yield prediction (Massart, Vandeginste, Buydens, De jong, Lewi, & Smeyers-Verbeke, 1997, Qingwu, Zhu, Musick, Stewart, & Dusek, 2006. The degree of dependence of one variate on the other is measured by regression coefficient. Regression coefficient was estimated on the basis of parent-offspring regression. Parent progeny regression is a method commonly used for estimating the amount of genetic potential transferred from parent to progeny. Inter-generation correlation studied by using parent offspring regression which helps in estimating the extent of transferring the genetic potentials of the character from one generation to other generation. Ibrahim and Quicke (2001) evaluated F 3 and their F 4 progeny derived from two crosses involving heat tolerant and heat sensitive genotypes to study the genetic variability for heat tolerance measured by MTS. The present investigation was aimed at studying the response of selection for yield and its component characters through parent progeny correlation and regression method between F 5 and F 6 generations.

MATERIALS AND METHODS Planting material
A total of 238 progenies each of the F 5 and F 6 generations along with parents (WH 711 / WH 542) were planted at the experimental area of Wheat and Barley Section of the Department of Genetics and Plant Breeding CCS Haryana Agricultural University, Hisar, during Rabi season of 2016-2018under two planting conditions, timely (November 6, 2016and November 8, 2017 and late sown (December 9, 2016 andDecember 8, 2017) conditions. The seed of each progeny was sown in a row of 2.5 meter length, with a spacing of 20 cm in a randomized complete block design (RBD) with two replications. Recommended package of practices for the agro-climatic zone were followed to raise the crop experiment.

Statistical analysis Regression coefficient
The degree of dependence of one variable on the other is measured by regression coefficient (b). For regression analysis, parent progeny regression coefficients were calculated by regression of the mean value of a character in the progeny (F 6 ) upon the value of a character in the parent (F 5 ) (Lush, 1940). The data was standardized by dividing individual values by respective standard deviation values and then the regression coefficient was calculated: The way to test the null hypothesis (b=0) is through the application of -t‖ test: Then, this t value was compared with t value from the table at the desired level of significance with the degree of freedom of the S.E. (b) Narrow sense heritability Based on parent offspring regression, narrow sense heritability was calculated as suggested by (Smith & Kinmann, 1956).
H (n.s.) = 2 x b Where, b = regression coefficient Intergeneration correlations Intergeneration correlation coefficients (r) were calculated for each trait between F 5 and F 6 generations. In each case progeny means (y) of F 6 generation were regressed on the individual plants (x) of F 5 generation. The observed value of correlation coefficient was compared with the tabulated value for (n-2) degree of freedom. The tabulated values used are of Fisher and Yates (1963). If the observed value is more than the tabulated one, the correlation coefficient is said to be significant.

RESULTS AND DISCUSSION Narrow sense heritability
Narrow sense heritability estimates helps to understand the nature of inheritance of a trait from generation to generation. Narrow sense heritability based on parent offspring regression analysis of F 6 generation over F 5 generation for 19 traits under timely sown and late sown conditions are presented in Table 1

Regression coefficient and Intergeneration correlation
The success of early generation selection usually depends on high correlation between the performance of the selected genotypes in one generation and their performance in the next generation (Barma, Islam, Hakim, & Sarker, 2011). Intergeneration correlation studies by parent-offspring regression method gave an idea about the extent of transmission of the genetic information from one generation to another generation. The parent progeny correlation and regression between the two generations is very useful for the selection of superior recombinants in the segregating population, to develop new and improved genotypes (Kumar, Kumar, Singh, Samita, Chaudhary, Sheokand, & Kumar, 2020). The present investigation was aimed at studying the response of selection for yield and its component traits through parent progeny correlation and regression method between F 5 and F 6 generations. The degree of dependence of one variable on the other is measured by regression coefficient. The parent progeny regression analysis was calculated on F 6 over F 5 . The regression coefficient b was calculated by using the formula suggested by Lush (1940).  (Table 2). Intergeneration correlation analysis was done using parent-offspring regression to transfer of characters from one generation to other generation. Under timely sown conditions highly significantly intergeneration correlation was observed for grain yield per meter (0.344**), grain length ( Table 2. The F 5 generation depicted highly significant and positive intergeneration correlation and regression with F 6 generation for plant height, number of tillers per meter, grain weight per spike, grain length, grain breadth, grain yield per meter, biological yield per meter and harvest index under both environments which indicated that these traits had high heritability, with more additive effect. Laala, Benmahammed, Oulmi, Fellahi, & Bouzerzour (2017) observed significant intergeneration correlation for days to heading, plant height, 1000-grain weight and harvest index in wheat. Similar findings were also observed by Wagoire, Ortiz, Hill, & Stolen (1999) for grain yield and Suwarto, Untung, & Siti (2015), Singh et al. (2017) for plant height in F 2 and F 3 generations. Shanatava, Veerghanti, Suresh, Mahadevu, & Hittalmani (2014) reported significant intergenerational correlation between F 2 and F 3 generations for almost all the characters except plant height. Manohara & Shashidhar (2018) reported non-significant and low intergeneration correlation for grain yield indicated that early generation selection for grain yield was not effective.