Introgression Progress for Phenotypic Traits and Parent-progeny Diversity at Advanced Segregation Population From Oryza Barthii and Oryza Glaberrima/oryza Sativa Crosses

Rice is a cereal staple of global fame and importance. Oryza barthii, a wild species holds signicant traits and its utilization in rice breeding is rare. This study traced introgression trend of heritable traits in the offspring of O. barthii with an Africa-Asian progenitor to F 8 and assessed diversity between the parents and the F 8 population. Signicant (P<0.05) genotypic variation existed for all the traits except tiller number, panicle/meter squared, grains/panicle and 1000 grain weight. Grains/panicle and days to 50% owering had respective least (3.34%) and highest (96.32%) broad sense heritabilities. All traits had lower GCV compared to PCV. The least (5.28% and 8.05%) and the highest (90.8% and 98.1%) GCV and PCV were respectively from grains/panicle and tiller number. Clear variations on the panicles and grains include: variations in sizes, shapes, colours, presence or absence of awns. The total variance explained by ve principal component axes was 80.1%. Plant height at maturity was the only trait with signicant (p ≤ 0.01) correlation and regression between F 6 and F 7. Progenies resemblance to Parent 1(IRGC 104084) retrogressively declined but parent-offspring to parent 2 (TGS 25) progressively increased from F 6 to F 8 . Three visible groups of rice type in this study were: the O. barthii (11%), O. sativa (67%) and the intermediate group (22%). This research has added to rice genetic resources; an investigation of the nutritional status of the progenies would be an interesting research.


Introduction
Rice is a global staple which is cultivated on every continent except the Antarctica. It is adaptable to numerous climates, soil, altitudes, terrains etc. It is cultivated in more than 115 countries and feeds over 50% of the population in the world (Liu et al., 2015). Rice is a prince among cereals, producing the highest quantum of calories per unit of land (Gujja and Thiyagarajan, 2009) and it is the backbone of India's economy, providing direct employment to about 70 per cent working people (Vanniarajan and Ramalingam, 2011). It is digni ed as relished culinary in different cultures and languages. It comes rst among the important commercial food crops of the world (IRRI, 2005). Four hundred and eighty one million tonnes was the projected rice production for 2020 (Mohanty, 2009); attainment of this will be subject to the pandemic in uence of COVID 19.
The cultivated rice, genus Oryza whose chromosome number is 12 (Matsuo et al., 1997) has over 20 species. Only two species, Africa rice -Oryza glaberrima (Steud.) and Asian rice -Oryza sativa (Linn.) are the cultivated species. O. sativa originated from South-East Asia, particularly India and Indo-china, where its richest diversity exists (Li, 1970;Sampath, 1973). The species is well distributed throughout the tropics and parts of the temperate regions of the world (Oka, 1988). The primary and secondary centers of diversity for O. glaberrima is the swampy basin of the upper river Niger and the southwest near the Guinean coast (Maclean et al., 2002). The cultivation of O. glaberrima is con ned to West Africa.
Crop wild relatives (CWR) harbours extremely valuable resources for crop breeding which through introgression can lead to considerable proportion of alleles sharing among rice cultivars (Jin et al., 2018). O. barthii is the progenitor of O. glaberrima (Linares, 2002;Sarla and Swamy 2005). It has long been recognized as a wild species of rice (Li et al., 2011) whose features (long ag leaf, presence of awns, long panicles, diverse grain sizes and weight) according AfricaRice (2012). A report (AfricaRice, 2012) hinted that the ag leaf shields the panicles from the sight of ying birds while the long awns could prevent insects from accessing the grains. Other very useful traits for which O. barthii is notable include: tolerance to drought, highly vigorous, high weed competitiveness, early maturing and production of many tillers (National Research Council, 1996).
Crosses followed by sel ng leads to the generation of segregating populations which allows gene expression for particular traits (Govintharaj et al., 2017). Parent offspring correlation and regression between two generations according to Vanniarajan and Ramalingam (2011) are usually undertaken to estimate the genetic proportion of gene transferred from one generation to other; it is noteworthy that parent-progeny correlation and regression are lesser in uenced by the environment and it is a very useful method for selection in segregating population (Govintharaj et al., 2017). While available rice genetic resources needs to be sustainably conserved, continuous generation of variation remains a strong course of pursuit in plant breeding to enhance increased productivity and alleviate poverty and hunger. In the present study, our choice of the male and the female parent following Lin et al. (2020) was based on the identi cation of genetic variation between them.
The unique adaptive features in O. barthii may have enhanced its survival well over 3,500 years (AfricaRice, 2012). The same wild species holds signi cant features and wide diversity, yet it has been greatly underutilized in rice breeding programs. The present investigation seek to identify the possible introgression of heritable traits in the offspring of the wild species (O. barthii) with and Africa-Asian rice and to access signi cant phenotypic diversity between the parents and the F 8 population. Moreover, the study seek to evaluate the level of diversity in the 8 th segregating populations derived from the cross.

Materials And Methods
Crosses were made between IRGC 104084 (Oryza barthii) and TGS 25 [(Oryza glaberrima x Oryza sativa) x Oryza sativa] to generate F1 hybrid. Through a three year sel ng program involving seven cycles, 27 progenies were generated (See procedure in Figure 1). The 27 progenies and the two parents (See list in Table 1) were evaluated on the eld. The experiment was laid out in an Augmented Randomized Complete Block Design at the Africa Rice regional station, International Institute of Tropical Agriculture (IITA), Ibadan (Latitude 7 0 30'N and Longitude 3 0 45'E), Nigeria. Each of the test entries including the two parents were evaluated in single plots of 5 rows of 5 meter. The seeds were hand-dibbled at even depth and uniform spacing of 20 x 20cm apart.  fertilizer was applied at the rate of 200 Kg/ha as basal application immediately after planting. Subsequently, 100 Kg/ha urea (46% N) was applied in two equal split doses at tillering and panicle initiation stages. Weeding was carried out as at when due. The 27 progenies and the two parents were evaluated using Standard Evaluation system for Rice (SES, 2002) on the following traits: days to 50% owering, days to 85% maturity, plant height at maturity, panicle length, fertility percentage, lodging score, shattering score, phenotypic acceptability, panicle exsertion, a thousand (1000) grain weight, number of panicles per plant, number of spikelets per panicle and grain yield per plot from which yield per hectare was estimated.
Analysis of variance (ANOVA) was carried out on the quantitative and transformed scored data following the procedure of Scott and Milliken (1993 Where: Y ij is the treatment, μ is the mean, b j denotes the block effect, ci is the check effect and X i (C i ) denotes the entry effect.
Gower genetic distance was carried out using the 29 x 15 matrix mean values of genotypes and phenotypic traits. The obtained paired similarity distance was employed and subjected to principal component and clustering analysis. The parent progeny correlation and regression analysis between F6 with F7 and F7 with F8 was carried out following the procedure of Govintharaj et al. (2017) and Aananthi (2018). To further identify the introgressive trend for each genotype for the combined 15 phenotypic traits, similarity of each genotype to the two parents was performed for F 6 , F 7 and F 8 data using Gower genetic distance in SAS (version 9.4 (SAS, 2011). Moreover, within the Statistical Tool for Agricultural Research (STAR, 2014) software, genotype by trait interaction was investigated and presented as a biplot graph. Table 2 presents descriptive and variance statistics and genetic estimates of 15 phenotypic data used in the evaluation of the two parents and 27 F 8 progenies of IRGC 104084 x TGS 25 crosses. For the 29 genotypes, the means with the standard error and range of performances for the 15 traits for the 29 genotypes was presented in Table 2. Signi cant (P<0.05) genotypic variation existed for all the traits except tiller number, panicle/meter squared, grains/panicle and 1000 grain weight, moreover, tiller number had the least (0.91) R 2 and grains/panicle had the least (3.34%) broad sense heritability (Table 2). Generally, the GCV were lower than the PCV, the least (5.28% and 8.05%) GCV and PCV were from grains/panicle while tiller number had the highest (90.8% and 98.1%) for the two estimates ( Table 2). The morphology of the ripe panicle of the two parents are presented in Plate 1. The panicle of O. barthii had awn and the panicle colour was black. The colour of the panicle of the O. sativa was straw (Plate 1). Phenotypic variability was observed on the panicles and the grains of the two parents and some of the progenies in Plate 2.

Results
Clear variations on the panicles and grains include: variations in sizes, shapes, colours, presence or absence of awns etc.; moreover, some of the progenies combined the features in the two parents in various proportion (Plate 2).
The ve principal component axes in Table 3 had approximately 1.0 eigenvalues and above. The highest eigenvalues and correspondence variance proportion to the total variance was in PC1 and of the eigenvalues and contributions to total variance consistently decrease from PC1 to PC5 (Table 3). The total variance explained by the ve PC axes was 80.1% (Table 3). Tiller number, panicle/meter square, panicle/plant, fertility percentage, panicle exsertion, lodging scoring, days to 50% owering, days to 85% maturity and yield were prominent in their contribution to the variance proportion in PC1; panicle length, grain/panicle, shattering score, phenotypic acceptability and 1000 grain weight were prominent in PC2 while, plant height at maturity had the highest eigenvector loading in PC3 ( Table 3). The eigenvector loadings of the mentioned variables (for each PC) were higher than 0.2 (Table 3). Six clusters were visible at 0.05 similarity coe cient and four clusters at 0.10 points of in ection ( Figure 2). P1 stood alone in cluster I, cluster III (with only two genotypes) was closest to it with Gower genetic similarity of 0.95 (Table not shown Mean performances of the different groups of genotypes in the various clusters is presented in Table 4. P1 (IRGC 104084) which solely occupied cluster I had the least value for plant height at maturity, panicle length, grain/panicle and yield. However, the same genotype highest value for tiller number, shattering score, phenotypic acceptability, panicle exsertion and 1000 grain weight ( Table 4). The twenty genotypes in cluster II had the least 1000 grain weight but the second best nal grain yield. G3 and G21 which were the only two members of cluster III had the highest mean for: plant height, PAM, panicle length, panicles/plant, fertility percentage, grains/panicle, lodging score and the highest yield, the genotypes in the cluster had the lowest value for: shattering score, PA, PE, days to 50% owering and days to 85% maturity (Table 4). Cluster IV was distinguished for the lowest tiller number, PAM, panicle length, fertility % and zero lodging but owered and matured latest ( Table 4).
The total variance which captured the display of genotype by trait interaction in Figure 3 by the rst two PC axes was 52.3%. The interactions featured in the four quadrants. Parent 1 and 2 were separately located at quadrants one and three respectively (Figure 3). Panicle exsertion, phenotypic acceptability and shattering scores were the prominent traits in the rst quadrant, P1, G1, G12 and G14 were the genotypes with corresponding highest values for them ( Figure 3). In quadrant two, G3, G9, G20, G21, G22, G25, G26 and G27 had signi cant higher performances for: 1000 grain weight, logging score, tiller number, panicle length and panicle/metre square (Figure 3). Prominent traits associating with the nine genotypes in quadrant three were: fertility %, height at maturity, panicle length, grains/panicle and yield. Days to 50% owering and days to 85% maturity were signi cantly correlated in quadrant four and genotypes with signi cant association with them include: G11, G12, G13, G15, G16, G17 and G23 (Figure 3).
Quantitative similarities/resemblance of the 27 progenies to the two parents was through Gower genetic distance was presented in Table 5. Generally, similarity of the 27 progenies to P1 declined linearly from F6 to F8 while the similarity of the same 27 progenies to the P2 rose from F6 to F8 in a positive linear trend (Table 5). Individual similarity of the 27 progenies to the two parents differed and four notable trend responses were identi ed which include: positive linear, negative linear, positive quadratic and negative quadratic. With P1, 40.8% of the progenies exhibited negative linearity, 37% exhibited positive quadratic and 22.2% exhibited negative quadratic trend response from F6 to F8 (Table  5). Furthermore in Table 5 with P2, the respective percentage response of the similarity of the 27 progenies were: 14.8% (positive linear), 14.8% (negative linear), 40.8% (positive quadratic) and 29.6% (negative quadratic). Among the eight phenotypic traits measured for the three generations (F 6 , F 7 and F 8 ) in Table 6, only plant height at maturity had signi cant (p ≤ 0.01) correlation and regression between F 6 and F 7 .

Discussion
Continuous sel ng of the earlier generation of progeny to advanced generations is aimed at obtaining higher homozygotic status in the progenies. The main objective of single seed descent method is to rapidly advance the generation of crosses and at the end a random sample of homozygous genotype is obtained (Agriinfo, 2011; Kanbar et al., 2011; Janwan et al., 2013). Oryza barthii derivatives are useful sources of positive alleles especially for one thousand grain weight, high number of grains per panicle, high tillering ability, early owering and high milling yield (Maricel, 2010).
The gross similarity of the 27 progenies to the two parents at F 6 , F 7 and F 8 were at opposite linear trend: progenies resemblance to P1 declined but increased to P2 from F 6 to F 8 . Therefore with P1 (IRGC 104084) cytoplasmic Genotypes with lower score for phenotypic acceptability and panicle exsertion tend to be tall plants. Hence these two traits could be used to guide selection for shorter plants with less likelihood for lodging. Number of tillers per plant, panicle length and number of grains per panicle were positively correlated indicating that these traits are good selection indices for grain yield. These results were in agreement with the ndings of Prasad et al (2001) and Sürek 1988. High fertility was associated with high number of grains per panicle, high grain yield and high one thousand grain weight. This is also in agreement with the ndings of Prasad et al (2001). The panicle exsertion having a negative correlation with number of grains per panicle and grain yield shows that genotypes with lower scores for panicle exsertion (well exserted panicles) also yield better. Lower phenotypic acceptability score, lower panicle exsertion score and lower lodging percentage also indicate higher yield. Rice genotypes with a higher number of tillers per plant, high panicles per meter square and panicle length were observed to have higher yields. This is in agreement with the ndings of Prasad et al (2001). These traits could therefore be used as selection indices for higher yield. Higher fertility, higher number of grains per panicle, lower phenotypic acceptability score and lower panicle exsertion score also positively in uenced number of grains per panicle, grain yield and one thousand grain weight and could also be used to select for higher yielding genotypes. Heritability is a very important genetic estimate (Yadav et al., 2007 andPrajapati et al., 2011) with immense utility in trait-based genotype selection. High heritability estimate was observed in plant height at maturity, fertility percent, panicle length, shattering score, phenotypic acceptability, panicle exsertion, days to 50% owering and days to 85% maturity; this corroborates the ndings of Ogunbayo et al. (2014). High heritability indicates that the traits are more in uenced by genetic contribution. (88%) of our measured traits may be conforming to the above since our single plant selection program commenced at F 6 . However, there are some traits whose effective selection would be most appropriate at the advanced generations (Kahani and Hittalmani, 2016;Aananthi, 2018). Our study identify plant height at maturity as one of such phenotypic traits whose most effective selection can be achieved at F 6 -F 7 intergeneration and indicating F 6 as a good indicator  , Fert-Fertility %, Grnpan-Grain per panicle, Log-lodging Score, Flw-Days to 50% Flowering, Mat-Days to 85% maturity, GRNWT-1000 grain weight, Shatt-Shattering score, PA-Phenotypic Acceptability, PE-Panicle exsertion, YLD-Yield g/m2 † *, **, *** -signi cance at p ≤ 0.05, 0.01 and 0.001     Figure 1 The schematic procedure of the generation of F8 progenies from IRGC 104084/ TGS 25 within three years †P1 -IRGC 104084 (female) and P2 -TGS 25 (male) Figure 3 Twenty-nine genotypes by fteen traits interaction display within the rst two principal components † Plhtmat-Plant height at maturity, Tilno-Tiller number, PAM-Panicle per meter square, Panpl-Panicle per plant, Panlt-Panicle length, Fert-Fertility %, Grnpan-Grain per panicle, Log-lodging Score, Flw-Days to 50% Flowering, Mat-Days to 85% maturity, GRNWT-1000 grain weight, Shatt-Shattering score, PA-Phenotypic Acceptability, PE-Panicle exsertion, YLD-Yield g/m2