Genotypic performance , adaptability and stability in special types of irrigated rice using mixed models 1

There is an increasing demand for special types of rice that have black or red pericarp, low amylose (subspecie japonica) or aromatic. However, intense breeding process for obtaining and indicating genetically promising new cultivars and adapted to cultivation environments are demanded. In this sense, for evaluating the special types of irrigated rice genotypes, and determining the adaptability and stability of these using mixed models. First, a preliminary agronomic evaluation of genotypes for special types of irrigated rice was performed, and later, the agronomically promising genotypes were evaluated in multi-site trials, aiming at verifying the interactions with the environment. Statistical analyzes were performed considering mixed models, using the SELEGEN-REML/BLUP software. High genotypic variability among genotypes for special types of irrigated rice was observed, obtaining some agronomically promising and with good adaptability and stability, with high accuracy and selection efficiency using mixed models. The observed wide diversity and rice genetic variability presented new prospects and opportunities for producers for acquiring food of higher added value to consumer market.


INTRODUCTION
Rice (Oryza sativa L.) is staple food for more than half the world's population (MUTHAYYA et al., 2014).Globally, there is a wide variation in the consumption habits of this cereal, mainly due to intrinsic characteristics associated with cultural and socioeconomic factors.The Brazilian market is based on the consumption of polished white rice or parboiled of long-thin class (subspecies Indica), which have a high amylose content, low gelatinization temperature, thus showing, grains with high softness and disintegration after cooking process.However, over the past few years, in general the pattern of food consumption has been changing worldwide, mainly due to globalization and the increasing number of studies concerned to composition and nutritional constituents of functional foods.Rice with black pericarp, is a cereal rich in fiber and polyphenols, especially anthocyanins (XIA et al., 2006) and may be good for cardiovascular health (CHIANG et al., 2006;CHOI;KIM;FRIEDMAN, 2013;GINTER;SIMKO, 2012;PANDEY;RIZVI, 2009;YANG et al., 2011).Salgado et al. (2010) found that white rice diet, common in Brazilian diet, has few nutrients compared to black rice, and the phenolic compounds and proteins values in black rice are five to ten times higher than in white rice.Besides black rice, Chen et al. (2012) identified high levels of proanthocyanidins in rice bran with red pericarp, with potential for use as a functional food supplement for human consumption.
It is increasingly evident the increasing range of scientific information pointing to the nutritional complexity of genetic rice diversity (HEUBERGER et al., 2010), therefore, the future trend is that these functional foods are no longer just market niches, becoming on new prospects and opportunities.Barni et al. (2015) aimed at meeting the market acceptance for special types of rice with black and red pericarp in the state of Santa Catarina, performed sensory analysis in 1,068 people and have obtained good consumer acceptance, with high appeal of the attributes of phenolic compound presence.Special types of rice are still in breeding processes much lower than the cultivars produced nationwide, and show lower yield potential.However, they have greater added value in the consumer market, being economic possible its production.
In order to obtain genotypes potentially more productive, it aims at selecting genotypes that have favorable agronomic characteristics, thus determining a greater genotypic efficiency.In this sense, it is evident the contribution of plant breeding to achieve significant increases in yield potential in rice crop, being mainly obtained through changes in plant architecture.The modern architecture of plants is derived from intensive breeding processes that resulted in plants with high capacity to have tillers, robust, short stems and erect leaves.
In these improvement stages is extremely important that the statistical methods used provide informative genetic parameters and with high accuracy.The REML/ BLUP method, which is the estimation of the variance components by restricted maximum likelihood (REML) (RESENDE;DUARTE, 2007), is a genotypic evaluation technique that has been constantly evolving with high efficiency in information about the correct ordering of genotypes for selection purposes and also on the effectiveness of inference about its genotypic value.
In order to act more effectively on this new reality, the authors aimed at evaluating agronomically and determining the adaptability and stability of special types of irrigated rice breeding program of Embrapa using mixed models.

MATERIAL AND METHODS
The experiment consisted of two stages, where, at first a screening related to agronomic evaluation of genotypes for special types of irrigated rice was performed, and later, the promising agronomically genotypes were evaluated in tests of crop value and use (VCU), in order to verify its interactions with the environment.
The preliminary agronomic evaluation was conducted in the experimental field in the municipality of Capão do Leão, State of Rio Grande do Sul (31º48'45,51" S latitude and 52º28'21,45" W longitude), during a period of five years in the harvest of 2009/2010, 2010/2011, 2011/2012, 2012/2013 and 2013/2014.The experiment was conducted in a randomized block design with four replications and plots consisting of 6 rows of 5 m length with 0.17 m spaced between rows.The usable area of plot was consisted of 4 m at the center of four internal rows, in order to exclude any incident effect on boundary.Thus, a total of 63 genotypes for special types of rice were evaluated (Table 1) which have features of interest and seven witness cultivars, four cultivars with long thin grains (BRS Pampa, BRS 7 'Taim', BR IRGA 409 and IRGA 417), and two with grains of subspecies japonica (IAS 12-9 Formosa and BRS Bojuru) and an aromatic cultivar (EMPASC 104).
In this experiment grain yield adjusted to 13% moisture; plant height (cm) at maturation stage, by measuring the length of the soil main stem until the end of panicle; days to flowering, considering the number of days from emergence to 50% of the exposed panicles; and percentage of intact grains after the industrial refining process were evaluated.Genotypes that showed promising agronomic characteristics were allocated in multi-site trials being installed in three locations of the State of Rio Grande do Sul (Capão do Leão, Santa Vitória do Palmar and Capivari do Sul) during the harvest of 2011/2012, 2012/2013 and 2013/2014.A randomized block design was used with four replications and plots consisted of 9 rows of 5 m length with 0.17 m spaced between rows.The usable area of plot consisted of 4 m in the center of 5 internal rows.
In this experiment, the evaluated characteristics were: grain yield adjusted to 13% moisture; plant height at maturity stage, by measuring the length of the soil main stem until the end of the panicle; days to flowering, considering the number of days from emergence to 50% of the exposed panicles; percentage of intact grains after the industrial refining process; tolerance to plant lodging, leave scald (Gerlachia oryzae), brown spot (Bipolaris oryzae) and spot of grains (caused by a fungi and bacteria complex); characterization of grain concerned to length, width and thickness of caryopsis, percentage of chalkness grains, grain percentage with white belly, percentage of chalkness grain area, percentage of grain with color defects, amylose content and gelatinization temperature.Considering that the intrinsic physical characteristics of the grains were measured using the S-21statistical rice analyzer.
For both experiments the basis fertilization was 300 kg ha -1 of NPK (05-20-20 formula) and 90 kg ha -1 of nitrogen as urea, applied 50% of the dose in the V4 stage (early tillering) and the remainder at the R1 stage (panicle differentiation).The permanent flood irrigation system to R9 was used (final stage of genotypes maturation).
Statistical analyzes were performed considering mixed models through REML/BLUP using SELEGEN- REML/BLUP software (RESENDE, 2007).The statistical model used to estimate the harmonic mean of the relative performance of genetic values (MHPRVG) was obtained from the model y = Xr + Zg + Wi + ε, wherein y is the data vector, r is the block effect vector, considered as fixed, added to the general average; g is the genotypic effect vector, considered as random; i is the genotype x environment (random) interaction effect vector; ε is the error or residue (random) vector; and X, Z and W represent the incident matrices for these effects.

RESULTS AND DISCUSSION
Estimates of genetic parameters (Table 2), considering the combined analysis of agronomic evaluation of 70 genotypes of special types of irrigated rice, showed high genetic variability for evaluated characters and evidenced by the high magnitude of coefficient estimates of genotypic variation and individual heritability.Low environmental variation coefficients with high obtained accuracies demonstrate that the experiment reached high efficiency in reducing environmental effects and showed high inferences of genotypic averages.Tests that have accuracy values above 0.90 can be considered as excellent experimental precision.
The genotypic values with the average interaction among genotypes and environment (u + g) on the analysis of all five agricultural yields (Table 3) indicated high variations in genotypic patterns of special types of rice.The character of days to flowering, showed fluctuations of magnitude between 70 and 134 days, i.e., genotypes was obtained with cycles ranging from extremely early and later.The CNAi 9917 genotype had the lowest plant height with only 75.28 cm, in contrast, Moti was considered the genotype with higher height (145.03cm).
Regarding the yield predicted genotypic values (Table 3), the BRS Pampa witness cultivar of wide cultivation under irrigation system, stood out for its high yield genotypic value, with genetic potential to exceed 11000 kg ha -1 , as established by the superior limit of confidence interval.It should be noted that seven genotypes for special types of rice (Nippombare, BRS 358, SCS119 Rubi, Jasmine, BRS AG, BRS 902 and LTB 12057), showed yield genotypic magnitude higher the best witness for this purpose (EMPASC 104) as well desirable agronomic attributes.The Nippombare genotypes (subspecies japonica), BRS 358 (subspecies japonica) and SCS119-Rubi (red pericarp) had higher values even the witnesses of traditional crops BR IRGA 409 and IRGA 417.The Nipponbare genotype that shows short type of grain (subspecies japonica), is known worldwide for this feature, whose genome was completely sequenced.
Table 2 -Deviance, variance components, accuracy and coefficient of genotypic and residual variation of the characters: days to flowering (DTF), height of plants (HP), percentage of intact grain after processing (PIGAP) and grain yield, obtained through individual REML, considering combined analysis of 70 genotypes of special types of irrigated rice (Oryza sativa L.) evaluated in five harvest V g -genotypic variance; V ga -variance of genotype x years; V e -residual variance; V f -individual phenotypic variance; h 2 g -heritability of individual parcels in the wide sense, ie, the total genotypic effects; c 2 ga -coefficient of determination of the effects of genotypes x years; h 2 mg -heritability in the wide sense the level of middle genotypes; A c g -accuracy of selection at the level of middle genotypes; r ga -genotypic correlation between genotype x year; CV g -genotypic variation coefficient; CV e -residual variation coefficient   (WICKERT et al., 2014).
Other genotypes that stood out in terms of yield and agronomic standards were: Jasmine genotype that presents as highlight the aroma presence in grains; BRS AG, nicknamed "Giant" is a result of a crossing conducted at Embrapa Temperate Climate involving genes of introduced genotype SLG1 (super large grain) (TAKITA, 1983), whose grain dimensions are larger than those of conventional rice and weight of thousand grains of approximately 52 g (twice of the traditional cultivars), therefore, due to its excellent agronomic performance become the first cultivar of irrigated rice released as raw Genotypic performance, adaptability and stability in special types of irrigated rice using mixed models material for ethanol production and also for animal feed; BRS 902, released in 2015 by Embrapa by showing red pericarp; and LTB 12057 that have characteristic grains to the Japanese culinary.
Several genotypes also stood out in relation to IAS 12-9 Formosa and BRS Bojuru witnesses for special types, being promising for future releases or selection for use in hybridizations to obtain superior genotypes for interest features.
The estimate of individual accuracy, which measures the correlation between predicted genetic values and phenotypic values of each genotype, varied between high and very high (0.88 to 0.97).Thus, this analysis has been highly informative concerned to data and with good statistical reliability.
Table 4 shows the variance components and the interactions of genotypes that have shown promising agronomic characteristics to nine environments (3 sites x 3 years), through four evaluated agronomic characteristics.And for days to flowering (DTF), plant height and yield variables, the statistical model provided high genotypic variance, associated to high genotypic heritability of character and accuracies higher than 0.97 (very high).Therefore, we can infer that, for these variables, there was high genetic variability of rice genotypes for special types, associated to high efficiency on phenotypic evaluations performed at field.The identification of genotypes with high yield, production stability and wide adaptability to various environments is one of the main targets of rice breeding programs.Thus, Table 5, through the harmonic mean of relative performance of genetic values (MHPRVG), efficiently provided the simultaneous interaction conception of genetic values for yield, adaptability and stability conditioned to evaluation environments.Regitano Neto et al. (2013) andCarbonell et al. (2007) also found that MHPRVG method was advantageous for presenting results on the same measurement scale of assessed character and simultaneously for production, stability and adaptability and it can be performed efficiently in the context of mixed models.
Among the evaluated genotypes of irrigated rice, it was found that Jasmine (aromatic), SCS 119 Ruby (red pericarp), BRS 358 (subspecies japonica) and BRS AG (high amylose) were genetically superior in terms of yield to IAS 12-9 Formosa cultivar (witness for special types of rice), combining both high yield, adaptability and genotypic stability.
Table 4 -Variance components, accuracy and coefficient of variation of characters -day to flowering (DTF), height of plants (HP), percentage of intact grain after processing (PIGAP) and grain yield, considering combined analysis of 15 genotypes of special types of irrigated rice (Oryza sativa L.) evaluated in nine environments (three sites and three harvest) V g -genotypic variance; V bloc -block variance V ga -variance of genotype x environment; V e -residual variance; V f -individual phenotypic variance; h 2 g -heritability of individual parcels in the wide sense, ie, the total genotypic effects; h 2 mg -heritability in the wide sense the level of middle genotypes; A c g -accuracy of selection at the level of middle genotypes; c 2 bloc -coefficient of determination of the effects of blocks; c 2 ga -coefficient of determination of the effects of genotypes x environment; r ga -genotypic correlation between genotype x environment; CV g -genotypic variation coefficient; CV e -residual variation coefficient  The current breeding programs of irrigated rice, aims at combining genotypes with high yield that have high genotypic plasticity in order to minimize the vulnerability to main biotic and abiotic factors, thus mitigating interference in the expression of its genetic potential.
Table 6 shows the genotypes responses concerned to lodging and principal diseases inherent to crop.Genotypes that presented high yields and good agronomic attributes in Table 5 also showed good tolerance to lodging and principal crop diseases.The best performances were obtained by already commercial genotypes (BRS 358, BRS AG, IAS 12-9 Formosa, BRS 902, Jasmine, SCS 119 Rubi, SCS 120 Ônix) due to rice breeding programs for special types combine high yield with good agronomic attributes and good tolerance to stresses in its genotypes.The unique grain pattern for special type that has not been much worked in Brazil to genetic performance improvement are the Italians types (Arbório, Carnaroli), which showed low genotypic performance for both grain production and agronomic attributes.
The attributes related to grain provide important parameters for the final product knowledge after cooking process.The most polished rice components consists of starch (until 95% of dry weight), protein (5% to 7%) and lipids (0.5% to 1%).The amount of these compounds affects sensory (FITZGERALD et al., 2008) and food appearance properties.
Physical properties include grain yield after processing, uniformity, whiteness and grain translucency.The defects associated to rice combined grain quality characteristics dictate its market value and have a key role in the adoption of new cultivars.Thus, in Table 6 it can be seen that there is a great variation in grain dimensions, which results in the different grain forms, as well the main physical properties.Regarding the caryopsis dimensions, it was found that eight of the fifteen genotypes have more rounder shape, i.e. it showed the relationship between length and width, less than 2.0 (<1.50 = rounded; 1.50 -2.00 = half-rounded), characteristic of Japonica subspecies.Moreover, these genotypes by genetic constitution had high chalkness area index.
In addition to grain physical characteristics, amylose content and gelatinization temperature are critical to determine sensory and cooking characteristics of rice.The amylose content is considered one of the main parameters for the technological quality and rice consumption and due to this reason, this characteristic should be assessed during cultivars development (WALTER;MARCHEZAN;AVILA, 2008).Table 6 shows that evaluated genotypes showed amylose content Genotypic performance, adaptability and stability in special types of irrigated rice using mixed models Genotype PL (1) (1-9) Esc (1) (1-9) BS (1) (1-9) GS (1) (1-9) CL (mm) CW (mm) CT (mm) LWR ( 2 (AC) between low to intermediate, while the genotypes of japonica subspecies showed low amylose, i.e., soft and "sticky" grains after cooking.In contrast, genotypes of Indica subspecies both aromatic and color pericarp types, showed intermediate amylose levels, with loose and soft grains after cooking process, being highly compatible with market requirements.
In addition to amylose content, gelatinization temperature (GT) is a very important quality parameter for rice, due to its evaluation of cooking resistance index.In Table 6 it was found that nine genotypes showed GT between 63 °C and 68 °C (low) and six genotypes between 69 ºC and 73 °C (intermediate).
Even demonstrating yield and agronomic characteristics inferior compared to BRS Pampa cultivar, which is one of the main traditional cultivars of irrigated crop in southern of Brazil, the special types presented in great prominence, due to its added value, either for human consumption as well for other uses such as animal feed, ethanol production, among others.This inferior yield of special types of rice compared to traditional crops corroborates several previous studies worldwide.Ji et al. (2012) found that rice genotypes with black pericarp showed yield, weight of 1,000 grains and thickness of caryopsis significantly lower when compared with white pericarp genotypes in two crops.Furthermore, differences in physiological characteristics of plants were observed.Thus, the special types have wide source for new prospecting studies in all stages of the production chain, using the great diversity and genetic variability of rice in order to expand opportunities for farmers in the cultivation of foods with added value and with higher quality to consumer market.

CONCLUSIONS
1.A high genetic variability among genotypes for special types of rice for the main agronomic characteristics using mixed models was found; 2. The commercial genotypes BRS 358, BRS AG, IAS 12-9 Formosa, BRS 902, Jasmine, SCS 119 Rubi, SCS 120 Ônix were agronomically superior through simultaneous selection for yield, stability, adaptability and agronomic attributes for special types of irrigated rice; 3. The use of mixed models and MHPRVG method proved to be suitable, efficient and highly informative about the inference of genotypic values for the use in rice breeding programs aiming at assessing special types of rice.

Table 1 -
Genotypes used in preliminary agronomic evaluation for special types of irrigated rice (Oryza sativa L.), considering grain class (GC), color of pericarp (CP) and special characteristics (SC)

Table 3 -
Predicted genotypic values (u + g) of agronomic attributes of days to flowering (DTF), height of plants (HP), percentage of intact grains after processing (PIGAP), accuracy, phenotypic mean and inferior (LIIC) and superior limits (LSIC) of the confidence interval of grain yield evaluated in special types of irrigated rice (Oryza sativa L.) in five harvest

Table 5 -
Genotypic values (u+g+gem) of days to flowering (DTF), plant height (HP), percentage of intact grains (PIGAP), production adaptability and stability parameters of the genotypic values (MHPRVG) and production adaptability and stability parameters multiplied by the overall average of all trials (MHPRVG*MG) of special types of irrigated rice (Oryza sativa L.)