Environment Interaction in Cold Tolerant Genotypes of Rice (Oryza sativa L.) for Vigour Index

INTRODUCTION

Rice (Oryza sativa L.) is a means of subsistence as a food crop across Asia where half of the world’s poorest people reside. The crop is unique because it can be cultivated in wet conditions which may not be possible for other crops to survive. To combat the elevating food demand, breeding for high yielding varieties of rice coupled with tolerance to abiotic stresses is extremely important. Among the abiotic stresses, cold stress is a highly significant factor that disturbs the rice development and metabolism cultivated in temperate and high-elevated regions. It interferes with the development of plant morphologically and reduces the yield performance. When seedlings are prone to cold, improper seed germination, injury to seedlings, weak crop stand coupled with diminishing productivity and stability of yield is witnessed. Even rolling of leaves, yellowing, necrosis and dwarfing of seedlings were reported [1]. In north-eastern India, rice cultivation during the winter season is frequently subjected to cold stress affecting the seed growth, tillers formation, fertility and ultimately the yield [2]. In Telangana, especially during the Rabi season, the productivity of rice in northern districts is hampered by cool temperatures of <10^oC during December and January resulting in stunted growth and poor tillering of the seedlings. Thus, under such conditions, cultivation of cold tolerance genotypes will be useful in producing optimum yield. Since cold tolerant genotypes normally belong to japonica subspecies of rice, the need for the development of cold tolerance in indica lines suited to rabi and higher-latitude zones persists.

On the other hand, seedling vigour is the product of size of seedling, internal condition, and rate of development related to genetic and environmental impacts. It is regarded as a main element in seed standards that deciphers precise information on the field performance potential of seed lots. Coupled with cold stress, low vigour in seeds may aggregate reduced speed of germination with low seedling emergence uniformity, affecting the initial and final crop establishment, consequently minimizing the seed production [3]. Since the seed health and quality are highly influenced by the time of sowing, edapho-climatic factors, growing environment, the genetic architecture of species [4], the search for greater adaptable cultivars may hasten the performance of rice seeds with good physiological functioning. 

For assessment of stability, optimum performance of cultivars is very much necessary [5]. A cultivar could be considered superior with an ability to perform well in a favorable climate coupled with phenotypic stability. Selection and evaluation of plant cultivars are usually influenced by the interaction of Genotype and environment (G x E). There have been several statistical techniques employed for Genotype by Environment interaction analysis [6]. However, the most commonly followed techniques rely on the statistics of regression, parameters of variance along with non-parametric approaches. Preliminary base for considering regression model being a technique of stability assessment was put forward [7], and it was revamped [8]. Tai with his companions assessed phenotypic stability with the help of regression coefficient (b) and deviations of mean square from the regression (s²d) for some key characters as an altered model of what Eberhart and Russell (1966) procured [9]. The ecovalence as a measure of stability which includes genotype’s role in the total sum of squares of Genotype by Environment interaction [10]. Shukla gave the famous parameter of ‘stability variance’ (S²) by altering the Wricke’s ecovalence model for the purpose of providing an impartial outcome of the Genotype by Environment variance pertaining to all genotypes [11]. Francis and Kannenberg considered the coefficient of variation measure (cvi) for every cultivar as a parameter of adaptation [12], while it was recommended as a superiority index (Pi) measure, by taking in account the average of squared interval in between the cultivar’s reaction and the average of highest response accounted from all the environmental conditions [13]. However, among all these stability models, Eberhert and Russell’s statistic (1966) is extensively used since it considers both the linear as well as non-linear parts of GEI for determining stable genotypes. As per this stability statistic, cultivar having top mean performance, coefficient of regression being unity and deviation from regression not differing from zero is considered more stable. Thus, an experiment comprising of 38 rice genotypes was performed under low temperature stress at the germination stage in four different experimental conditions to determine the stability and G x E interactions using various parametric stability models.

MATERIALS AND METHODS

Plant material

A collection of 38 cold tolerant genotypes (Table 1) procured from Vivekananda Parvatiya Krishi Anusandhan Sansthan, Almora, Regional Research Station, Palampur and Sher-e-Kashmir University of Agricultural Sciences and Technology, Kashmir in addition to three checks that were used as a basic material for the study.

Cold tolerance

The cold tolerance experiment was taken up under controlled conditions at Quality control Lab, PJTSAU, Rajendranagar, Hyderabad [14]. Twenty seeds of each genotype were initially rinsed with 70% aqueous ethanol (30 sec) accompanied by soaking in 5% aqueous solution of sodium hydrochlorite (20 min) to avert any microbial risk. Then the seeds were rinsed with distilled water allowed to grow in petri plates with double layered germination paper dipped in distilled water. Genotypes were tested for cold tolerance by germinating them in a BOD seed germinator at different experimental temperatures viz., 28°C for a week (control), 13°C and 8°C for a period of 28 days (treatments) and during Rabi season. By employing the mean data of seedling length (cm) and germination (%), seedling vigour index was calculated.

Parametric stability models

Various parametric stability models were utilized for detecting stable genotypes grown at different temperature conditions as follows:

[10] conceptualized the contribution of every cultivar to the total mean squares of GEI as the ecovalence. Ecovalence or stability (Wi) is determined as the association of i^th cultivar with environment described as:

In which, Yij is overall mean performance of cultivar (i) in j^th environmental condition, Yi. With Y.j are the deviations of cultivar and environmental means and Y.. is overall mean. Because of it, cultivars having a lower Wi ecovalence are supposed to have minimum deviations from the environmental means and thus regarded as highly stable.

Coefficient of Variation stability parameter (CVi) was deciphered with combination of coefficient of variation statistic and overall mean performance of cultivar [12].

Splitting up the effect of GEI of i^th genotype into important Tai’s measures, viz., α and λ, that measure genotype’s linear type of response to the effects of environment and the amount of deviation from the linear response, using the component of GE interaction. 

Superiority Index (Pi) value was computed [13].

In which, Xij is performance of cultivar (i) in j^th environment, Mj being the high performance of cultivar in (j) environment and E is total number of environmental conditions.

The  variance of cultivar’s stability as its performance across environmental conditions following the elimation of environmental mean main effects [11].

In which, Yij is overall mean performance of (i^th) cultivar in (j^th) environment, Yj. is mean of the genotype i in all environments, Y.j is mean of all genotypes in j^th environments and Y.. is mean of all genotypes in all environments.

            A stability parameter to quantify non-linear sensitivity to the environmental variations was given by Perkins and Jinks with consideration on the GEI constituent of every cultivar as a linear function of additive environmental variation. In 1966, Eberhart and Russell described combining the environmental mean sum of squares and GE interactions and partitioning it into a linear effect between environments and for G by E interaction. For determining index of stability from regression statistic, residual mean sum of squares pooled over all environmental conditions is considered with a stable variety as the cultivar which has less deviation from regression mean squares (Sdi²). The equation depicting Eberhart and Russell’s model includes:

In which, Yij is mean value of cultivar (i^th) in environment (j^th), m is overall means of all cultivars over entire environmental conditions, bi is regression coefficient of cultivar on the environmental index, Ij is environmental index, which is the mean deviation of performance of all the varities from the total mean value and δij is deviation from regression.

RESULTS AND DISCUSSION

Mean performance of genotypes

Seed vigour is a prime characteristic feature of seed health, determining effective germination of the seed, seedling growth, seed perpetuity and hardening to adverse conditions [15]. High vigour index containing seeds can notably revamp the uniformity and speed of germination leading to an increased field emergence. This, in turn, paves the way for better performance of crop eventually producing higher yield under various situations. In this study, mean vigour index (Table 3) ranged from 5.90 (VL DHAN-221) to 81.60 (HPR-2336). Twenty-two genotypes in Rabi, 20 at 28 ^oC, 20 at 13^oC and 18 genotypes at 8^oC showed above average mean vigour index. Among the environments, 13^oC and 8^oC conditions were far below the average vigour index. The highest vigour index was recorded at 28^oCfollowed by Rabi with a mean value of 94.86 and 60.54 respectively. Nineteen genotypes (SKAU-382, JHELUM, SKAU-5, SKAU-389, CHENAB, HIMALAYA-1, HIMALAYA-741, HIMALAYA-2216, RP-2421, HPR-1068, SUKARADHAN-1, HPR-2336, VIVEK DHAN-82, VIVEK DHAN-62, VIVEK DHAN-65, VL DHAN-86, V L DHAN-206, VL DHAN-207, VL DHAN-209) have recorded higher individual mean vigour index as compared to the overall genotype mean vigour index across environments of which HPR-2336 and RP-2421 were the best ones which are expected to tolerate temperatures, humidity and produce a good yield.

Analysis of G×E Interaction

Genotype × environment interaction effects are accorded exceptional attention for recognizing the most adaptable cultivars befitted to suitable environmental conditions. A genotype is represented as stable if it performs statically across different environments. Combined Analysis of Variance (ANOVA) for Vigour index across four environmental conditions exhibited highly significant differences for mean squares (P < 0.01) in terms of genotype, environment as well as for GE interaction suggesting variable behaviour of genotypes coupled with a wider range of diversity across the tested environments (Table 2). Vigour index was comparatively more affected by environment, which explained 70.79% of overall variation in contrast with genotypes describing 20.13% and GE interaction interpreting 8.86% of the entire variation. The high sum of squares of environment pointed that environmental conditions were distinct with huge variations in mean of environments contributing majority of vigour index variation. The first and second interaction Principal Component Axis explained most of the variation, 78.18% and 19.6%, respectively, while the remaining residuals were not significant. Therefore, IPCA1 and IPCA2 together predicted the total variation among the tested genotypes. Environmental sum of squares was 3.5 times larger than the genotype sum of squares, while the magnitude of the GEI sum of squares was twice smaller than that for genotypes, suggesting smaller variations in  response of genotypes. The small mean square error of combined ANOVA signifies that almost all of the variation was accounted. For estimating the magnitude of GEI, though the standard ANOVA procedure is useful, but the relative contribution of each and every genotype to the total GE interaction cannot be known [16]. Thus, in order to dissect the contribution of individual genotypes to the total GEI, various statistical parameters have been deployed to study the adaptability and stability of genotypes.

Stability Statistics

Different parametric based stability models were utilized for identifying stable genotypes for vigour index.Wricke’s Ecovalence considers GEI mean square as the thumb-rule for the assessing the stability of genotypes in addition to the relative proportion of each cultivar to the overall interaction of genotype by environment [10]. Lower the Wi² values, higher is the stability of genotypes. The ecovalence values (Wi²) for 38 cold resistant rice genotypes over four environments were presented in Table 3, which ranged from 3.52 for the genotype RAJENDRA to 2825.79 for VL DHAN-221. Since cultivars with the least ecovalence (W_i²) values are regarded as stable, RAJENDRA was designated as highly stable, followed by K-429. Though genotypes were stable for ecovalence, they had average vigour index values than overall mean index indicating that stability for this parameter is not highly correlated with vigour index. Cultivar general superiority also known as superiority measure was given by Lin and Binns, 1988. It explains the mean square distance between the responses of genotypes to the peak response over varied environmental conditions. This parametric measure largely focuses on genotype’s performance rather than a stability measure over environments. Lesser the superiority measure, shorter is the way to the genotype, leading to the emergence of promising genotype with maximum vigour index. Genotypes are identified as adaptable and desirable if they possess a lower superiority measure value. This stability measure described the highly stable genotypes (Table 3) for Pi and vigour index being HPR-2336 followed by the genotypes RP-2421 and HPR-1068, which ranked second and third respectively. This parameter was hugely associated with vigour index, since, genotypes with high vigour index were also observed to be possessing high stability, according to this parameter.

More stable are those varieties with low environmental variance (S²_i) and coefficient of variation (CV_i) as proposed by Francis and Kannenberg in 1978. In this study, VIVEKDHAN-65, TELLAHAMSA, SKAU-341, and CHINA-1039 genotypes had lesser values for environmental variance (S²_i) with coefficient of variation (CV_i) compared to the remaining for the vigour index indicating their high adaptability (Table 3, Fig. 1). On the other hand, genotypes SUKARADHAN-1, CHINA-988 and VL DHAN-209 had higher rank for CV_i, thus were classified to be unstable coupled with their lower vigour index. In Shukla’s stability variance, with regard to each and every genotype, the Genotype by Environment sum of squares is divided into variance components (σi²). A genotype is supposed to be adaptable and stable with its stability variance value (σi²) being close to the variance of environment (σo²) on the basis of the variance components which indicated by Sh-σi² equal to zero. Stable genotypes have smaller stability variance (σi²), while higher value points the lesser stability of genotypes. This stability parameter aids in the recognition of factors of environment that provides contribution to the heterogeneity in GEI and thus is considered practically important. Results of the present study revealed that RAJENDRA, K-429 and CHINA-1007 were the highly stable genotypes concerning their rank. For this parameter, the stable genotypes had moderate vigour index indicating that high vigour index is always not related to the stability of genotypes.

Two of the genotypic stability statistics that were put forward by Tai (1971) include linear response of environmental effects as quantified by α while the divergence from linear response of the error variance magnitude is reflected by λ. The distribution of 38 rice genotypes on the Alpha-Lambda space depicting distinct regions of stability was shown in Figure 2 with the horizontal axis as lambda (λ) and the vertical axis as alpha (α). Perfect stable genotype has consistent across all environments which is equivalent to determine α = -1 and λ = 1. Genotypes with α = 0 and λ = 1 are regarded as moderately stable, α > 0 and λ = 1 as below average stable, while with α < 0 and λ = -1 are determined to be having above average stability.HPR-2336, RP-2421, HIMALAYA-2216 and VIVEKDHAN-62 were average stable cultivars according to this model. Numerous reasons are responsible for the dissimilarity between the results of Tai’s and Shukla’s parameters. Shukla’s method is considered only when there are a large number of genotypes and is dependent on interaction totals, while Tai’s model utilizes the interaction means for the purpose of analysis. By taking into account the GEI constituent of every genotype as a linear function of the additive environmental constituent, a statistical analysis to quantify the non-linear response to the varaiations of environment was given by Perkins and Jinks. In this stability model, for each genotype and environment, the fixed effect is considered as the deviation from regression line with βi= 0 and σ²i= 0 as stable. In the current study, genotypes HPR1068, HPR2336 and RP-2421 were regarded as stable. The genotypes with lesser b values were least responsive to environmental fluctuations with least environmental correlations.

Cultivars possessing higher mean performance, coefficient of regression (bi) near to one and deviation from regression (S²di) close to zero are regarded as stable [8].  The values of regression coefficient for genotypes varied from 0.134 to 1.581 (Table 3, Fig. 3). Nineteen genotypes were with bi-values greater than unity (> 1.0), indicating better buffering of these genotypes to rich environments and might be sensitive to environmental changes. The remaining 19 genotypes had bi scores smaller than unity (< 1.0), suggesting specific adaptability to low and poor performing environments as suggested by Das et al., 2010. On the other hand, some genotypes namely, VIVEKDHAN-85, SKAU-389, HIMALAYA-741, VLDHAN-209, MTU-1010 and VIVEKDHAN-62 had bi-values close to or equal with unity, indicating wide adaptability to the environmental conditions. Further, considering mean vigour index and bi value of the genotypes jointly, four genotypes (VIVEKDHAN-85, SKAU-389, SKAU-339, SHALIMAR-1 and HPR-2513) were found to be having greater buffering and adaptability to the environment. The deviation from regression (S²di) value among the genotypes ranged from 1.060 to 246.72. Of the total 38 genotypes, none of them had deviation from regression (S²di) value equal with zero showing poor adaptability of these genotypes in different environments. Overall, when the combined adaptability statistics of higher mean vigour index, coefficient of regression along with mean square deviation were taken into account, VIVEKDHAN-85, SKAU-389, SKAU-339, SHALIMAR-1 and HPR-2513 of the entire genotypes exhibited general adaptability. Ranking of genotypes based on vigour index and stability parameters was given in the Table 4.

A cultivar that delivers good performance and is consistent in environmental fluctuations is said to be adapted [17]. This capacity in a stable cultivar is a consequence of perfect combination of variable characters that help to perform coordinately in difficult situations. Thus, the current study identified the genotypes RAJENDRA, K-429 to be stable according to Shukla’s Stability Variance and Wricke’s Ecovalence, RP-2421, HPR 2336 genotypes adaptable as per Superority measure, Tai Stability Analysis and Perkins and Jinks model. Of all the genotypes, VIVEKDHAN-85, SKAU-389, SKAU-339, SHALIMAR-1 and HPR-2513 emerged as stable cultivars according to Eberhart and Russell model. The genotype was VL DHAN-221 found to be least stable from all the stability models. 

CONCLUSION

From the present investigation multiple models were fitted to assess the level of adaptability, which compiled with vigour index mean analyses, deciphered a satisfactory understanding of the stability level of rice genotypes in varied environmental conditions. Vigour index stability across different environments varied among the rice genotypes. It was highly correlated with Superiority measure and Perkins and Jinks stability models. On the whole from the study, it was evident that among the promising genotypes that were identified, RAJENDRA, K-429, RP-2421 and HPR 2336, were more stable for most of the stability statistics in union with high vigour index which could be utilized as a germplasm collection for upcoming breeding programmes. But as per Eberhart and Russell model, VIVEKDHAN-85, SKAU-389, SKAU-339, SHALIMAR-1 and HPR-2513 were found to be having wider adaptability to the environmental conditions, while VL DHAN-221 was found to be least stable with least vigour index. Thus, the current study was useful for identifying potentially well performing and stable cold tolerant rice genotypes for vigour index through genotype x environment interactions.

REFERENCES

[1] Andaya VC, Mackill DJ (2003) Mapping of QTLs associated with cold tolerance during the vegetative stage in rice. Journal of experimental botany 54(392):2579-85.

[2] Thakur R, Mishra S (1990) Cold tolerance in rice breeding in north Bihar for spring season. InPoc. of Planning Meeting of IRRI Collaborative Research Network on Cold Tolerance in Rice in East and Northeast India. IRRI, Manila, Philippines: 23.

[3] Peske ST, Villela FM, Meneghello GE (2012) Sementes: fundamento scientíficos e tecnológicos. 3^rd edition Pelotas: UFPel 573.

[4] Marcos-Filho J, Fisiologia (2015) de sementes de plantas cultivadas. 2^nd edition Londrina: ABRATES 660.

[5] Becker HC, Leon J (1988) Stability analysis in plant breeding. Plant breeding 101(1):1-23.

[6] Crossa J (1990) Statistical analyses of multilocation trials. Advances in agronomy 44:55-85.

[7] Finlay KW, Wilkinson GN (1963) The analysis of adaptation in a plant breeding program. Australian Journal of Agricultural Research 4:42-754.

[8] Eberhart ST, Russell WA (1966) Stability parameters for comparing varieties 1. Crop science 6(1):36-40.

[9] Tai GC (1971). Genotypic stability analysis and its application to potato regional trials. Crop science 11(2):184-90.

[10] Wricke G (1962) on a method of understanding the biological diversity in field research. Z. Pflanzenzucht 47(1):92-6.

[11] Shukla GK (1972) Some statistical aspects of partitioning genotype environmental components of variability. Heredity 29(2):237-45.

[12] Francis TR, Kannenberg LW (1978) Yield stability studies in short-season maize. I. A descriptive method for grouping genotypes. Canadian Journal of Plant Science 58: 1029-1034.

[13] Lin CS, Binns MR, Lefkovitch LP (1986) Stability analysis: where do we stand? 1. Crop science 26(5):894-900.

[14] Cruz RP, Milach SC (2004) Cold tolerance at the germination stage of rice: methods of evaluation and characterization of genotypes. Scientia Agricola 61:1-8.

[15] Sun Q, Wang JH, Sun BQ (2007) Advances on seed vigor physiological and genetic mechanisms. Agricultural Sciences in China 6(9):1060-1066.

[16] Ghaderi A, Everson EH, Cress CE (1980) Classification of Environments and Genotypes in Wheat 1. Crop Science 20(6):707-10.

[17] Sinha SK, Biswas S (1986) Adaptability of some rice varieties in unilocational testing programme during aus season. Oryza 23:166-72.