elocation-id: e3688
Yellow rust is the biotic factor that most affects the yield of rainfed wheat in the High Valleys of Mexico, and currently, the most effective management strategy for this disease is the use of resistant varieties, which must also be stable in grain yield when exposed to this rust. In order to understand the correlation between yield stability and the severity of yellow rust, between 2017 and 2020, trials were established in 37 localities in the High Valleys of Mexico, where 14 wheat genotypes were evaluated. The correlation between yield and the incidence of yellow rust was analyzed and different stability parameters were determined. A moderate negative correlation was found between yield and disease incidence. The advanced lines Kone ‘s’, Nely ‘s’, Terre ‘s’, and the varieties Texcoco F2016 and Valles F2015 were the most stable in yield and the most resistant to rust, whereas the least stable and the most susceptible to the disease were Náhuatl F2000, Triunfo F2004, and Nana F2007, so they should no longer be recommended for sowing.
Puccinia striiformis, Triticum aestivum, correlation, High Valleys.
Globally, wheat production has been affected by high input costs, extreme droughts, and diseases, which has put the population’s food security at risk for the acquisition of this grain (The World Bank, 2024). In Mexico, yellow rust (Puccinia striiformis W) is the disease that most affects the crop in the High Valleys of central Mexico, this region being one of the most important in terms of production of this cereal. Due to the presence of new yellow rust races in 2014 that caused yield losses of 84% (Díaz et al., 2018), wheat production was reduced to 65 551 t reported in 2023 (SIAP, 2024).
To resume national production, it is necessary to have varieties that are resistant to yellow rust and also have good grain yield and that this trait is not affected when planted in varied climatic conditions.
The development of varieties that are stable in yield and resistant to yellow rust is one of the purposes of the INIFAP Wheat Program since this ensures the production of this grain in different environments, even when the incidence of rust is high (Villaseñor, 2015). In Mexico, there is no information on the correspondence between the stability of wheat genotypes yield and yellow rust.
Therefore, this study aimed to understand the correlation between the parameters of grain yield stability and the percentage of yellow rust severity measured in different rainfed environments. This stability analysis can be useful as a tool for the selection of outstanding genotypes.
Fourteen rainfed wheat genotypes (11 varieties and three advanced lines) from the INIFAP Wheat Program were evaluated (Table 1). Genotypes were established in trials in 37 localities in the states of Mexico, Tlaxcala, Oaxaca, and Puebla, during the spring-summer (SS) cycles from 2017 to 2020 (Table 2). These localities are classified as favorable and intermediate environments for the crop due to their average annual rainfall of between 400 and 600 mm (Villaseñor and Espitia, 2000).
| Genotypes† | |
|---|---|
| Temporalera M87 | Nana F2007 |
| Romoga F96 | Don Carlos M2015 |
| Náhuatl F2000 | Valles F2015 |
| Tlaxcala F2000 | Texcoco F2016 |
| Rebeca F2000 | Kone ‘s’ |
| Triunfo F2004 | Nely ‘s’ |
| Altiplano F2007 | Terre ‘s’ |
The trials were established under a randomized complete block design (RCBD) with two replications. The experimental plot consisted of four rows 0.3 m apart, with a seedbed of 1.5 m × 3 m long, with the useful plot being the total of the experimental plot (4.5 m2).
The considered variables were grain yield in kg ha-1 (YIE) and the percentage of yellow rust (Yr), which was recorded with the modified visual Cobb scale (Roelfs et al., 1992), taking the first reading at the flag leaf stage and subsequently taking readings every 10 to 15 days until physiological maturity, reporting the final reading for the present study.
An analysis of variance combined with their interactions and a Tukey comparison of means (α= 0.05) were performed. An analysis of correlation between the variables, a stability analysis for grain yield using the standard deviation (Si) and the coefficient of variation (CVi) proposed by Francis and Kannenberg (1978), Eberhart and Russell’s (1966) stability indices, and Lin and Binns’s (1988) superiority indices, were also performed using the SAS package version 9.2 for Windows.
Table 3 shows the analysis of variance of the 37 localities, where there were highly significant differences among localities, genotypes, and in the locality × genotype (G×E) interaction for the variables evaluated. As mentioned by Hortelano et al. (2013), this analysis confirmed the contrast that exists between rainfed wheat production environments, where genotypes behave differently when changing environments.
Eberhart and Russell (1966) results mention that G×E interactions are important for the development of varieties; however, when they are planted in different environments, their categorization differs, which causes difficulty in observing their superiority. To solve this problem, it is necessary to detect stable genotypes that interact less with the environments where they are planted and that are classified as such by stability parameters such as those determined below.
Table 4 presents the comparison of means for the different localities, where it is observed that the average yield varied from 2 to 7.4 t ha-1 and the percentage of yellow rust ranged from 6.8 to 43.8%.
With these results, three response groups were formed for localities: 1) with the highest yields, x̅= 5 932 kg ha-1, eg., Juchi1F20, Tlal19, and Luc1F19 with an average percentage of yellow rust of 18%; 2) with intermediate yields, x̅= 3 658 kg ha-1, Juch17, Nan1F20, and Terre18 with an average percentage of yellow rust of 19%; and 3) with the lowest yields, x̅= 2 543 kg ha-1, Mira17, Luc1F20, and Cha1F17 with an average percentage of yellow rust of 23%. This suggests that higher percentages of rust lead to lower yields and vice versa, a trend observed by Ramírez et al. (2016) when evaluating rainfed wheat and by González and Rodríguez (2023) in barley attacked by leaf rust.
The Table 5 shows the grain yield and the percentage of yellow rust observed in the genotypes, where Kone ‘s’, Nely ‘s’, and Terre ‘s’ surpassed Nana F2007, Triunfo F2004, and Náhuatl F2000 by up to 1.4 tons, the latter varieties are considered outstanding varieties for rainfed environments, but they are currently already susceptible to yellow rust (Díaz et al., 2018).
These higher yields are due to the fact that these three advanced lines presented the lowest percentages of yellow rust (<3%) and, as mentioned by Ramírez et al. (2016); Villaseñor et al. (2021), as part of the advances of the Wheat Program, the obtaining of genotypes with higher yields and greater resistance to rusts should be gradual and ascending over the years. For this reason, Kone ‘s’, Nely ‘s’ and Terre ‘s’, as candidates to be released as varieties, showed these results.
Genotypes considered to be resistant, with percentages of yellow rust <5% (Roelfs et al.,1992) (Table 5), were Valles F2015, Texcoco F2016, Kone ‘s’, Nely ‘s’, and Terre ‘s’, which had the highest yields and were considered the most stable in terms of this trait and, as mentioned by Ramírez et al. (2016), they would be the genotypes that have a wide range of adaptation.
These same authors mention that the varieties released in the 2000s are the ones with the lowest yield and most susceptible to yellow rust, coinciding with what was observed in this study since Náhuatl F2000, Triunfo F2004, and Nana F2007 showed a greater severity of the disease and lower yields (Table 5).
For the correlation analysis between YIE and Yr, a moderate negative correlation was obtained (-0.41821, <0.0001) (Schober and Boer, 2018); that is, the lower the percentage of yellow rust, the higher the grain yield and according to Table 4, in most of the localities where the highest yields were obtained, the percentages of yellow rust were lower, and this same trend was observed among varieties (Table 5). Similar results were reported by Sánchez-Martín et al. (2013) in oats and crown rust and by Solomon et al. (2023) in bread wheat and stem rust.
On the other hand, considering that small values of the coefficient of variation (CVi) and standard deviation (Si) indicate greater stability (Francis and Kannenberg, 1978) and that the desirable genotypes would be those with higher yield and lower CVi (De Vita and Maggio, 2006; Hortelano et al., 2013), then, Valles F2015 followed by Terre ‘s’, Nely ‘s’, and Kone ‘s’ were the most stable in all the environments evaluated (Table 6); these three lines were also the most stable in the environments with the highest incidence of yellow rust (Table 7); likewise, these genotypes presented a lower percentage of yellow rust in the four years of evaluation (Table 5).
The stability parameter βi resulted in regression coefficients less than one (the lowest) in Nana F2007 and Romoga F96 for all environments (Table 6) and in Rebeca F2000 and Romoga F96 for environments with the highest incidence of yellow rust (Table 7), indicating that these varieties present relative adaptation to unfavorable environments (Hortelano et al., 2013).
For all environments in Table 6, it was observed that seven genotypes had βi values between 0.96 and 1.05 and that five genotypes did so for the environments with the highest incidence of rust (Table 7); this range mentioned by Hortelano et al. (2013) was applied to those genotypes with stable yields across environments and years. Rodríguez-González et al. (2014); Aula et al. (2023) mention that genotypes with βi and S2di greater than 1 and equal to 0, respectively, have a good response to favorable environments and are stable. In this sense, Rebeca F2000, Triunfo F2004, Texcoco F2016, Kone ‘s’, and Nely ‘s’ presented a βi greater than one (the highest) in all environments (Table 6) and Don Carlos M2015, Valles F2015, and Texcoco F2016 did so in the environments with the highest incidence of rust (Table 7), indicating that all these genotypes respond positively to environmental production improvements (Hortelano et al., 2013).
The variances of the regression deviations (S2di) in both analyses (Tables 6 and 7) were different from zero and, as mentioned by Hortelano et al. (2013), the linear model used is not appropriate to describe the response of genotypes as a function of the environment.
According to Lin and Binns (1988) superiority parameters, the genotypes with the lowest Pi values are considered the most stable and present a behavior parallel to the maximum response across environments, being the genotypes with the highest yield and stability in terms of this trait (Rodríguez et al., 2002; El-Hashash and Agwa, 2018). For this study, for all environments (Table 6) and environments with the highest incidence of rust (Table 7) were the Kone ‘s’, Nely ‘s’, and Terre ‘s’ lines followed by the Texcoco F2016 and Valles F2015 varieties, a categorization that coincided with that observed with the coefficient of variation (CVi) in both analyses.
These genotypes were the ones that had the highest yields and presented the lowest percentages of yellow rust (Table 5) and as they obtained the lowest values of the MS(G×E) parameter, they have good adaptation (stability) to all the environments studied (Rodríguez et al., 2002; Sabaghnia et al., 2013).
On the other hand, for all environments, Nana F2007 was the variety that exceeded the value of the critical cut-off point for the mean square of the G×E interaction, so it has specific adaptation only in certain environments (Rodríguez et al., 2002), with the lowest yield and the highest percentage of yellow rust (Table 5). For the environments with the highest incidence of rust (Table 7), Nana F2007 together with Temporalera M87, Romoga F96, Tlaxcala F2000, Rebeca F2000, and Triunfo F2004 exceeded the value of the critical cut-off point for G×E, so since they have specific adaptation to certain environments, their yield could be drastically decreased if they are planted in localities with a high incidence of yellow rust.
The advanced lines Kone ‘s’, Nely ‘s’, Terre ‘s’, and the varieties Texcoco F2016 and Valles F2015 were the most stable genotypes in grain yield, even in the localities where there was a higher incidence of yellow rust. These genotypes had the lowest percentages of rust across environments, so there was a moderate negative correlation between yield and disease incidence.
Due to their stability and resistance to yellow rust, Texcoco F2016 and Valles F2015 can be recommended for planting in the High Valleys of Mexico. On the other hand, Náhuatl F2000, Triunfo F2004, and Nana F2007, which presented lower stability, low yields, and high susceptibility to yellow rust, are no longer recommended for planting.
The determination of yield-stable genotypes with resistance to yellow rust can be a useful tool for the selection of outstanding genotypes. The higher yield of the three advanced lines is an indication that INIFAP Wheat Program continues to make progress in the generation of genotypes with higher yields and resistant to yellow rust.
SIAP. 2024. Servicio de Información Agroalimentaria y Pesquera. Anuario Estadístico de la Producción Agrícola, Producción mensual agrícola, Avance de siembras y cosechas. https://nube.siap.gob.mx/avance-agricola/.
Solomon, T.; Sime, B.; Dabi, A.; Alemu, G.; Geleta, N.; Negash, T.; Delesa, A.; Zegeye, H.; Duga, R.; Yirga, F.; Asefa, B.; Ashake, D.; Getamesay, A.; Zewdu, D. ; Kasa, D. and Degefa, N. 2023. Association between dominant wheat stem rust races found in Ethiopia and grain yield in advance bread wheat genotypes. International Journal of Bio-resource and Stress Management. 14(9):1305-1312.
The World Bank. 2024. Food Security Update. https://thedocs.worldbank.org/en/doc/40ebbf38f5a6b68bfc11e5273e1405d40090012022/related/Food-Security-Update-CV-April-25-2024.pdf.