Effective selection criteria for screening drought tolerant landraces of bread wheat (Triticum aestivum L.)

Transcription

1 Available online at cholars esearch Library Annals of Biological esearch, 0, ():0- ( IN 0- CODEN (UA): ABNBW Effective selection criteria for screening drought tolerant landraces of bread wheat (Triticum aestivum L.) Ezatollah Farshadfar *, Zahra Moradi, arvin Elyasi, Bita Jamshidi, oghaye Chaghakabodi College of Agriculture, azi University, Kermanshah, Iran Department of lant Breeding, Islamic Azad University, Kermanshah Branch, Kermanshah, Iran _ ABTACT In order to evaluate the ability of several selection criteria to identify drought tolerant landraces of bread wheat 0 landraces were evaluated in a randomized complete block design with three replications under irrigated and rainfed conditions. Fourteen drought tolerance indices including stress tolerance index (TI), geometric mean productivity (GM), mean productivity index (M), stress susceptibility index (I), tolerance index (TOL), yield index (YI), yield stability index (YI), drought response index (DI), drought resistance index (DI), modified stress tolerance index (MTI), relative drought index (DI), abiotic tolerance index (ATI), stress susceptibility percentage index (I) and stress non-stress production index (NI) were calculated and adjusted based on grain yield under drought (Ys) and irrigated conditions (Yp). ignificant positive correlation was found between grain yield in the stress condition (Ys) with criteria TI, GM, M, YI, YI, DI, K TI, K TI, DI, NI, I and DI indicating that these indices discriminate drought tolerant genotypes in the same manner. rincipal component analysis (CA), exhibited that first and second CA accounted for.0% of the variation. creening drought tolerant genotypes using mean rank, standard deviation of ranks and biplot analysis, discriminated genotypes (), () and () as the most drought tolerant. Therefore they are recommended to be used as parents for genetic analysis, gene mapping and improvement of drought tolerance in common wheat. Key words: Bread wheat, drought tolerance indices, screening methods. _ INTODUCTION Global warming and concomitant increase in drought effected areas limit plant production in the world. Wheat production is also restricted by drought exposed areas and this loss led to considerable economic and social problems because of its great importance on human nutrition []. electing wheat genotypes based on their yield performance under drought conditions is a common approach, therefore some drought stress indices or selection criteria have been suggested by different researches [, ]. The impact of water shortage (availability at farm gate) and lower rainfall during the sowing period seems to be the main reason for lesser acreage under wheat crop and reduction in wheat production. Therefore, breeding for drought tolerant wheat is an important task and objective in the present scenario []. Breeding for drought resistance is complicated by the lack of fast, reproducible screening techniques and the inability to routinely create defined and repeatable water stress conditions when a large amount of genotypes can be evaluated efficiently []. Achieving a genetic increase in yield under these environments has been recognized to be a difficult challenge for plant breeders while progress in grain yield has been much higher in favourable environments []. Thus, drought indices which provide a measure of drought based on yield loss under drought conditions in comparison to normal conditions have been used for screening drought-tolerant genotypes []. cholars esearch Library 0

2 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- tress tolerance index (TOL) and mean productivity (M) were defined as the difference in yield and the average yield between stress and non-stress environments, respectively []. Other yield based index is geometric mean productivity (GM) that is often used by breeders interested in relative performance since drought stress can vary in severity in the field environment over years []. Another selection criterion for a high yielding cultivar under drought conditions is stress susceptibility index (I) proposed by Fischer and Maurer []. Lan [0] and Fernandez [] defined new indices of drought resistance index (DI) and stress tolerance index (TI), which were commonly accepted to identify genotypes producing high yield under both stress and nonstress conditions. Fischer et al. [] introduced another index as relative drought index (DI). Bidinger et al. [] suggested drought response index (DI) with its positive values indicating stress tolerance. Other yield based estimates of drought resistance are yield index (YI) [] and yield stability index (YI) []. To improve the efficiency of TI a modified stress tolerance index (MTI) was introduced by Farshadfar and utka [] as ki TI, where ki is a correction coefficient which corrects the TI as a weight. Therefore, k TI and k TI are the optimal selection indices for stress and non-stress conditions, respectively. Fernandez [], divided the manifestation of plants into the four groups of (I) genotypes that express uniform superiority in non-irrigated and irrigated conditions (group A), (II)- genotypes which perform favorably only in nonstress conditions (group B), (III) - genotypes which yield relatively higher only in stress conditions (group C) and (IV) - genotypes which perform poorly in non-irrigated and irrigated conditions (group D). Therefore, as Fernandez stated, the best index for stress tolerance selection is one that can be able to separate group A from others. We believe that the best index for relative tolerance or relative resistance depends on the selection aims (only selection for stability without attention to high yield or selection for commercial aims with attention to stable and high yield) and the conditions of selection ( the selection aim is for non-irrigated or irrigated conditions). Moosavi et al. [] recommended testing of new indices (ATI) that can select group C with more emphasis on Y than I and TOL for identification of relative tolerant genotypes (stable yield in non-irrigated and irrigated conditions), I for better understanding of yield changes and identification of relative tolerant genotypes (stable yield in non-irrigated and irrigated conditions) and NI for selection of relatively resistant genotypes with relatively stable and high yield in non-irrigated and irrigated conditions. The objectives of the present investigation were to screen bread wheat landraces for drought tolerance with high yield potential and stability under water stress conditions. MATEIAL AND METHOD Twenty landraces of bread wheat (Triticum aestivum L.) listed in Table were provided from eed and lant Improvement Institute of Karaj, Iran. They were assessed in a randomized complete block design with three replications under two irrigated and rainfed conditions during 00-0 growing season in the experimental field of the College of Agriculture, azi University, Kermanshah, Iran ( N, E and m above sea level). Mean precipitation in 00 0 was 0.0 mm. The soil of experimental field was clay loam with ph.. owing was done by hand in plots with four rows m in length and 0 cm apart. The seeding rate was 00 seeds per m for all plots. At the rainfed experiment, water stress was imposed after anthesis. Non-stressed plots were irrigated three times after anthesis, while stressed plots received no water. At harvest time, yield potential (Yp) and stress yield (Ys) were measured from rows m in length. Drought resistance indices were calculated using the following relationships: (Y Y ) ) tress susceptibility index = I = []. (Y Y ) ) elative drought index = DI= (Ys/Y p )/ ( Y / Y ) []. ) Tolerance = TOL = Y - Y []. ) Y Y Mean productivity = M = + []. Y Y ) tress tolerance index = TI = []. ) Geometric mean productivity = GM = (Y Y []. Y cholars esearch Library 0

3 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- Y ) Yield index = YI = []. Y ) Yield stability index = Y YI = []. ) Drought response index = DI= (Y A -Y E ) /( E ) []. 0) Drought resistance index (DI) = Ys (Ys/Yp)/ Y [0]. Y ) Modified stress tolerance index = MTI = ki TI, k =Y p / Y k = Y s / Y [] where ki is the correction coefficient. ) Abiotic tolerance index = ATI =[(Yp-Ys) / ( Y / Y )] [ ] []. ) tress susceptibility percentage index = I=[Yp-Ys /( Y )] 00 []. ) tress non-stress production index = = NI = [ ] [ ] []. In the above formulas, Y, Y, Y and Y represent yield under stress, yield under non-stress for each genotype, yield mean in stress and nonstress conditions for all genotypes, respectively. Y A, Y E and E are representative of yield estimate by regression in stress condition, real yield in stress condition and the standard error of estimated grain yield of all genotypes, respectively. For screening drought tolerant genotypes a rank sum () was calculated by the following relationship: ank sum () = ank mean ( ) + tandard deviation of rank (D) and D= ( i) 0.. tatistical analysis Correlation analysis and principal component analysis (CA), based on the rank correlation matrix and biplot analysis were performed by ver., TATITICA ver. and Minitab ver.. EULT AND DICUION Data concerning yield (Yp and Ys ) and indices are given in Table. The estimates of stress tolerance attributes (Table ) indicated that the identification of drought-tolerant genotypes based on a single criterion was contradictory. For example, according to TI, genotypes, and were the most drought tolerant, whereas genotypes 0, and the least relative tolerant genotypes. With regard to GM genotypes, and were the most relative tolerance and according to M genotypes, and showed the most relative tolerance. Mevlüt and ait [] indicated that the genotypes with high TI usually have high difference in yield in two different conditions. They reported in general, similar ranks for the genotypes were observed by GM and M parameters as well as TI, which suggests that these three parameters are equal for selecting genotypes. According to TOL, the desirable drought-tolerant genotypes were, and. As to I, the desirable droughttolerant genotypes were 0, and. According to YI and YI genotypes and were the most and 0 and the least relative tolerant genotypes (Table ). According to DI and DI indices selected the genotypes and as the most relatively tolerant genotypes while for DI the genotypes, and were the most relative tolerant. According to K TI the genotypes, and and according to K TI the genotypes, and were the most relative tolerant. ATI and I discriminated genotypes,, and as the best and as the worst relatively tolerant genotypes, while for NI the genotype,,,, exhibited the best and landraces 0,, displayed the worst relatively resistant genotypes respectively. Correlation analysis Correlation analysis between grain yield and drought tolerance indices (Table ) can be a good criteria for screening the best genotypes and indices used. Yield in stress (Ys) condition was significantly and positively correlated with TI, GM, M, YI, YI, DI, K TI, K TI, DI, NI, I and DI (Table ) therefore these indices are identified as drought tolerance criteria and discriminate drought tolerant genotypes in the same manner. Mollasadeghi [] showed that correlation between M, Yp and Ys was positive. Akçura et al. [0] reported that YI, YI, TI, GM, M and HM were significantly and positively correlated with stress yield and these indices showed that cultivars may be ranked only on the basis of their yield under stress and so does not discriminate genotypes of group A. It cholars esearch Library 0

4 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- was interesting to note positive correlation between I and Yp indicating that stress susceptibility was positively correlated with non-stressed yield [, ]. This suggested that some characteristics that contribute to yield potential may act to increase susceptibility to stress and that selection for both I and Yp may counteract each other. However, Ehdaie and hakiba [] in wheat found that there was no correlation between stress susceptibility and yield under optimum environments. Clarke et al. [] showed that yield-based I index did not differentiate between potentially drought resistant genotypes and those that possessed low overall yield potential. The geometric mean (GM) is often used by breeders interested in relative performance, since drought stress can vary in severity in field environments and over years []. The TI, GM and M were used for screening drought tolerant high yielding genotypes in the both conditions [, ]. Three dimensional plot A three-dimensional plot between Yp, Ys and TI (Fig. ) was used to distinguish the group A genotypes from the other three groups (B, C and D) [, ]. In this case the most desirable genotypes for irrigated and rainfed conditions were, and. Fernandez [] reported that M also was able to differentiate genotypes belong to A-group (including genotypes with high yield performance in both conditions, from the others (B, C or D groups). M is related to yield under drought stress if it is not too severe and the difference between Yp and Ys is not too large. In these cases, genotypes with a high M would belong to A-group []. Golmaghani et al. [] reported that the potential of indices M, GM, TI and YI to identify genotypes with high yield is higher than TOL and I. Khalil Zadeh and Karbalayi Khiavi [], and Farshadfar et al. [] believe that the most suitable indices for selection of drought tolerant cultivars, is an indicator which has a relatively high correlation with grain yield in both stress and non-stress conditions. Therefore the correlation between indices of stress tolerance and yield in both conditions, identify the most suitable indicators for screening drought tolerant genotypes. creening drought tolerant genotypes and indices -rincipal component and biplot analysis for screening drought tolerance indicators The relationships among different indices are graphically displayed in a biplot of CA and CA (Fig. ). The CA and CA axes which justify.0% of total variation, mainly distinguish the indices in different groups. In rincipal component analysis (CA), the Cs axes divided the indices into four groups. Group (G) included only the parameter DI. In group (G), the parameters DI, DI, NI, Ys, K TI, YI, YI, GM, TI, M and I are strongly correlated with yield under rainfed condition indicating that these criteria are suitable for identification of drought tolerat genotypes in the stress condition (group C of Fernandez). Indices K TI and Yp we refer to group (G), and ATI, I and TOL were separated as group (G). In general indices in the same group distinguish drought tolerant genotypes in the same manner. This procedure was also employed in durum and bread wheat [, 0] for screening selection criteria of different climate and water regime conditions. everal studies conducted in Iran measuring drought response of improved wheat varieties [], pure lines derived from winter wheat landraces [], and spring wheat landraces [] revealed that indices such as I and TOL were not efficient to be used in selecting genotypes with high yield capacity in either stressed or non-stressed environments. aba et al. [] reported that TI and were identified as efficient indices. I and TOL indices only assess the plasticities of the genotypes under study, whereas a variety may rank first in both environments but still have higher I and TOL than other varieties. Based on their studies, it seemed that I and TOL were not useful indices to select for drought tolerant genotypes in plant breeding programs, because, I exhibited negligible heritability and TOL was less heritable than other indices usually not identifying genotypes with both high yield and drought tolerance characteristics. On the other hand indiex like TI was moderately heritable and is usually able to select high yielding genotypes in both environments. Golabadi et al. [] reported that selection for TOL will be worthwhile only when the target environment is no-drought stressed. Hohls [] thought that M should increase yield in both environments unless the genetic variance under stress is more than double that under non-stress, and the genetic correlation between yields in contrasting environments is highly negative. Bouslama and chapaugh [] stated that cultivars with a high YI were expected to have high yield under both stress and non-stress conditions. However, io-e Mardeh et al. [] found that cultivars with the highest YI exhibited the least yield under non-stressed conditions and the highest yield under stressed conditions. The two indices namely ATI and I revealed a relative tolerance of a cultivar to drought stress. The nature of ATI and I are such that they rely on crop survival mechanisms in stress conditions although these genotypes can have either high or low yields in two conditions so, they have not exhibited a significant correlation with high Y but cholars esearch Library 0

5 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- have shown a significant correlation with Y. The yield stability is more importance than high yield in rainfed and irrigated conditions for these indices. Although NI and TI are very similar and highly correlated, but in addition to high yield in stress and non-stress conditions, stable yield and high Y are more emphasized in NI than in TI, and these characteristics, make NI a better index than TI for identifying genotypes with stable and high yield in both stress and non-stress conditions []. Therefore, this index is an indicator of the relative stress resistance (because this index select tolerant genotypes with high yield in stress and non-stress conditions) while, indices ATI and I show relative stress tolerance. - Biplot diagram election based on combination of indices may provide a more useful criterion for improving drought resistance of wheat but study of correlation coefficients are useful in finding out the degree of overall linear association between any two attributes. Thus a better approach than correlation analysis and principal component is needed to identify the superior genotypes for both stress and non-stress environments. elationship between genotypes and resistance to drought was used as a biplot for identification of drought tolerant landraces (Fig. ). Biplot diagram showed that the first component was higher and the second component was lower for genotypes,,, and. Thus, selection of these landraces with high C and low C are suitable for both rainfed and irrigation conditions (Fig. ). io-e Mardeh et al. [] and Golabadi et al. [] obtained similar results in multivariate analysis of drought tolerance in different crops. - anking method The estimates of in vivo indicators of drought tolerance (Table ) indicated that the identification of drought-tolerant genotypes based on a single criterion was contradictory. Different indices introduced different landraces as drought tolerant. To determine the most desirable drought tolerant genotype according to the all indices, mean rank and standard deviation of ranks of all in vivo drought tolerance criteria were calculated and based on these two criteria the most desirable drought tolerant genotypes were identified. In consideration to all indices, genotypes (=WC-), (=hishtaz) and (= WC-) exhibited the best mean rank and low standard deviation of ranks in stress condition, hence they were identified as the most drought tolerant genotypes, while genotypes (0= WC-), (= WC-) and (= ishgam) as the most sensitive, hence they are recommended to be used as parents for genetic analysis, gene mapping and improvement of drought tolerance in common wheat. Table. Genotype codes Genotype WC- WC-0 hishtaz ishgam WC- WC- WC- WC- WC-0 WC- Code 0 Genotype WC- WC-00 WC- WC- WC- WC- WC-00 WC- WC- WC- Code 0 cholars esearch Library

6 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- cholars esearch Library Table. anks (), ranks mean ( ) and standard deviation of ranks (D) of drought tolerance indicators TOL I M GM TI Y Y Genotypes Table continued. ATI DI I DI DI YI YI Genotypes

7 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- cholars esearch Library Table continued. D K TI K TI NI Genotypes Table. Correlation coefficients between drought tolerance indices K TI DI I YI YI TI GM M TOL Yp Ys Ys 0.* Yp 0.* -0. TOL ** 0.** M 0.** ** 0.0** GM 0.** 0.0** ** 0.0** TI 0.** 0.** 0.** * 0.** YI 0.** 0.** 0.0** 0.** ** 0.** YI 0.** 0.** 0.** 0.** 0.** ** 0.** I 0.** 0.** 0.** 0.** 0.0** 0.** -0.** ** DI 0. 0.** 0.** 0.* 0.** 0.** 0.** 0.** 0.** 0.* K TI 0.* 0.** 0.** 0.** 0.** 0.0** 0.** 0.** * 0.** K TI ** 0.* 0. 0.** ** ** DI 0.0** ** 0.** -0.0 ATI 0.** -0.** ** 0.** -0. I 0. 0.** 0.0** 0.** 0.** 0.** 0.** 0.** -0.** ** NI ** 0.** 0.** 0.** 0.* 0.* ** ** DI Table continued DI NI I ATI DI K TI K TI 0.0** DI -0.** -0.0 ATI 0.** -0.** -0. I -0.0** ** 0.** NI 0.** -0.** -0.** 0.0** 0.** DI

8 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- Fig.. Three-dimensional plot between Yp, Ys and TI.0 rincipal components analysis (CA) 0. DI DI DI NI G Factor :.% Ys K TI 0.0 YI YI I GM TI -0. M G -.0 Yp K TI G ATI G TOL I Factor :.% Fig.. creening drought tolerance indicators using biplot analysis. cholars esearch Library

9 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- Biplot of Ys,..., DI econd Component (. % ) I ATI TOL 0 KTI Yp M TI GM I YI YI KTI Ys NI DI DI DI First Component (.% ).0. Fig.. Biplot based on first and second components of drought tolerance indices. EFEENCE [] E İlker, Ö. Tatar, F. Aykut Tonk, M. Tosun, Turk J of Field Crops, 0,():-. [] Talebi, F. Fayaz, A.M. Naji, General and Applied lant hysiol, 00, : -. [] A ireivatlou, B.D. Masjedlou,.T. Aliyev, African J of Agricult es, 00, : -. [] J Anwar, G. Mahboob ubhani, H. Makhdoom, A. Javed, H. Mujahid, M. Muhammad, ak J Bot, 0, ():-0. [] amirez, J.D. Kelly, Euphytica,, :-. [] A ichards, G. J. ebetzke, A. G. Condon, A. F. Herwaarden, Crop ci, 00, :. [] J Mitra, Curr ci, 00, 0:-. [] AA osielle, J. Hamblin, Crop ci,, ():-. [] A Fischer,. Maurer, Aust J Agric es,, :-. [0] J Lan, Acta Agricult Bor-occid inic,,:. [] GCJ Fernandez, ublic Tainan Taiwan,, 0. [] A Fischer, J. T. Wood, Aust J Agric es,, 0: [] F Bidinger, V. Mahalakshmi, G. D.. ao, Aust J Agric es,, :-. [] Gavuzzi, F. izza, M. alumbo, Canad J of lant ci,, :. [] M Bouslama, W.T. chapaugh, Crop ci,, :. [] E Farshadfar, J. utka, Acta Agron Hung, 00, 0():. [] Moosavi, B. Yazdi amadi, M.. Naghavi, A. A. Zali, H. Dashti, A. ourshahbazi, DEET, 00, :-. [] A Mevlüt, Ç. ait, Zemdirbyste Agricul, 0, ():-. [] V Mollasadeghi, Thesis of M.c in plant breeding.islamic Azad University, Ardabil branch, 00. [0] M Akçura, F. artigoç, Y. Kaya, J of Animal & lant ci, 0, ():00-0. [] Ceccarelli,. Grando, Euphytica,, : -. [] B Ehdaie, M.. hakiba, Cereal es Commun,, :-. [] JM Clarke,. M. Deauw, T. T. Fownleymith, Crop ci,, :. [] Mohammadi, E. Farshadfar, M. Aghaee, J. hutka, Cereal es Comm, 00 : -. [] E Farshadfar, M. Ghannadha, M. Zahravi, J. utka, lant Breeding, 00, : -. cholars esearch Library

10 Ezatollah Farshadfar et al Annals of Biological esearch, 0, ():0- [] Mohammadi, M. Armion, D. Kahrizi, A. Amri, Int J lant ro, 00, (): -. [] Golmoghani Asl, H. Kzemi Arbat, M.Yarnia, G. Aminzade, L. Golmoghani Asl, T. Ghanifathi, Advances in Environ Biol, 0, (): -. [] GH Khalilzade, H. Karbalai-Khiavi, In proc, of the th Irainan congress of crop sciences. Gilan, Iran, 00, -. [] M Mohammadi,. Karimizadeh, M. Abdipour, Aust J Crop ci, 0, ():. [0] E Farshadfar,. Elyasi, M. Aghaee, Amer J of ci es, 0, :0-. [] D Hassanpanah, M. Moghaddam, M. Valizadeh,. Mahfoozi, In: Abstracts of th Iranian Cong. of Crop ro. and lant Breed. Aug. ep. II, Karaj, Iran., 0-. [] A Tarinejad, M. Moghadam, M.. hakiba, H. Kazemi, M. H. adr-moossavi, In: Abstracts of th Iranian Cong. of Crop rod. and lant Breed. Aug. - ep. II, Karaj, Iran,, -. [] M Moghaddam,.. Alavikia, M.. hakiba, M.. Nishabouri, In: Book of Abstracts, rd Internat. Crop. ci. Cong. ICC. - Aug, Hamburg, Germany, 000, 0. [] J aba, M. Moghaddam, K. Ghassemi, M.. Nishabouri, J. Agric. ci. Technol, 00, : -. [] M. Golabadi, A. Arzani,. A. M. Maibody, Afr. J. Agric. es, 00, : -. [] T Hohls, Euphytica, 00, 0(): -. [] A, io-e -Mardeh, A. Ahmadi, K. oustini, V. Mohammadi, Field Crops es, 00, :-. cholars esearch Library