Introduction
Chickpea is an important grain legume crop with an indeterminate growth habit. Plant height is an important trait in chickpea-breeding programmes, not only because it indirectly contributes to yield but also due to its significance from a mechanical harvesting point of view (Singh et al., Reference Singh, Gaur, Chaturvedi, Hazra and Singh2019). However, because of heat stress, there could be a reduction in yield due to a decline in plant height. In quantitative trait loci (QTL) mapping studies, often the data on plant height scored at maturity (the endpoint phenotype) are subjected to analysis. It is now known that in nature, many developmental traits, including plant height, are controlled by genes that are expressed at different time points in response to prevailing environmental conditions (Gupta et al., Reference Gupta, Kulwal, Mir, Gupta and Varshney2013; Kulwal, Reference Kulwal, Varshney, Pandey and Chitikineni2018). These are known as ‘dynamic traits’. Therefore, studies involving dynamic traits where QTLs are identified using the data of only endpoint phenotypes, can miss several important QTLs which are expressed during the different growth phases of the plant. This is because the genotypic differences for plant height which are visible during the early developmental phases are often neutralized at maturity. This thus necessitates multistage phenotyping for such dynamic traits and can facilitate identification of the growth stage having highest heritability as well as stage-specific QTLs. The aim of this work was to study the heritability for plant height in chickpea at different growth periods under heat stress and to identify marker–trait associations (MTAs).
Experimental
Experimental material comprised of 49 diverse desi chickpea genotypes comprising released varieties and advanced breeding lines from different zones of India (online Supplementary Table S1). Sowing was carried out at three different sowing dates in a lattice design with two replications at 30 cm × 10 cm spacing in two rows of 3 m for each genotype. The first trial was sown in the first week of November (timely sown trial), followed by two delayed sowings in the first weeks of December and January during the seasons of 2015–16 and 2016–17. Delayed sowings were carried out to study the effect of increasing heat stress. Observations on plant height were recorded at 1 month intervals after sowing as well as at maturity in all the six trials. Data of individual trials and on a pooled basis were subjected to analysis. Broad sense heritability was computed following Hanson et al. (Reference Hanson, Robinson and Comstock1956).
Thirty-five simple sequence repeat markers from different linkage groups (LGs) reported earlier to be associated with heat and drought stress-related QTLs were used for MTA analysis (online Supplementary Table S2). DNA extraction and polymerase chain reaction amplification were carried out as per Borse et al. (Reference Borse, Kulwal, Mhase and Jadhav2017). MTA analysis was performed following general linear model (GLM; Q) and mixed linear model (MLM; K/Q + K) approaches using TASSEL 4.0 (Bradbury et al., Reference Bradbury, Zhang, Kroon, Casstevens, Ramdoss and Buckler2007) as described in Jadhav et al. (Reference Jadhav, Rayate, Mhase, Thudi, Chitikineni, Harer, Jadhav, Varshney and Kulwal2015). The P-values found significant (P < 0.05) after 1000 permutations are reported.
Results and discussion
At maturity, plant height was highest in the November sown trial, followed by December, and lowest in the January sown trials, indicating a significant negative effect of an increase in temperatures on plant height (⩾30°C) (online Supplementary Table S3). The maximum growth rate for plant height was observed up to 1 month after sowing in the November sown trial, while in the December and January sown trials it was observed after 1 month of sowing for a period of 2 months. This was also reflected in heritability. The effect of heat stress was most visible in the January sown trial since genotypes matured about 3 weeks earlier than that of the November sown trial during 2015–16 and 4 weeks earlier during 2016–17 (Table 1).
Broad sense heritability estimates (%) and maturity duration (d) for plant height recorded at different growing periods during 2015–16, 2016–17 and using pooled data under timely and late sown conditions
a % Reduction in heritability calculated at maturity compared to heritability at 1 month stage.
Heritability for the November sown trial was highest at 2 months after sowing, while for the December and January sown trials it was highest at 1 month after sowing (Table 1). This was expected because November is the normal period for sowing chickpea in India and the genotypes had a longer period of cooler temperatures for completing their growth than the delayed sown trials (Table 1). Moreover, with the delay in sowing (December and January sown trials) accompanied by an increase in temperatures during the growing period, genotypes attained maximum plant height in the initial month. These differences in heritability could be resolved due to multi-stage phenotyping. It was also observed that reduction in heritability from the 1 month stage to maturity was 21–45% during 2015–16 and 7.77–11.7% during 2016–17 in the November to January sown trials (Table 1).
In association analysis it was observed that all the MTAs identified using endpoint phenotypes were also identified in first and second month plants (Table 2). However, few novel growth stage specific MTAs were identified using data of monthly intervals only. For instance, using endpoint phenotypes, significant MTAs were identified on LGs 1, 3, 4 and 8, while using monthly interval data, in addition to these, new MTAs were identified on LGs 5, 6 and 7 (Table 2). This was expected because the heritability was highest in the initial stage of crop growth as compared to that at maturity. An important and significant MTA with marker TA200 was consistently identified at the 1 month stage in all the seasons, but not at maturity (Table 2). Similarly, one MTA with marker NCPGR203 from LG6 was identified at the 1 and 3 month stages. In addition, a solitary MTA was identified on LG7 with marker TA76 at 1 month stage. It is thus evident that if we score only endpoint phenotypes, then we can miss important MTAs/QTLs which are expressed during the initial growth stages but disappear at maturity. Similar observations were also made in triticale where common as well as some stage-specific QTLs were identified for plant height measured at different growth stages (Würschum et al., Reference Würschum, Liu, Busemeyer, Tucker, Reif, Weissmann, Hahn, Ruckelshausen and Maurer2014).
Details of MTAs identified for plant height at different durations of crop growth using different criteria
H1M, plant height at 1 month; H2M, plant height at 2 months; H3M, plant height at 3 months; HatM, plant height at maturity.
*N, November; D, December; J, January; more than one alphabet in the parenthesis indicate that P-values were significant in those seasons and the season for which P-values are lowest are highlighted in bold.
Q, GLM; K, MLM; Q + K, MLK (with kinship as well as principal components).
R 2, phenotypic variation explained.
In the current study, the maximum numbers of MTAs were identified on LG4 and LG8. These LGs were earlier identified to contain QTL clusters for different drought stress related traits in chickpea by Varshney et al. (Reference Varshney, Thudi, Nayak, Gaur, Kashiwagi, Krishnamurthy, Jaganathan, Koppolu, Bohra, Tripathi, Rathore, Jukanti, Jayalakshmi, Vemula, Singh, Yasin, Sheshshayee and Viswanatha2014). All the MTAs identified in this study using the Q-approach were also identified following the K and/or Q + K approach, except for two MTAs on LG4 during 2016–17 (Table 2). Therefore, in the current study, we not only validated earlier identified MTAs, but also identified stage-specific novel MTAs.
It is thus concluded that environment plays a great role in influencing heritability of any dynamic trait and important QTLs can be missed if only endpoint phenotypes are scored. The findings also suggest that plants having maximum height/growth rate in the first 30 d should be preferred while selecting for promising genotypes under heat stress in chickpea.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1479262120000349
Acknowledgements
During the course of data analysis and writing this manuscript, PLK received funding from the Ministry of Agriculture, Cooperation and Farmers Welfare, Government of India and TVB received a National Fellowship from the University Grants Commission, Govt. of India for the Ph.D. programme. Thanks are also due to the Editor in Chief of PGR for critical comments and suggestions.
