Phenotypic evaluation of agronomic and root related traits for drought tolerance in recombinant inbred line population derived from a chickpea cultivar (C. arietinum L.) and its wild relative (C. reticulatum)

Abstract

Drought is a major abiotic stress that drastically reduces chickpea yields. The present study was aimed to identify drought-responsive traits in chickpea by screening a recombinant inbred line population derived from an inter-specific cross between drought cultivar of GPF2 (C. arietinum L.) and drought sensitive accession of ILWC292 (C. reticulatum), at two locations in India. Twenty-one traits, including twelve morphological and physiological traits and nine root-related traits were measured under rainfed (drought-stress) and irrigated conditions (no-stress). High genotypic variation was observed among RILs for yield and root traits indicated that selection in these germplasms would be useful in achieving genetic progress. Both correlation and principal component analysis revealed that plant height, number of pods per plant, biomass, 100-seed weight, harvest index, membrane permeability index, and relative leaf water content were significantly correlated with yield under both irrigated and drought stress environments. Root length had significant positive correlations with all root-related traits except root length density in drought-stressed plants. Path analysis and multiple and stepwise regression analyses showed that number of pods per plant, biomass, and harvest index were major contributors to yield under drought stress conditions. Thus, a holistic approach across these analyses identified number of pods per plant, biomass, harvest index, and root length as key traits for improving chickpea yield through indirect selection for developing drought-tolerant cultivars. Overall, on the basis of yield components morphological and root traits, a total of 15 promising RILs were identified for their use in chickpea breeding programs for developing drought tolerant cultivars.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12298-022-01218-z.

Introduction

Chickpea (Cicer arietinum L.) is a cool-season, self-pollinated grain legume crop from south-eastern Turkey (Ladizinsky 1975). The major chickpea-producing countries are India, Australia, Pakistan, Turkey, Ethiopia, Myanmar, Mexico, Canada, Iran, and the United States. India alone generates about 70% of global production (FAOSTAT 2017). Chickpea seeds are high in protein (20–22%) and carbohydrates (~ 60%). The seeds also contain good levels of essential amino acids in a readily digestible form (Williams and Singh 1987; Jukanti et al. 2012) and vitamins, minerals, and dietary fiber, which help to control blood sugar levels and improve insulin secretion (Karim and Fattah 2007).

Despite growing international demand, chickpea yield is low and unstable (Millan et al. 2006) due to its sensitivity to biotic and abiotic stresses particularly drought (Thudi et al. 2014; Kushwah et al. 2021a), particularly drought (Krishnamurthy et al. 2010). About 90% of chickpea is sown under rainfed conditions with stored soil moisture (Kumar and Abbo 2001). Drought stress is well-known for reducing the crop growth duration in various crops, thus affecting yield components, i.e., total biomass, pod number, seed number, seed weight, seed quality, and yield per plant (Krishnamurthy et al. 2013; Toker et al. 2007). However, early flowering in chickpea can be an advantage as more pods are set before the occurrence of drought and heat stress, due to a comparatively longer period of reproductive growth (Kushwah et al. 2020b). Drought stress promotes evapo-transpiration, which reduces conserved soil water as the season progresses, shortens the growing season, and reduces yields (Toker et al. 2007). Up to 50% of annual yield losses in chickpea are attributed to drought stress alone (Sabaghpour et al. 2006; Varshney et al. 2010), emphasizing the urgent need to focus on increasing chickpea productivity under drought stress. Thus, it is essential to develop cultivars that can attain their maximum potential under drought stress or rainfed environments.

Drought tolerance is the comparative ability to maintain adequate biomass and crop yield under a limited water supply (Serraj and Sinclair 2002). It is a complex quantitative trait prone to significant genotype by environment (G × E) interactions (Kashiwagi et al. 2008; Kushwah et al. 2020a). The direct selection of genotypes with high yields under stress conditions is largely hampered by G × E interactions in the field (Kushwah et al. 2021b). Understanding yield stability under drought stress becomes more difficult due to the presence of various underlying mechanisms used by plants to survive, such as drought escape, drought avoidance, and drought tolerance (Tuberosa and Salvi 2006). These mechanisms led to the development of early maturing varieties and the selection of genotypes with extensive root systems for better productivity under drought stress (Kashiwagi et al. 2005). Genetic improvement in traits responsible for drought tolerance can be a long-lasting strategy for improving yield under drought stress. Since grain yield is profoundly affected by high G × E interactions and low heritability, a trait-based breeding strategy is preferred over the yield-based breeding approach to conquering G × E interaction effects on yields under drought stress (Ludlow and Muchow 1990; Kushwah et al. 2021c).

Phenotypic attributes of the root system are expected to have a direct effect on transpiration in plants under drought stress. A profuse root system is expected to absorb more soil water than a less extensive root system under drought stress. Substantial progress has been made using the strength of a larger and deeper root system to improve chickpea yields under drought stress (Purushothaman et al. 2016). Screening of chickpea mini core germplasm collection at an early growth stage reflected that total root biomass, rooting depth, root length density, and root prolificacy contributed toward higher yields under terminal drought as these traits help to extract more soil moisture (Kashiwagi et al. 2005, 2006a; Varshney et al. 2011). Shoot biomass production and harvest index have also been correlated with a profuse root system with greater rooting depth and root prolificacy (Zaman-Allah et al. 2011; Kashiwagi et al. 2013; Purushothaman et al. 2016). Yet, some studies suggest that profuse rooting is not essential for higher grain yield or biomass production due to the needless energy loss from its vigorous respiration (Vadez et al. 2008). Thus, research is needed to resolve these contradictions on the contribution of root traits to drought tolerance. Detailed studies on various root traits are difficult due to low heritability and complex mechanisms of these traits, variable expression across soil environments, and the labor-intensive nature of the studies (Gaur et al. 2008; Varshney et al. 2011). High-throughput phenotyping of root-related traits using polyvinyl chloride (PVC) pipe-based cylinder culture system has been used successfully for studying root-related traits in a large number of chickpea genotypes and advanced breeding lines (Kashiwagi et al. 2005, 2006b; Krishnamurthy et al. 2010; Upadhyaya et al. 2012). The presence of strong correlations between profuse root systems and grain yield has been reported in some studies (Zaman-Allah et al. 2011; Kell 2011; Schoppach et al. 2013; Bishopp and Lynch 2015).

Other morphological and physiological characteristics are equally important for providing drought tolerance in chickpea. The association of these putative traits with grain yield under drought stress has been confirmed previously (Purushothaman et al. 2016). High heritability and fewer G × E interaction effects make harvest index (HI) an important trait for improving yield under drought stress (Hay 1995). However, considering HI alone as the selection criterion for drought tolerance could be problematic due to its association with low biomass potential (Wallace et al. 1993). Several physiological traits such as membrane permeability index, photosynthetic efficiency, relative leaf water content, chlorophyll content, proline accumulation, and ABA content and morphological traits such as days to germination, days to flowering, plant height, biomass, and 100-seed weight have been proposed for the selection of drought-tolerant genotypes (Maqbool et al. 2016; Kushwah et al. 2020b).

Breeding for drought tolerance requires knowledge of the mechanisms adopted by plants to sustain productivity under stress, as well as the genetic basis of the numerous morphological and physiological traits responsible for drought tolerance (Purushothaman et al. 2017). Despite several studies in this context, the importance of the traits responsible for drought tolerance remains unclear, resulting in poor exploitation of critical characteristics in drought-tolerance breeding programs. The genetic advancement for yield is low in chickpea due to the presence of inadequate genetic variation (van Rheenen et al. 1993). The exploitation of wild species through interspecific hybridization broadens the genetic base of chickpea. It facilitates the transfer of desirable alleles from wild accessions into the cultivated ones and consequently, plays an important role in the genetic enhancement of yield (Singh et al. 2008). Though the wild accession of ILWC292 (C. reticulatum) used to develop RIL population is drought sensitive, yet it has more root length density, root to shoot ratio and membrane permeability index as compared to the cultivated parent of GPF 2 which has superior root related traits as well as yield contributing traits. Therefore, an inter-specific RIL population was developed by crossing chickpea cultivar of GPF2 (C. arietinum) and wild accession of ILWC292 (C. reticulatum) to combine the superior drought related traits and yield contributing traits of both parents. The present study assessed RIL population derived from inter-specific cross under irrigated and drought stress conditions to estimate genetic variability and heritability of the desirable traits, identify the key traits responsible for drought tolerance, understand the relationship of important traits with yield and obtain drought tolerant lines with high yield potential.

Materials and methods

Plant materials and experimental sites

A set of 202 RILs (F₈-generation) was developed from an inter-specific cross of drought tolerant cultivar GPF2 (C. arietinum L.) with a drought-sensitive accession ILWC292 (C. reticulatum) using the single seed descent method. Chickpea cultivar GPF2 is a semi erect, medium tall cultivar recommended for cultivation in Punjab state and in North Western Plains Zone of India. This is a drought tolerant high yielded chickpea cultivar resistant to fusarium wilt and ascochyta blight. Another parent of RILs, ILWC292 (C. reticulatum) having semi prostrate growth habit. It is sensitive to drought and susceptible to ascochyta blight disease, but resistant to botrytis grey mould disease and chickpea cyst nematode. In spite of drought sensitivity, it possesses some desirable drought related traits such as more root length density, root-to-shoot ratio and membrane permeability index. In 2017, seeds of RIL population were assessed, together with their parents in an alpha lattice design (17 × 12) under irrigated (non-stress) and rainfed (drought-stress) conditions at two locations (Ludhiana and Faridkot) with three replications. Each RIL was planted in 2 m long paired-rows at 30 cm × 10 cm spacing. The Ludhiana (30.9010° N, 75.8573° E) and Faridkot (30.6769° N, 74.7583° E) sites are categorized as a semi-arid sub-tropical region and semi-arid dry region, respectively. Both sites comprise loamy sand with 59.8% sand and 16.5% clay (Typic Ustorthents). The average annual rainfall is 700 mm at Ludhiana and 450 mm at Faridkot, of which more than 70% occurs from July to September.

Sowing was done on 30th October on the residual moisture which was sufficient for good germination, as recommended for chickpea sowing in this region. Soil moisture was measured at the time of sowing, 70 days, 90 days, 110 days and 130 days of planting and at the time of maturity in irrigated and rainfed conditions at both locations (Table 1a). Essential irrigation was applied to the irrigated plots at regular intervals as and when required, while no irrigation was applied to the rainfed plots. In the case of rainfed plots, the soil moisture was ideal for drought conditions for chickpea crop. As a result of drastic reduction in soil moisture content at 90, 110 and 130 days of planting in rainfed plots as compared to irrigated plots induced sufficient drought stress at reproductive (flowering, pod formation) and pod development stages at both locations.

Table 1.

Soil moisture contents in the (a) experimental fields at Ludhiana and Faridkot, (b) PVC cylinders at Ludhiana and Faridkot

Stage of recording soil moisture	Soil moisture content (%)
	Ludhiana		Faridkot
	Irrigated (Control)	Rainfed (Drought)	Irrigated (Control)	Rainfed (Drought)
(a)
Initial moisture content in the soil	21.5	21.2	22.8	21.6
Moisture of the soil at 70 days of planting	16.3	15.8	19.3	14.3
Moisture of the soil after 90 days of planting	20.6	13.7	21.6	13.2
Moisture of the soil after 110 days of planting	21.4	15.3	22.3	12.6
Moisture of the soil after 130 days of planting	19.2	12.4	20.4	11.3
Moisture of the soil at maturity	13.7	10.6	15.3	9.2
(b)
Initial moisture content in the soil	19.5	19.5	20.6	20.6
Moisture of the soil at 20 days of planting	18.3	15.6	19.3	15.7
Moisture of the soil after 40 days of planting	18.6	12.7	18.9	12.8

Phenotyping under field conditions for morphological and physiological traits

Phenotypic data were collected for 12 morphological and physiological traits, such as days to germination (DG), days to flowering initiation (DFI), days to 50% flowering (DFF), days to 100% flowering (DHF), plant height at physiological maturity(PH), number of pods per plant (NPP), biomass (BIO), seed yield (YLD), 100-seed weight (HSW), harvest index (HI), membrane permeability index (MPI) and relative leaf water content (RWC). The MPI and RWC were measured at 50% flowering stage.

In each plot, five randomly taken plants were used to record PH, NPP, BIO, and YLD observations. The data for DG, DFI, DFF, DHF, and HSW was recorded on a plot basis. HI was calculated as:

$$\text{Harvest}\text{index}=\left(\text{seed}\text{yield}/\text{total}\text{shoot}\text{biomass}\right)\times 100$$

Membrane permeability index was determined at 50% flowering stage according to the modified method of Premchand et al. (1990), as follows:

$$\text{Membrane}\text{permeability}\text{index}\text{(MPI)}=[1-\left(\text{C1}/\text{C2}\right)]\times 100$$

where C1 is the initial electrical conductivity at 40 °C and C2 is the final electrical conductivity at 100 °C.

RWC was calculated at 50% flowering stage using the formula given by Slavik (1974):

$$\text{RWC}(\%)=\frac{\left(\text{FW}-\text{DW}\right)}{\left(\text{TW}-\text{DW}\right)}\times 100$$

where FW is fresh weight, DW is dry weight, and TW is turgid weight.

Phenotyping for root and related traits

For measuring root traits, PVC pipe-based cylinder culture approach was used for recording root- and shoot-related traits. Chickpea plants were grown in PVC cylinders (18 cm diameter, 120 cm height) with three replications at Ludhiana and Faridkot locations. The PVC cylinders, except for the top 15 cm, were filled with an equi-mixture (w/w) of vertisol and sand, mixed with 0.07 g kg⁻¹ diammonium phosphate. A mixture of soil and sand was used to decrease the soil bulk density and facilitate root growth and extraction. The soil water content of the mixture was equilibrated to 70% field capacity to create conditions similar to those in the field at sowing when the soil is not fully saturated with water. Need based irrigation was applied in each cylinder under irrigated conditions, while no irrigation was given under drought conditions. Soil moisture was measured at the time of sowing, 20 days and 40 days of planting (Table 1b). At 40 days of planting, the soil moisture was reduced to a level sufficient for generating drought stress under drought conditions, while in irrigated conditions, sufficient soil moisture was maintained by applying essential irrigations at regular intervals. Sampling was done 40 days after sowing (DAS), avoiding physically damaged plants, as Krishnamurthy et al. (1996) have shown that maximum variations in root-related traits among genotypes in this environment are best detected at this stage and that the variation decreases after 41 DAS.

Root samples were collected using steel soil-coring tubes (50 mm diameter) to a depth of 120 cm in each PVC cylinder at the flowering stage. Each RIL sample comprised three soil cores, which were pooled to increase the sample size. The soil cores for each sample were soaked overnight in water, and the roots recovered by passing the suspension through a 2 mm wire mesh sieve. Chickpea roots were separated manually from debris and weed roots. Phenotypic data were collected for 9 root and shoot traits, such as root length (RL), shoot length (SL), root to shoot ratio (RSR), root length density (RLD), fresh root weight (FRW), fresh shoot weight (FSW), root dry weight (RDW), shoot dry weight (SDW), and the ratio of root dry weight to total plant dry weight (RDW/TDW).

Total RL and FRW were measured, and then before measuring RDW, roots were oven-dried at 70 °C for 72 h. Likewise, FSW was measured, and then before measuring SDW, shoots were oven-dried at 70 °C. RLD (cm cm⁻³) was measured as root length (cm)/volume of soil core (cm³), while the root to shoot ratio (RSR) was calculated using root and shoot lengths.

Statistical analyses

Analysis of variance (ANOVA) was undertaken for each environment using a mixed model analysis to estimate the contribution of each factor to total variation using SAS-software version 9.3 (SAS Institute, Cary NC). The variability of each trait was estimated by simple statistical measures, such as mean, range, phenotypic and genotypic variances, and coefficients of variation. Variances and coefficients of variation were calculated as per Singh and Chaudhary (1985). Broad sense heritability (h_b²) was calculated using the formula given by Allard (1960).

$${\text{h}}_{\text{b}}^2=({\mathrm{\sigma }}_{\text{g}}^2/{\mathrm{\sigma }}_{\text{p}}^2)\times 100$$

where σ²_g is the genotypic variance and σ²_p is the phenotypic variance.

Expected genetic advance (GA) was calculated using the formula given by Allard (1960):

$$\text{GA}=\text{k}\times {\text{h}}_{\text{b}}^2\times \surd {\mathrm{\sigma }}_{\text{p}}^2$$

The GA as a percentage of the mean was estimated as:

$$\text{GA}(\%\text{of}\text{mean)}=\left(\text{GA}/\mathrm{\mu }\right)\times 100$$

where k is the selection differential (2.06) at 5% selection intensity, h_b² is broad sense heritability, σ²_p is phenotypic standard variation, and μ is the mean.

A matrix of simple correlation coefficients between seed yield and its components was computed using SAS-software version 9.3 to determine the relationship between the examined traits and seed yield. The path coefficients were analyzed according to the method suggested by Dewey and Lu (1959) using the matrix method in which yield was taken as the dependent variable and the rest of the traits were considered as independent variables. A matrix inverse of correlation coefficients was used to calculate the direct and indirect effects by INDOSTAT. Multiple linear regression analysis, step-wise linear regression analysis, and principal component analysis were performed using SAS-software version 9.3 in which all the traits were considered as independent variables while yield was taken as dependent variable.

Results

Phenotypic performance of RIL population and parents

The RILs, along with parents, were evaluated in the irrigated (non-stress) and rainfed (drought-stress) treatments at two research locations (Ludhiana and Faridkot, Punjab, India). Significant variation was observed in the RILs and parents for most of the morphological, physiological, and root-related traits in both the irrigated and rainfed treatments (Tables 2, 3). All of the measured morphological and physiological traits were significantly affected by drought stress, except for HI. In the rainfed treatment, the RILs had slightly higher mean values (8.68%) for days to germination (DG) than the irrigated one. As expected, the RILs had lower mean values for days to flowering initiation (DFI), days to 50% flowering (DFF), and days to 100% flowering (DHF) in the rainfed treatment by 16.11%, 14.48%, and 13.76%, respectively than the irrigated treatment. Similarly, significantly lower mean values of plant height (PH) (30.91%), number of pods per plant (NPP) (42.50%), biomass (BIO) (40.15%), and yield (YLD) (44.18%) were observed under rainfed treatment. The rainfed treatment caused a moderate reduction in the mean performance of RILs for 100-seed weight (HSW) (18.28%), relative to the irrigated treatment. Harvest index (HI) was least affected by the rainfed treatment (0.09% reduction), as compared to the irrigated treatment. The rainfed treatment significantly increased the mean values of RILs for membrane permeability index (MPI) (16.34%) and decreased the mean values of relative leaf water content (RWC) by 19.53%, relative to the irrigated treatment.

Table 2.

Mean performance of a chickpea RIL population for various morphological and physiological traits in the irrigated (IR) and rainfed (RF) treatments at both locations (Ludhiana and Faridkot)

Trait		DG	DFI	DFF	DHF	PH	NPP	BIO	YLD	HSW	HI	MPI	RLWC
ILWC292 (Susceptible parent)	IR RF	12.34 13.70	90.33 78.15	94.26 82.14	98.08 84.93	42.42 23.55	43.53 18.09	76.78 40.72	27.91 11.78	11.27 10.33	36.63 28.89	42.23 50.24	65.28 48.11
GPF 2 (Tolerant parent)	IR RF	8.12 8.53	82.80 70.10	86.46 73.57	89.78 76.45	58.87 48.52	68.54 43.71	113.32 77.51	49.74 36.77	16.18 16.09	43.55 47.96	28.81 36.17	88.31 78.16
Contrast analysis between parents	IR RF	44.33** 194.89**	78.87** 28.44**	171.50** 28.91**	171.50** 30.18**	65.43** 538.06**	133.10** 169.97**	55.17** 450.56**	232.07** 200.39**	1629.73** 714.84**	15.81** 54.76**	150.71** 91.06**	178.91** 429.24**
Mean (RILs)	IR RF	9.33 10.14	85.99 72.14	89.46 76.51	93.10 80.33	45.68 31.56	47.39 27.25	81.33 48.68	32.14 17.94	14.22 11.62	38.98 35.37	39.42 47.12	74.85 60.23
CV	IR RF	5.05 10.10	2.01 9.45	2.08 8.75	1.96 8.20	10.47 26.12	21.75 35.25	15.47 33.28	27.07 50.90	16.11 20.33	16.48 19.94	12.12 13.02	9.37 15.97
Range	IR RF	8.12–12.34 8.27–13.89	78.91–90.33 55.53–87.61	82.35–94.26 60.19–91.55	86.67–98.08 64.44–95.80	33.82–58.87 12.74–48.52	25.13–75.07 12.69–50.09	51.55–113.70 18.68–83.48	14.13–54.69 7.31–45.79	9.79–18.42 7.16–17.10	22.49–52.86 23.55–56.56	28.70–50.76 31.56–58.10	59.06–89.94 44.80–79.65
Genotypic variance	IR RF	1.50** 2.35**	1.96** 15.24**	2.24** 25.76**	2.28** 16.30**	4.05** 21.30**	18.64** 37.02**	10.35** 35.06**	18.16** 30.72**	20.41** 25.40**	12.52** 8.36**	15.49** 19.61**	6.07** 14.39**
G × E variance	IR RF	0.51 0.15	0.15 0.61	0.14 0.52	0.20 0.59	3.85** 4.58**	5.82** 4.12**	4.12** 6.58**	6.12** 9.12**	18.35** 9.89**	5.31** 7.53**	8.51** 3.08**	10.17** 8.96**
H2 (Broad-sense)	IR RF	0.56 0.42	0.35 0.88	0.25 0.89	0.30 0.88	0.79 0.89	0.85 0.89	0.84 0.90	0.88 0.89	0.90 0.90	0.86 0.87	0.89 0.89	0.87 0.89
GCV (%)	IR RF	4.30 6.34	1.08 7.86	0.83 7.24	0.86 6.72	9.74 21.42	17.33 28.43	13.10 27.00	22.96 41.01	13.92 17.03	14.10 17.65	10.15 10.81	9.13 13.82
PCV (%)	IR RF	5.73 8.77	1.82 8.36	1.65 7.69	1.58 7.15	10.93 22.70	18.77 30.04	14.27 28.52	24.47 43.42	14.65 17.94	15.17 18.86	10.78 11.48	9.82 14.66
GAM (%)	IR RF	6.66 9.45	1.32 15.22	0.86 14.04	0.97 13.01	17.88 41.62	32.97 55.41	24.76 52.68	44.37 69.79	27.27 33.31	27.00 34.02	19.69 20.98	17.51 26.84

DG = days to germination, DFI = days to flowering initiation, DFF = days to 50% flowering, DHF = days to 100% flowering, PH = plant height (cm), NPP = Number of pods per plant, BIO = biomass per plant (gm), YLD = yield per plant (gm), HSW = 100-seed weight (gm), HI = harvest index (%), MPI = membrane permeability index, RLWC = relative leaf water content (%), CV = coefficient of variation, G × E = genotype by environment interaction, H² = broad-sense heritability, GCV = genotypic coefficient of variation, PCV = genotypic coefficient of variation, GAM = genetic advancement as percentage of means at 5% level of significance

**Significant at 1% probability level

Table 3.

Mean performance of a chickpea RIL population for root- and shoot-related traits in the irrigated (IR) and rainfed (RF) treatments at both locations (Ludhiana and Faridkot)

Trait		RL	SL	RSR	RLD	FRW	FSW	RDW	SDW	RRDWTDW
ILWC292 (susceptible parent)	IR RF	84.82 88.98	26.60 17.07	3.19 5.20	9.14 10.52	8.99 7.93	9.40 5.71	2.00 1.39	2.49 1.97	0.45 0.42
GPF 2 (tolerant parent)	IR RF	109.60 125.71	35.65 31.82	3.08 3.97	6.71 6.64	11.46 12.26	15.04 17.32	3.00 4.69	3.61 4.25	0.46 0.52
Mean (RILs)	IR RF	90.84 92.10	28.72 24.72	3.26 3.86	9.33 9.35	9.72 9.64	11.93 8.90	2.30 2.44	2.99 2.57	0.43 0.47
CV	IR RF	17.08 17.86	19.22 21.95	18.92 20.50	23.97 20.21	17.06 17.39	27.94 43.68	28.62 45.20	22.43 30.18	12.58 21.63
Range	IR RF	57.37–128.17 56.10–127.46	16.84–47.99 13.56–41.36	1.90–4.89 1.99–6.94	5.67–13.85 4.67–14.24	6.37–13.73 5.71–13.62	7.68–25.04 4.79–24.92	1.17–3.85 0.37–5.80	2.13–5.47 1.60–5.60	0.25–0.55 0.18–0.74
Genotypic variance	IR RF	24.25** 20.23**	7.20** 10.81**	4.50** 6.40**	12.29** 28.89**	10.46** 17.26**	33.77** 31.96**	10.28** 15.49**	31.62** 26.99**	5.09** 8.46**
G × E variance	–	1.69**	2.13**	1.74**	7.26**	1.61**	5.35**	3.83**	5.03**	3.22**
H2 (Broad-sense)	IR RF	96.30 95.40	88.00 92.20	78.80 85.50	92.60 96.90	91.30 94.90	97.40 97.10	91.20 94.20	97.20 96.70	81.30 89.20
GCV (%)	IR RF	16.73 17.44	17.86 20.76	16.32 18.67	23.02 19.97	16.11 16.87	27.63 42.41	27.05 43.20	22.11 29.25	11.31 20.28
PCV (%)	IR RF	17.05 17.85	19.04 21.62	18.38 20.19	23.92 20.29	16.85 17.32	28.00 43.03	28.32 44.52	22.42 29.75	12.54 21.47
GAM (%)	IR RF	33.82 35.09	34.51 41.07	29.85 35.56	45.64 40.48	31.71 33.84	56.16 86.09	53.20 86.36	44.89 59.25	21.01 39.45

RL = root length, SL = shoot length, RSR = root-shoot ratio, RLD = root length density, FRW = fresh root weight, FSW = fresh shoot weight, RDW = root dry weight, SDW = shoot dry weight, RRDWTDW = ratio of root dry weight and total plant dry weight, CV = coefficient of variation, G × E = genotype by environment interaction, H² = broad-sense heritability, GCV = genotypic coefficient of variation, PCV = genotypic coefficient of variation, GAM = genetic advancement as percentage of means at 5% level of significance

^**Significant at 1% probability level

The contrast analysis of parents for the morphological characteristics identified highly significant differences between parents in both the irrigated and rainfed treatments. The pooled ANOVA, analyzed using both locations, for all measured traits showed highly significant differences between genotypes at both locations in the irrigated and rainfed treatments. The interactions of genotype and environment were highly significant for all traits, except DG, DFI, DFF, and DHF (Tables 2, 3).

The rainfed treatment significantly reduced the mean values of RILs for shoot length (SL) (13.93%), fresh shoot weight (FSW) (25.40%), and shoot dry weight (SDW) (14.05%), and slightly reduced for fresh root weight (FRW) (0.82%), relative to the irrigated treatment. The rainfed treatment significantly increased the mean performance of RILs for root to shoot ratio (RSR) (18.40%), moderately increased root dry weight (RDW) (6.09%) and ratio of root dry weight to total plant dry weight (RDW/TDW) (9.30%), and slightly increased for root length (RL) (1.39%) and root length density (RLD) (0.21%), relative to the irrigated treatment. In both treatments, the frequency distribution of most traits, including root-related traits, was normal (Figs. S1, S2, S3 and S4).

Genetic variability, heritability, and genetic advance

Estimates of the phenotypic coefficient of variation (PCV) are generally higher than those of the genotypic coefficient of variation (GCV), which is in accordance with our results. The GCV and PCV values in the rainfed treatment were high for PH, NPP, BIO, and YLD (21.42–41.01% for GCV and 22.70–43.42% for PCV), moderate for HSW, HI, MPI, and RWC (10.81–17.65% for GCV and 11.48–18.86% for PCV), and low for DG, DFI, DFF, and DHF (6.34–7.86% for GCV and 7.15–8.77% for PCV) (Table 2). In the irrigated treatment, the GCV and PCV values were low to moderate for all measured traits. For the root-related traits in the rainfed treatment, the GCV and PCV values were high for FSW, RDW, and SDW (29.25–43.20% for GCV and 29.75–44.52% for PCV) and moderate for RL, SL, RSR, RLD, FRW, and RDW/TDW (16.87–20.76% for GCV and 17.32–21.62% for PCV). In the irrigated treatment, the GCV and PCV values were high for RLD, FSW, RDW, and SDW, and moderate for the remaining traits (Table 3).

In the irrigated treatment, PH, NPP, BIO, YLD, HSW, HI, MPI, and RWC had higher broad-sense heritability (79.40–90.40%), while DG, DFI, DFF, and DHF showed low to moderate heritability values (25.10–56.40%). In the rainfed treatment, all traits had high heritability (87.50–90.10%), except for DG, which had moderate heritability (42.40%) (Table 2). The root-related traits had a high broad sense of heritability in both irrigated and rainfed treatments (Table 3). The genetic advance as percent of means (GAM) for all measured traits ranged from 0.86 to 44.37% in the irrigated treatment and from 9.45% to 69.79% in the rainfed treatment (Table 2). In the irrigated treatment, GAM values were moderately high for all traits, except for DG, DFI, DFF, and DHF. In the rainfed treatment, GAM values were high for NPP, BIO, and YLD, moderate for PH, HSW, HI, MPI, and RWC, and low for DG, DFI, DFF, and DHF. The GAM values for the root-related traits ranged from 21.01 to 56.16% in the irrigated treatment and from 33.84 to 86.36% in the rainfed treatment (Table 3). In the irrigated treatment, GAM values were high for FSW and RDW, and moderate for the remaining traits. Likewise, GAM values were high for FSW, RDW, and SDW and moderate for the remaining traits in the rainfed treatment.

Association among traits and path analysis

The significance of independent indirect traits in the selection process under drought stress can be analyzed through associations with dependent traits, such as yield. The correlation coefficients among measures traits indicated that YLD had significant and positive correlation with PH, NPP, BIO, HSW, HI, and RWC in both irrigated and rainfed treatments. In contrast, YLD had significant and negative association with DG, DFI, DFF, DHF, and MPI (Fig. 1). Significant and positive correlations occurred consistently among DG, DFI, DFF, DHF, and MPI in both irrigated and rainfed treatments. A significant positive association also occurred consistently among PH, NPP, BIO, YLD, HSW, and RWC in both irrigated and rainfed treatments.

Fig. 1.

Phenotypic correlation coefficients for various morphological and physiological traits in a chickpea RIL population in irrigated treatment (a) and rainfed treatment (b) at both locations (Ludhiana and Faridkot). Positive correlations are in blue and negative correlations in red color. White color represents non-significant correlation. Significance is set at 5% probability level (note: dg = days to germination, dfi = days to flowering initiation, dff = days to 50% flowering, dhf = days to 100% flowering, ph = plant height, npp = Number of pods per plant, bio = biomass per plant, yld = yield per plant, hsw = 100-seed weight, hi = harvest index, mpi = membrane permeability index, rlwc = relative leaf water content)

For root-related traits, the correlation analysis indicated that RL had a significant and positive correlation with SL, RSR, FRW, FSW, RDW, SDW, and RDW/TDW and an unexpected significant negative association with RLD (Fig. 2).

Fig. 2.

Phenotypic correlation coefficients for various root-related traits in a chickpea RIL population in the irrigated (IR) treatment (a) and rainfed (RF) treatment (b) at both locations (Ludhiana and Faridkot). Positive correlations are in blue and negative correlations in red color. White color represents non-significant correlation. Significance is set at 5% probability level (note: rl = root length, sl = shoot length, rsr = root-shoot ratio, rld = root length density, frw = fresh root weight, fsw = fresh shoot weight, rdw = root dry weight, sdw = shoot dry weight, rrdwtdw = ratio of root dry weight and total plant dry weight)

Path analysis identified that BIO and HI were the major contributors to YLD in the irrigated and rainfed treatments at both locations. BIO had the greatest positive indirect effect on YLD through NPP, followed by PH and HSW in both treatments at both locations (Table 4). BIO and HI had the greatest negative indirect effect on YLD through MPI in each moisture condition averaged over two environments (two locations).

Table 4.

Direct (highlighted in bold) and indirect contribution of various morphological and physiological traits on grain yield in RIL population of chickpea in the irrigated and rainfedtreatments at both locations (Ludhiana and Faridkot)

Traits		DG	DFI	DFF	DHF	PH	NPP	BIO	HSW	HI	MPI	RLWC
DG	IR RF	0.006 0.019	0.003 0.008	0.002 0.008	0.002 0.007	–0.001 –0.006	–0.001 –0.005	–0.001 –0.006	0.001 –0.006	0.000 –0.001	0.000 0.005	–0.001 –0.005
DFI	IR RF	–0.024 –0.016	–0.051 –0.041	–0.048 –0.041	–0.046 –0.041	0.013 0.014	0.010 0.011	0.011 0.012	0.003 0.013	0.005 0.008	–0.007 –0.013	0.013 0.009
DFF	IR RF	0.022 –0.037	0.074 –0.093	0.079 –0.093	0.077 –0.093	–0.016 0.032	–0.014 0.026	–0.014 0.028	–0.008 0.030	–0.009 0.020	0.010 –0.030	–0.017 0.021
DHF	IR RF	–0.007 0.046	–0.023 0.122	–0.025 0.123	–0.025 0.123	0.006 –0.042	0.004 –0.033	0.005 –0.036	0.002 –0.039	0.002 –0.027	–0.003 0.040	0.006 –0.027
PH	IR RF	0.006 –0.011	0.008 –0.012	0.006 –0.012	0.007 –0.012	–0.032 0.035	–0.018 0.029	–0.025 0.032	–0.016 0.029	–0.006 0.014	0.018 –0.030	–0.015 0.028
NPP	IR RF	0.005 –0.005	0.012 –0.005	0.011 –0.005	0.011 –0.005	–0.034 0.016	–0.060 0.019	–0.056 0.018	–0.048 0.017	–0.036 0.013	0.052 –0.018	–0.040 0.016
BIO	IR RF	–0.085 –0.195	–0.141 –0.202	–0.114 –0.200	–0.119 –0.198	0.495 0.633	0.581 0.646	0.627 0.680	0.492 0.613	0.293 0.358	–0.525 –0.632	0.408 0.594
HSW	IR RF	0.001 0.008	–0.001 0.008	–0.001 0.008	–0.001 0.008	0.007 –0.021	0.011 –0.022	0.011 –0.023	0.014 –0.025	0.010 –0.017	–0.012 0.023	0.009 –0.021
HI	IR RF	–0.001 –0.021	–0.050 –0.103	–0.061 –0.106	–0.048 –0.110	0.092 0.210	0.308 0.331	0.240 0.264	0.374 0.333	0.514 0.503	–0.413 –0.368	0.327 0.342
MPI	IR RF	–0.001 0.005	–0.008 0.006	–0.008 0.006	–0.007 0.006	0.032 –0.016	0.051 –0.018	0.049 –0.018	0.052 –0.018	0.047 –0.014	–0.059 0.019	0.041 –0.017
RLWC	IR RF	–0.003 0.012	–0.005 0.011	–0.004 0.011	–0.004 0.011	0.008 –0.042	0.011 –0.044	0.011 –0.045	0.011 –0.042	0.011 –0.035	–0.012 0.046	0.017 –0.051

Residual effect of irrigated treatment = 0.075 and rainfedtreatment = 0.105, DG = days to germination, DFI = days to flowering initiation, DFF = days to 50% flowering, DHF = days to 100% flowering, PH = plant height, NPP = number of pods per plant, BIO = biomass per plant, YLD = yield per plant, HSW = 100-seed weight, HI = harvest index, MPI = membrane permeability index, RLWC = relative leaf water content

Regression analysis

The analysis revealed that NPP, BIO, and HI contributed positively, while MPI, HSW, and RWC contributed negatively to the total variation of YLD in the rainfed treatment at both locations. In the irrigated treatment, BIO and HI contributed positively, while MPI and DHF contributed negatively to the total variation in YLD averaged over two locations (Table 5).

Table 5.

Regression coefficients (b), standard errors (SE), t-values and probabilities of estimated variables of a multiple linear regression analysis to predict chickpea seed yield in the irrigated (IR) and rainfed (RF) treatments at both locations (Ludhiana and Faridkot)

Entered variable		Regression coefficient (b)	SE	t-value	p-value
DG	IR RF	–0.170 –0.002	0.036 0.025	–4.74** –0.10	< .0001 0.9235
DFI	IR RF	–0.075 0.073	0.039 0.038	–1.91 1.93	0.0562 0.0535
DFF	IR RF	0.184 0.029	0.056 0.063	3.29** 0.46	0.0010 0.6470
DHF	IR RF	–0.222 –0.138	0.042 0.047	–5.30** –2.91	< .0001 0.0036
PH	IR RF	–0.030 –0.002	0.010 0.011	–3.09** –0.16	0.0020 0.8739
NPP	IR RF	–0.009 0.044	0.010 0.013	–0.96 3.32**	0.3373 0.0009
BIO	IR RF	0.393 0.333	0.010 0.011	38.97** 31.35**	< .0001  < .0001
HSW	IR RF	0.008 –0.141	0.026 0.035	0.27 –4.09**	0.7852  < .0001
HI	IR RF	0.672 0.505	0.009 0.008	78.84** 64.42**	< .0001  < .0001
MPI	IR RF	–0.204 –0.192	0.011 0.080	–18.90** –24.01**	< .0001  < .0001
RLWC	IR RF	0.010 –0.049	0.005 0.007	1.85 –6.80**	0.0653  < .0001
RL	IR RF	–0.077 –0.265	0.154 0.082	–0.50 –3.23**	0.6153 0.0013
SL	IR RF	0.084 –0.033	0.288 0.154	0.29 –0.22	0.7698 0.8291
RSR	IR RF	0.992 –1.048	2.384 0.942	0.42 –1.11	0.6774 0.2662
RLD	IR RF	0.905 1.308	0.373 0.169	2.43* 7.76**	0.0156  < .0001
FRW	IR RF	0.966 3.594	2.652 0.689	0.36 5.22**	0.7159  < .0001
FSW	IR RF	0.438 0.745	1.229 0.694	0.36 1.07	0.7214 0.2834
RDW	IR RF	–1.794 4.37	5.685 1.000	–0.32 4.37**	0.7524  < .0001
SDW	IR RF	0.569 –2.231	5.987 3.484	0.09 –0.64	0.9251 0.5222
RRDWTDW	IR RF	31.061 –24.570	25.220 8.342	1.23 –2.95**	0.2186 0.0033

DG = days to germination, DFI = days to flowering initiation, DFF = days to 50% flowering, DHF = days to 100% flowering, PH = plant height, NPP = number of pods per plant, BIO = biomass per plant, YLD = yield per plant, HSW = 100-seed weight, HI = harvest index, MPI = membrane permeability index, RLWC = relative leaf water content, RL = root length, SL = shoot length, RSR = root-shoot ratio, RLD = root length density, FRW = fresh root weight, FSW = fresh shoot weight, RDW = root dry weight, SDW = shoot dry weight, RRDWTDW = ratio of root dry weight and total plant dry weight

^*Significant at 5% probability level; **significant at 1% probability level

For root-related traits, the multiple linear regression analysis revealed that RLD, FRW, FSW, and RDW/TDW contributed positively, while RDW and SDW contributed negatively to the total variation in response traits, such as YLD. Thus, the multiple linear regression revealed that NPP, BIO, and HI of the shoot-related traits and RLD, FRW, FSW, and RDW/TDW of the root-related traits contributed significantly to the total variation in YLD in the rainfed treatment (Table 5).

The estimated results were used to formulate prediction equations for seed yield (SY as a dependent variable) in rainfed and irrigated treatment using the chickpea plant variables (as independent variables).

The regression equation for SY and other yield components in the irrigated treatment at both Ludhiana and Faridkot locations is:

$$\begin{array}{cc}\hfill \text{SY}& =4.61-0.170\text{DG}+0.\text{184DFF}-0.\text{222DHF}-0.030\text{PH}+0.\text{393BIO}+0.\text{672HI}\hfill \\ \hfill & -0.20\text{4MPI}-0.0\text{56PV},{\text{R}}^2=0.9988,\text{Adjusted}{\text{R}}^2=0.9988.\hfill \\ \hfill \end{array}$$

The regression equation for SY and yield component traits in the rainfed treatment at both Ludhiana and Faridkot locations is:

$$\begin{array}{cc}\hfill \text{SY}& =9.607+0.0\text{44NPP}+0.\text{333BIO}-0.\text{141HSW}+0.50\text{5HI}-0.\text{192MPI}\hfill \\ \hfill & -0.0\text{49RWC},{\text{R}}^2=0.9956,\text{Adjusted}{\text{R}}^2=0.9956.\hfill \\ \hfill \end{array}$$

The regression equation for SY and root-related traits for pooled condition including both irrigated and rainfed treatment is:

$$\begin{array}{cc}\hfill \text{SY}& =-9.336+1.\text{835RLD}+3.\text{591FRW}+7.\text{892FSW}-2.\text{478RDW}-35.\text{556SDW}\hfill \\ \hfill & +20.\text{377 RDW}/\text{TDW};{\text{R}}^2=0.8753,\text{Adjusted}{\text{R}}^2=0.8744.\hfill \\ \hfill \end{array}$$

Stepwise linear regression is used to identify a useful subset of predictors. For the irrigated treatment, MPI was first entered to the model and explained 77% of the total variation of YLD, followed by BIO and HI (Table S1). In the rainfed treatment, NPP significantly contributed to the variation in YLD, and explained 86% of the variation, followed by HI and BIO (Table S2).

Principal component analysis

The principal component analysis (PCA) provides information about the measured traits by elucidating the maximum variability present in the population under specific environments. The PCA analysis revealed that the first three principal components explained 80.76% and 88.25% of the total phenotypic variability in the irrigated and rainfed treatments, respectively.

The PCA revealed that PH, NPP, BIO, HSW, HI, MPI and RWC were the main contributing traits in PC1 for YLD and had strong associations with PC1 and PC2 (Fig. 3, Table S3) in both irrigated and rainfed treatments at both locations. MPI was the only major contributing trait in PC1 that occurred in a negative direction in both treatments.

Fig. 3.

Biplots based on principal component analysis (PCA) showing the relationship of yield (YLD) with agronomic traits under irrigated treatment (a) and rainfed treatment (b), likewise with root related traits under irrigated treatment (c) and rainfed treatment (d). PC1 = Principal component 1, PC2 = Principal component 2, dg = days to germination, dfi = days to flowering initiation, dff = days to 50% flowering, dhf = days to 100% flowering, ph = plant height, npp = Number of pods per plant, bio = biomass per plant, yld = yield per plant, hsw = 100-seed weight, hi = harvest index, mpi = membrane permeability index, rlwc = relative leaf water content, rl = root length, sl = shoot length, rsr = root-shoot ratio, rld = root length density, frw = fresh root weight, fsw = fresh shoot weight, rdw = root dry weight, sdw = shoot dry weight, rrdwtdw = ratio of root dry weight and total plant dry weight

For root-related traits, the PCA revealed that the first three principal components explained 87.48% and 85.94% of the total phenotypic variability in the irrigated and rainfed treatments, respectively. The main contributing traits in PC1 for YLD were RL, SL, RLD, FRW, FSW, RDW, SDW, and RDW/TDW in both the irrigated and rainfed treatments (Fig. 3, Table S4). Unexpectedly, RLD was the only major contributing trait that occurred in a negative direction in both treatments. Thus, the major yield contributing traits (except RLD) clustered together and contributed to the maximum variability for YLD in both the irrigated and rainfed treatments. Hence, phenotypic selection for these root related traits will be helpful for monitoring chickpea genotypes under drought stress.

Overall, on the basis of yield components, morphological and root traits, a total of 15 promising RILs were identified for their use in chickpea breeding programs for developing drought tolerant cultivars (Table 6). Among them, RIL 75 was the highest yielding followed by RIL 81 and RIL 154. Although, significant effect of genotype by environment interaction for various traits depicts that the stable genotypes may not show the uniform performance across different environments. The RIL population was also genotyped using ddRAD-seq (Peterson et al. 2012) and the linkage map was constructed as described previously (Kushwah et al. 2021a). Composite interval mapping identified eight consensus QTLs for six traits and five QTL clusters containing QTLs for multiple traits on linkage groups CaLG04 and CaLG06 (Table S5) (Kushwah et al. 2022).

Table 6.

List of promising recombinant inbred lines for yield and yield contributing traits under drought stress

RILs	DG	NPP	BIO	YLD	HSW	HI	MPI	RLWC	RL	FRW	RDW
7	9.33	42.99	73.74	35.82	15.65	48.75	36.88	72.56	97.31	10.05	2.71
9	9.02	40.51	69.76	36.27	15.47	52.37	36.69	75.85	90.08	9.51	2.45
12	9.11	47.05	77.85	35.94	16.58	46.51	36.88	74.52	108.08	11.72	3.93
13	9.96	45.84	77.04	36.90	15.55	48.03	37.34	74.80	114.54	12.40	4.46
15	9.39	40.31	70.91	35.59	16.83	50.47	37.81	76.97	109.93	11.69	3.81
16	10.66	43.96	74.11	37.04	14.66	50.35	34.74	76.77	114.56	12.24	4.26
26	10.69	50.07	79.78	38.66	17.09	48.95	35.21	76.53	122.07	13.10	4.70
41	9.61	46.27	73.20	38.91	16.22	53.97	34.47	75.95	105.02	11.12	3.19
56	9.93	43.44	75.33	37.11	15.24	49.38	35.51	78.02	87.41	9.41	2.36
75	8.37	50.08	81.46	45.79	15.69	56.56	31.87	78.09	127.45	13.62	5.18
77	9.06	44.35	79.20	36.16	16.95	45.79	37.01	77.58	93.47	10.28	2.61
80	8.68	43.72	73.64	35.68	13.88	49.47	36.91	75.94	100.88	10.65	3.10
81	8.26	47.98	76.98	42.68	17.05	56.01	32.58	77.59	84.88	9.31	2.17
154	9.30	49.70	83.48	42.04	16.49	50.87	31.55	79.37	106.68	11.11	3.20
180	10.44	45.95	73.38	35.13	14.03	48.16	36.62	71.25	84.66	9.21	2.20
GPF2	8.52	43.71	77.50	36.77	16.08	47.95	36.17	78.15	125.70	12.25	4.69
ILWC292	13.70	18.08	40.71	11.777	10.33	28.88	50.23	48.10	88.98	7.92	1.39

DG = days to germination, NPP = Number of pods per plant, BIO = biomass per plant (gm), YLD = yield per plant (gm), HSW = 100-seed weight (gm), HI = harvest index (%), MPI = membrane permeability index, RLWC = relative leaf water content (%), RL = root length, FRW = fresh root weight, RDW = root dry weight

Discussion

Genetic variability for various traits

Drought stress reduces chickpea yields by up to 50%, making it a significant factor for chickpea production worldwide (Kumar et al. 2015). It is imperative to study the complex nature of drought stress and identify the parameters responsible for drought tolerance for the selection of drought-tolerant genotypes. Drought stress significantly affected all of the measured morphological and physiological traits, except HI. Early flowering is a good option as it can escape drought stress but it is negatively correlated with seed yield. Hence it is necessary to develop drought tolerant high yielding chickpea genotypes having normal maturity. Genetic studies for drought tolerance in chickpea are limited due to the absence of effective selection criteria, including morphological, physiological, and biochemical responses (Sachdeva et al. 2018). Some studies showed that early flowering, phenological plasticity, and a profuse and deep root system could be beneficial under drought stress (Saxena 2003; Berger et al. 2006).

Moderate to high values of GCV, PCV, heritability, and GAM were observed for most yield-contributing traits and root-related traits under drought-stress, relative to non-stress, indicating that selection will be more effective under stress environment, as observed in other studies (Krishnamurthy et al. 2011; Paul et al. 2018). Higher GAM could be due to the presence of a high amount of additive gene action in the RIL population. Significant differences among genotypes for all the measured traits indicated that selection of genotypes with variable tolerance to drought stress is possible in the studied germplasm. Combined ANOVA in several studies also showed significant differences for various morphological and physiological traits (Hamwieh et al. 2013; Pang et al. 2017; Purushothaman et al. 2017; Sachdeva et al. 2018). Highly significant differences have also been observed for several root-related traits in pooled ANOVA (Kashiwagi et al. 2005; Varshney et al. 2014; Purushothaman et al. 2017).

Plants mature earlier under water stress, and sensitive genotypes tend to have fewer pods and seeds per plant, and lower seed weight, biomass, and harvest indices than tolerant genotypes, which is in accordance with our results (Zaman-Allah et al. 2011; Sachdeva et al. 2018). This is likely due to reduced soil water uptake and restricted transpiration, which reduces root growth. Reduced seed weights could be due to the adverse effects of drought stress on photosynthetic assimilates and the partitioning of these assimilates into seeds. Early flowering due to drought stress reduces the duration to crop maturity, which indicates that this trait is a useful selection criterion for drought tolerance (Krishnamurthy et al. 2010; Purushothaman et al. 2016). However, some previous studies observed that early flowering is negatively correlated with seed yield under drought stress (Turner et al. 2001; Kumar et al. 2005). The primary adaptive strategy identified by several chickpea breeding programs for tolerance to terminal drought stress is drought escape via early flowering (Kumar and Abbo 2001; Berger et al. 2006; Gaur et al. 2008; Purushothaman et al. 2014). The early flowering chickpea genotypes had a comparatively longer duration of reproductive growth, which is advantageous for seed set before the onset of drought stress (Krishnamurthy et al. 2013; Purushothaman et al. 2014). However, another study reported that earliness shortens the duration of reproductive growth in chickpea (Maqbool et al. 2016). Plants adapt themselves to complete their life cycle as soon as possible before the onset of drought stress and that response is called “drought escape”. Irrigation at flowering increases the duration of the vegetative stage and redirects photosynthate translocation into vegetative components, which reduces seed yield (Maqbool et al. 2015).

A profuse and deep root system could be responsible for higher seed yields under drought stress. Genotypes with a strong root system would also produce high biomass under drought stress. Several previous studies have shown that root traits, such as root length, root length density, root depth, and root dry weight could be promising for improving drought tolerance in chickpea (Kashiwagi et al. 2006b; Zaman-Allah et al. 2011; Purushothaman et al. 2017). Despite the significance of prolific root systems for drought stress, few improvements have been made in this direction, mainly because root studies are laborious and time-consuming. In the present study, drought stress caused a significant reduction in SL, RSR, FSW, and SDW the most, as reported elsewhere (Ganjeali et al. 2011). Prolific root systems are likely to influence transpiration, biomass and harvest index under drought stress through the utilization of deep soil moisture (Zaman-Allah et al. 2011; Kashiwagi et al. 2013; Purushothaman et al. 2016). Shoot and root vigor are reciprocally advantageous as the production of shoot biomass depends on the exploitation of soil moisture by the root system (Pinheiro et al. 2005; Purushothaman et al. 2017) and root vigor depends on the production of photo-assimilates by shoots (Wasson et al. 2014). Some studies have shown that root length density contributes little to chickpea yield under drought stress (Kashiwagi et al. 2006b; Zaman-Allah et al. 2011). This suggests that further improvements in root-related traits could improve drought-stress tolerance in chickpea as higher yields and harvest index can be attained with a strong root system. Further investigation of the function of these root-related traits requires precise phenotyping and genotyping (Gaur et al. 2008).

Response of shoot- and root-related traits to drought stress

The correlation coefficients indicated that YLD had a significant positive correlation with PH, NPP, BIO, HSW, HI, and RWC. However, the association of YLD was significantly negative with DG, DFI, DFF, DHF, and MPI at averaged over two locations. A significant positive association (more than 0.5) indicates that any positive increase will also accelerate seed yield. Similarly, other studies have reported positive correlations between seed yield and plant height (Anbessa et al. 2002), pod and seed numbers per plant, 100-seed weight, biomass, and harvest index (Jivani et al. 2013; Hamwieh and Imtiaz 2015) and negative correlations between seed yield and flowering time and days to maturity in chickpea (Krishnamurthy et al. 2010; Zaman-Allah et al. 2011; Varshney et al. 2014; Purushothaman et al. 2017). Drought stress during vegetative growth adversely affected shoot biomass production and grain yield; however, drought stress during reproductive growth ensures the shoot biomass production during the vegetative growth stage and eventually increases the grain yield (Fenta et al. 2012; Kashiwagi et al. 2015) by partitioning of shoot biomass into grain (Ainsworth et al. 2011; Krishnamurthy et al. 2013).

For root-related traits, the correlation coefficient analysis indicated that RL had significant and positive correlations with SL, RSR, FRW, FSW, RDW, SDW, and RDW/TDW, and a significant and negative association with RLD. The results are in accordance with other studies that exhibited significant positive correlations between root length and root dry weight, root depth, and root volume (Fageria et al. 2006; Ganjeali et al. 2005, 2011). Root length, root depth, root volume, and root dry weight decreased slightly in tolerant genotypes, but significantly decreased in susceptible genotypes under drought stress, indicating that a deep and dense root system increases drought tolerance in chickpea (Anbessa et al. 2002). Unexpectedly, RLD was not positively correlated with any of the measured root-related traits in the present study, which differs from some other studies (Kashiwagi et al. 2005; Varshney et al. 2014; Purushothaman et al. 2017). However, one study revealed that RLD did not differ between sensitive and tolerant chickpea genotypes under drought stress and had no correlation with seed yield or pod number per plant (Zaman-Allah et al. 2011). Eventually, plants with a superficial rooting system could have limited water uptake potential, reducing yield under drought stress, relative to a profuse root system (Wasson et al. 2012). Thus, the root system architecture and root-related traits play a significant role in water uptake and consequently improving the grain yield in chickpea under drought stress (Kashiwagi et al. 2005).

The present study indicated that BIO followed by HI was the major contributor to YLD in irrigated and rainfed treatments. BIO had the greatest positive indirect effect on YLD through NPP, followed by PH and HSW. Harvest index, followed by YLD, NPP, and HSW was the greatest positive direct contributor to grain yield in chickpea in several studies ( Renukadevi and Subbalakshmi 2006; Thakur and Sirohi 2009; Vaghela et al. 2009; Jivani et al. 2013). Thus, the combination of traits with the highest positive direct and indirect effects can be used as effective selection indices for drought tolerance in chickpea. Both the multiple linear regression and stepwise linear regression showed that NPP, BIO, and HI contributed the most to the total variation of YLD. While, among the root related traits, multiple linear regression explained that RLD, FRW, FSW, and RDW/TDW were the main contributors to the total variation of YLD. Regression analyses have shown that reductions in seed yield are significantly related to reductions in seed number (Zaman-Allah et al. 2011) and that the number of filled pods and total number of seeds contribute the most to the total variation of chickpea yield (Paul et al. 2018). In the present study, the PCA revealed that PH, NPP, BIO, HSW, HI, MPI, and RWC and the root-related traits, RL, SL, RLD, FRW, FSW, RDW, SDW, and RDW/TDW were the main contributors to YLD in both the irrigated and rainfed treatments. Other studies have indicated that harvest index, biological yield, and pod number per plant are the major contributors to seed yield and should be evaluated ahead of other traits to increase chickpea yield (Toker 2004; Toker and Cagirgan 2004; Canci and Toker 2009). Thus, these traits will be helpful for phenotypic selection of drought-tolerant genotypes in chickpea.

Finally, on the basis of the combined approach of all analyses used in this study, the traits such as number of pods per plant, biomass, harvest index, and root length were identified as key traits for developing high yielding drought tolerant cultivars for drought stress conditions. On the basis of yield components, morphological and root traits, a total of 15 promising RILs were found to be promising for developing drought tolerant chickpea cultivars (Table 6). Out of which, RIL 75 was the highest yielding followed by RIL 81 and RIL 154. Identified consensus QTLs and QTL clusters containing QTLs for multiple traits can be introgressed into elite cultivars using genomics-assisted breeding to enhance drought tolerance in chickpea.

Conclusions

This study illustrated the presence of large genetic variation in RILs for yield, yield components and root-related traits under drought stress. Some traits such as HI, RL, RLD, and FRW were least affected by drought stress. The correlation coefficient analysis showed that yield had a significant and positive correlation with PH, NPP, BIO, HSW, HI, and RWC and, for root-related traits, RL had a significant and positive correlation with SL, RSR, FRW, FSW, RDW, SDW, and RDW/TDW under irrigated and rainfed treatments. The path analysis and multiple and stepwise regression analyses showed that NPP, BIO, and HI were the major contributors to yield in the rainfed treatment at both locations. The principal component analysis also supports the results and revealed that PH, NPP, BIO, HSW, HI, MPI, and RWC were the major contributors to the total variation of yield in the rainfed treatment at both locations. Thus, these traits can be exploited as indirect selection criteria for improving chickpea yield under drought stress. Out of these significant traits, NPP, BIO, HI, and RL could be the most promising traits, which substantially increase the chickpea yield. A total of 15 promising RILs having greater tolerance to drought were identified for further use in chickpea breeding programs.

Supplementary Information

Below is the link to the electronic supplementary material.

Author’s contributions

AK, GS, SS, IS and SB designed and conducted the experiments. DB, AK and SV performed the data acquisition and data analysis. AK, DB, KHMS, HN and SS prepared and edited the final manuscript. All authors reviewed the manuscript critically and approved for submission.

Funding

The INSPIRE research grant provided to the AK by Department of Science and Technology (DST), New Delhi, India and research grant provided under the project ‘Consortia Research Platform on Molecular Biology' by Indian Council of Agricultural Research, New Delhi to SS for carrying out the research are highly acknowledged.

Availability of data and materials

Not applicable.

Code availability

Not applicable.

Declarations

Conflict of interest

All authors declare that there are no conflicts of interest.

Ethics approval

Not applicable.

Consent to participate

Not Applicable.

Consent for publication

Not Applicable.