Potato (Solanum tuberosum L.) is the fourth most important food crop produced after wheat, rice, and corn. Despite the limited genetic diversity of commercial varieties, potatoes are highly related species (Jansky,2006).

The most important goal in breeding programs is to increase yield, but the quantitative and polygenic aspects of this trait make it difficult to study (Heidari Roodbali et al.,2016). Little progress in the selection for tuber yield has encouraged breeders to focus on the selection of secondary traits (Landjeva et al.,2008) that have high heritability and are highly correlated with tuber yield (Sadeghi,2014). Development of potato genotypes that have drought tolerance is also necessary (Coleman,2008). The occurrence of drought stress at the end of the season will have less effect on these tolerant genotypes due to their evasion from drought stress and faster completion of the life cycle (Deshi et al.,2015). Potatoes are less sensitive to water shortage in the early vegetative period and maturity period, while water shortage during germination and tuber formation can cause serious damage to potato plants (Ayas & Korukçu,2010). Nouri et al. (2016) identified the ‘Satina’ and ‘Savalan’ cultivars, which are tolerant to mild and severe water deficit stress. Hassanpanah et al. (2016) conducted hybridization of potato cultivars with a diallel cross using the Griffing III method and selected 397045-13, 397031-16, and 397067-6 as a promising clones. These clones had uniform tubers, yellow to dark yellow skin color, yellow to light yellow flesh color, round and oval round tubers, shallow to medium eye depth, late maturity, and are used for the consumption chips and frying.

This research was performed to create genetic diversity in potato crops and assess suitable hybrids for agronomic traits that are both marketable and adaptable to the climatic conditions of Iran. Our study was conducted in laboratories and greenhouses at Zare-Ghostare-Arta Company over 3 years (2016–2018).

Material and Methods

Generating Genetic Diversity (2016)

This experiment used four cultivars: ‘Satina,’ ‘Caesar,’ ‘Savalan,’ and ‘Agria’ as parents for the genetic crossings. From each stem, two flowers were selected for crossing and all other flowers were removed. The experimental design followed a diallel cross using the Griffing III method (Griffing,1956; Pooni et al.,1984). The crosses were performed early in the morning or after the sunset when the air temperature was low. The true potato seed (TPS) produced from similar crosses were harvested together and were evaluated in 2017.

Out of 279 crosses between ‘Agria,’ ‘Satina,’ ‘Caesar,’ and ‘Savalan’ cultivars, 157 berries were harvested and 7,980 TPS were obtained from 12 breeding populations during 2016. Produced seeds included 1,322 seeds from cross-breeding ♀‘Agria’ × ‘Satina’♂ (16.57%); 485 seeds from cross-breeding ♀‘Agria’ × ‘Caesar’♂ (6.08%); 2,163 seeds from cross-breeding ♀‘Agria’ × ‘Savalan’♂ (27.11%); 1,495 seeds from cross-breeding ♀‘Satina’ × ‘Agria’♂ (18.73%); 1,224 seeds from cross-breeding ♀‘Satina’ × ‘Caesar’♂ (15.34%); 175 seeds from cross-breeding ♀‘Satina’ × ‘Savalan’♂ (2.19%); 106 seeds from cross-breeding ♀‘Satina’ × ‘Caesar’♂ (1.33%); 75 seeds from cross-breeding ♂‘Satina’ × ‘Caesar’♀ (0.94%); 101 seeds from cross-breeding ♂‘Savalan’ × ‘Caesar’♀ (1.27%); 212 seeds from cross-breeding ♀‘Savalan’ × ‘Satina’♂ (2.66%); 312 seeds from cross-breeding ♀‘Savalan’ × ‘Agria’♂ (3.91%), and 101 seeds from cross-breeding ♀‘Savalan’ × ‘Caesar’♂ (3.88%).

Evaluation of Clones in the Greenhouse (2017)

Breeding populations produced from the crosses were cultivated in plastic pots (10 × 10 cm) containing peat-moss and mineral pumice in a 1:1 volume ratio. During plant growth, irrigation and weeding were conducted on a regular basis. For pest control, Confidor poison (250 mL per 400 L water per ha) was used. For the control of fungal diseases, Mancozeb fungicide (1 g per 10 m2) was used. After harvesting, clones were evaluated for tuber number and weight per plant, eye depth, skin color, and tuber shape.

Evaluation of Clones in the Field Under Water Deficit Stress Conditions (2018)

The selected clones were cultured in a split-plot using randomized complete block design in triplicate. The main factor consisted of three levels of irrigation (100%, 80%, and 60% of field capacity) and a secondary factor of 12 populations (♀‘Agria’ × ‘Satina’♂, ♀‘Agria’ × ‘Caesar’♂, ♀‘Agria’ × ‘Savalan’♂, ♀‘Satina’ × ‘Agria’♂, ♀‘Satina’ × ‘Caesar’♂, ♀‘Satina’ × ‘Savalan’♂, ♀‘Caesar’ × ‘Satina’♂, ♀‘Caesar’ × ‘Agria’♂, ♀‘Caesar’ × ‘Savalan’♂, ♀‘Savalan’ × ‘Satina’♂, ♀‘Savalan’ × ‘Agria’♂, and ♀‘Savalan’ × ‘Caesar’♂). The rows were spaced 75 cm apart and the distance between the tubers was roughly 25 cm. The amount of water irrigation was determined at each stage of potato growth based on field capacity (FC), permanent wilting point (PWP), bulk density (Bd.D), available water (AW), and readily available water (RAW) (Rasoulzadeh & Raof,2013). The soil moisture content was 21.147% at irrigation start time. The soil moisture content during the growth stages was determined using a portable soil moisture meter (model PMS-714, Taiwan). The amount of water consumed at different stages of potato growth is shown in Table 1.

Table 1

Water consumed in different stages of potato growth provided by tape irrigation under normal conditions.

Stages of potato growthIrrigation numberWater consumed amount (m3 per ha)Effective rain amount (m3 per ha)Applied water volume* (m3 per ha)
Planting stage1159.06102261.06
Planting till tuberization3477.18339816.18
Tuberization till tuber maturity96,435.01216,556.0

[i] * Applied water volume = Water consumed amount + Effective rain amount.

[ii] The 100%, 80%, and 60% available water were 7,633.24, 6,106.59, and 4,579.94 m3 per hectare, respectively.

After calculating the water requirement for irrigation at 100%, the values were determined for 80% and 60% irrigation, respectively. Analysis of variance was performed after confirmation of the normality of the data obtained from trait measurements. Mean comparisons of measured traits were done using a least significant difference (LSD) test at a 5% probability level. Statistical computation and table construction used SAS 9.1 and Excel software. The SAS ver. 9.1 statistical software with the SAS-Griffing-Method3 program was used to estimate the effects of general and specific combining abilities and genetic components of the clones. The effects of the general combining ability (GCA) and specific combining ability (SCA) (Moghadam & Amiri-Oghan,2010), standard errors for the effects of general and specific combining (Kempthorne,1957; Vasal et al.,1992), the variance of general and specific combinations in potato (Bradshaw,2006; Thompson & Mendoza,1984), degree of dominance (Dick et al.,1988), and general and specific inheritance (Teklewold & Becker,2005) were estimated.


In this experiment, 279 crosses between ‘Agria,’ ‘Satina,’ ‘Caesar,’ and ‘Savalan’ cultivars were conducted to produce 7,980 true potato seeds from 12 breeding populations in 2016. From 2,313 seeds, we selected 244 hybrids, which had a mean tuber weight per plant of approximately 975.5 g and 7.5 tubers per plant. The results of the analysis of variance showed different levels of water deficit stress significantly affected the genotypes and their interactions with respect to tuber number and weight per plant, tuber yield, main stem number per plant, main stem diameter, and stolon length at a probability of 1% and the trait tuber dry matter percent between genotypes at a 1% probability level (Table 2).

Table 2

Results from the variance analysis of the studied traits in the potato genotypes grown under water deficit stress.

Sources of variationDfMS
Tuber number per plantTuber weight per plantTuber yieldMain stem number per plantMain stem diameterDry matter percentStolon length
Replication20.18 ns44.4444 ns12.48 ns1.00 ns1.56 ns1.07 ns1.00 ns
Water deficit stress (A)210.63**1,283,416.08**3,604.33**16.00**100.77**3.034 ns4.088**
Error 162.0314,444.4412.481.000.861.0671.00
Genotypes (B)113.113**83,877.72**235.55**2.909**13.24**9.994**12.28**
A × B221.15**19,796.26**55.59**1.18**6.99**0.571 ns2.82**
Error 2660.50549,141.4125.670.4540.8681.0680.696

[i] Note: ns and ** indicate nonsignificant and significant differences at a 1% probability level, respectively.

The highest tuber yields were observed in normal growth conditions (100% available water), in clones related to crosses of ♀‘Satina’ × ‘Agria’♂, ♀‘Agria’ × ‘Caesar’♂, ♀‘Agria’ × ‘Savalan’♂, ♀‘Satina’ × ‘Agria’♂, ♀‘Caesar’ × ‘Agria’♂, ♀‘Caesar’ × ‘Savalan’♂, ♀‘Savalan’ × ‘Satina’♂, ♀‘Savalan’ × ‘Agria’♂, and ♀‘Savalan’ × ‘Caesar’♂. In mild stress conditions, (80% of available water) clones related to ♀‘Agria’ × ‘Caesar’♂ and ♀‘Caesar’ × ‘Savalan’♂ produced the highest tuber yields, and in severe stress conditions (60% available water), clones related to ♀‘Agria’ × ‘Caesar’♂ produced the most tubers compared to the other clones tested. The highest tuber yield was reported in a clones of ♀‘Agria’ × ‘Caesar’♂ and ♀‘Caesar’ × ‘Savalan’♂ under normal, mild, and severe stress conditions.

The results of analysis of variance showed that the environmental condition, genotype, and their interactions had a significant effect on tuber yield at a 1% probability level. This result indicates the presence of sufficient genetic diversity in tuber yield between the crosses. The results of the variance analysis of general combining ability (GAC) and specific combining ability (SAC), and the reverse effects of tuber yield showed that in all three environmental conditions there was a significant difference at a probability level of 5% and 1% (Table 3).

Table 3

Analysis variance of general and specific combining ability of potato tuber yield in genotypes under three environmental conditions.

Sources of variationDfMS
Normal condition (100% available water)Mild stress condition (80% available water)Severe stress condition (60% available water)

[i] Significant differences at a 5% (*) and 1% (**) probability level.

Significant differences between clones indicated that there was also a significant difference between the general combining ability of the clones and the role of gene additive effects. The most significantly positive GCA effects were for tuber yield under normal growth conditions in ‘Agria’ and ‘Savalan’ cultivars, while under mild and severe stress conditions, these effects were most significantly positive in ‘Agria’ and ‘Caesar’ cultivars (Table 4). Therefore, these cultivars can easily transfer this trait to their offspring and increase the presence of this trait in future generations. The significance of the general combining ability effect indicates a role for additive effects in the control of trait expression.

According to the estimated GCA effects among the cultivars, the highest positive GCA for tuber yield was observed in ‘Agria’ and ‘Caesar’ cultivars under mild and severe water stress conditions and the lowest in ‘Satina’ and ‘Savalan’ cultivars (Table 4). In this study, two parents included ‘Agria’ and ‘Caesar’ cultivars with positive and significant GCA and two parents included ‘Satina’ and ‘Savalan’ cultivars with negative and significant GCA were.

In normal growth conditions, the highest SCA in the direct crosses for tuber yield was significantly positive in compounds of ♀‘Agria’ × ‘Caesar’♂ and significantly negative in compounds of ♀‘Agria’ × ‘Savalan’♂. In the reverse crosses, the SCA was significantly positive in compounds of ♀‘Satina’ × ‘Caesar’♂ and ♀‘Satina’ × ‘Savalan’♂ (Table 4).

In mild stress growth conditions, the highest SCA in the direct cross for tuber yield was in compounds of ♀‘Agria’ × ‘Caesar’♂ in the positive direction and ♀‘Agria’ × ‘Savalan’♂ in the negative direction. In the reverse cross, compounds of ♀‘Caesar’ × ‘Agria’♂ in the negative direction and ♀‘Savalan’ × ‘Caesar’♂ in the positive direction had the highest SCA for tuber yield (Table 4).

Table 4

Estimation of general combining ability (GCA), specific combine ability (SCA), and inverse (REC) potato tuber yield in normal, mild, and severe water stress conditions.

Mild stress2.0875**
Severe stress2.838**
Mild stress−5.7875**
Severe stress−5.213**
Mild stress4.8125**
Severe stress4.463**
Mild stress−1.1125**
Severe stress−2.08**
Mild stress−0.15865.217**−166.567**
Severe stress0.00044.48**−5.78**
Mild stress−0.583−0.158
Severe stress−3.33−2.20
Mild stress−0.158
Severe stress0.00
Mild stress
Severe stress
Mild stress
Severe stress
Mild stress−2.350
Severe stress0.150
Mild stress−5.450**1.550
Severe stress2.650−2.750
Mild stress2.0004.650**−5.050**
Severe stress4.950**1.250−0.700

[i] Standard error (SE) (GCA) = 0.317; critical difference (CD) = SE × t0.05 = 0.808; CD = SE × t0.01 = 0.994.

[ii] SE (SCA)= 1.423; CD = SE × t0.05 = 3.730; CD = SE × t0.01 = 4.593.

[iii] SE (REC)= 1.267; CD = SE × t0.05 = 3.231; CD = SE × t0.01 = 3.978.

[iv] Significant differences at a 5% (*) and 1% (**) probability level.

In severe water stress conditions, the highest SCA observed in the direct cross for tuber yield was in compounds of ♀‘Agria’ × ‘Caesar’♂ in the positive direction and ♀‘Agria’ × ‘Savalan’♀ in the negative direction; while in the reverse cross, compounds of ♀‘Savalan’ × ‘Agria’♂ in the positive direction had the highest SCA (Table 4). These crosses can, therefore, be used to produce hybrids with increased tuber yield under drought stress. In this study, the ‘Agria’ and ‘Caesar’ cultivars had a good composition for tuber yield that resulted in Additive × Additive and Additive × Dominance interactions when used as parents.

The direct cross of ♀‘Agria’ × ‘Caesar’♂ had the highest tuber yield. Among the studied genotypes, when ‘Agria’ and ‘Caesar’ cultivars were grown under mild and severe stress conditions, they had a significantly positive general combining ability for tuber yield, indicating that they are suitable parents for breeding programs. Indeed, the use of cultivars with significantly positive specific combining for tuber yield is emphasized. For example, crossing ‘Agria’ and ‘Caesar’ cultivars with ‘Satina’ and ‘Savalan’ cultivars had a significantly positive specific combining ability. Therefore, selection among the offspring resulting from these crosses, in addition to increased additive effects of genes, also increased the genetic efficiency of the selection (Table 4).

In Table 5, the general and specific heritability values are presented. The general heredity of tuber yield in 100%, 80%, and 60% available water was 93.88%, 93.32%, and 87.07%, respectively (Table 5). The specific heredity of tuber yield in 100%, 80%, and 60% available water was 71.94%, 80.16%, and 64.36%, respectively (Table 5). The dominance degree for tuber yield in these growth conditions was positive (Table 5). The positive sign indicates incomplete dominance in order to increase the relevant trait. The ratio of the variance of general combining ability to the variance of the specific combining ability for tuber yield in in 100%, 80%, and 60% available water was estimated to be more than one (Table 5).

Table 5

Genetic components of tuber yield in genotypes grown under three environmental stress conditions.

Combining ability*In 100% available waterIn 80% available waterIn 60% available water
δ2a = 2δ2g119.331687.8415846.25158
δ2d = δ2s12.0011731.8211711.40117
Degree of dominance0.3170.6010.496
General heritability93.8893.3287.07
Specific heritability71.9480.1664.36

[i] * δ2a – dditive variance; δ2g – variance of general combining ability; δ2d – dominance variance; δ2s – variance of the specific combining ability.


The importance of specific combining ability effects is the presence of a nonadditive effect on the control of traits. Both general and specific combining ability effects are important in controlling traits and additive and nonadditive effects. Therefore, in these cultivars, the frequency of genes with additive effects was high, which indicated that they can be used for breeding programs based on their selection (Ahmad et al.,2009). When the general combining ability (GCA) was negative for tuber yield in ‘Satina’ and ‘Savalan’ cultivars in mild and severe stress conditions, they could not efficiently transfer this trait to offspring. The cultivars that have a significant GCA for specific traits could easily transfer the desired traits to their offspring; in other words, these are suitable cultivars for breeding programs for the desired traits (Fathi et al.,2007). Therefore, depending on the direction and purpose of the breeding program, the parents could be selected (Dehghanpour,2013). Finally, based on our results, ‘Agria’ and ‘Caesar’ cultivars are the best combination for improving tuber yield in mild and severe water stress conditions, indicating that these parents may be used to improve different cross-combinations. Accordingly, ‘Agria’ and ‘Caesar’ varieties with a positive GCA and an ability to transfer the improved tuber yield trait to future generations can participate as one of the parents in breeding programs for tuber yield optimization. However, ‘Satina’ and ‘Savalan’ cultivars had negative GCA for tuber yield and should not be considered for breeding programs aimed to improve tuber yields. Given that GCA is based on additive gene effects, lines that have a relatively high GCA will also have high additive gene effects and can be used to produce synthetic varieties of potatoes (Sharma & Madhu,2013). These results are consistent with reports by Sharma & Madhu (2013) and Gopal (1996). Rowel et al. (1986) showed that for the production of hybrid plants with high SCA, parents with high GCA are not needed, but parental composition with low GCA usually produce hybrids with better SCA. Akhtar and Chowdhry (2006) reported that combining ability and gene effects have importance roles in breeding of genetic materials. Selection of a breeding method that exploits the genetic potential of different crop characteristics of a plant depends on the type of genes controlling a trait and their inheritance. The estimation of general and specific combining ability helps breeders determine what the breeding program and genotype selection strategies they use (de la Vega & Chapman,2006). Rather et al. (2007) conducted growth experiments in more than two environments and reported that variance due to the general and specific combining abilities indicates an importance for additive and nonadditive effects of inherited genes in all studied traits. Sanford (1960). further concluded that specific combining ability was relatively more important than the general combining ability for tuber yield. However, Mohammed (2009) reported that the effects of general combining ability are more important than those of specific combining ability in the grain yield traits in sorghum. The specific combining ability expresses the nonadditive contribution of genetic variance (Hosana et al.,2015), and the ratio of GCA to SCA + GCA can be used as a criterion for determining the importance of additive genetic effects (Baker,1978). If this ratio is close to one, it indicates the importance of additive effects in determining a trait (Fan et al.,2008).

The general heritability of the tuber yield in all three water stress environments ranged from 83% to 94%. Estimation of general heritability of traits indicates more of a role for genetic effects in controlling these traits and suggests that the additive effect of genes is approximately equal to the nonadditive effect on genetic control of all traits studied (Golparvar et al.,2004). The specific heritability of tuber yield in all three environments ranged from 64% to 80%. The low specific heritability is shown in the genetic control of the studied traits, while the role of nonadditive effects of genes (i.e., dominance and epistemic) had more of an impact than the additive effect of the genes. Yield is a quantitative trait that is controlled by a number of genes and therefore, selection based on yield is not useful for its improvement (Richards,1996). Morphological and phenological traits can be easily and precisely measured and have relatively high heritability, so selection based on these traits is a consistent and fast way to improve tuber yield (Yap & Harvey,1972). If the degree of dominance is more than one, it indicates the action of the genes controlling this trait is super dominant or false dominant (Sharifi et al.,2010).


In this study, we produced 279 crosses between ‘Agria,’ ‘Satina,’ ‘Caesar,’ and ‘Savalan’ cultivars and collected 7,980 true potato seeds from 12 breeding populations. Of these, 2,313 seeds were used to select 244 hybrids. The selected clones had a mean tuber weight per plant of about 975.5 g and 7.5 tubers per plant. In all three environmental conditions, the hybrids that produced the highest tuber yield and tuber weight per plant were related to ♀‘Agria’ × ‘Caesar’♂. The most significantly positive general combining ability for tuber yield in mild and severe water stress conditions were the ‘Agria’ and ‘Caesar’ cultivars. Therefore, these cultivars can easily transfer this trait to offspring and increase this trait within the crop population. The specific combining ability for tuber yield in normal, mild and severe stress conditions were significantly positive in the direct cross for the compounds of ♀‘Agria’ × ‘Caesar’♂, and in the reverse cross for the compounds of ♀‘Satina’ × ‘Caesar’♂, and ♀‘Satina’ × ‘Savalan’♂ in under normal conditions. Under mild stress conditions, compounds of ♀‘Savalan’ × ‘Caesar’♂ had significantly positive specific combining ability, while in severe stress conditions, the compounds of ♀‘Savalan’ × ‘Agria’♂ in the positive direction had the highest tuber yield. These crosses are suitable for producing clones with high tuber yield and drought stress tolerance.

Handling Editor

Joanna Kaczmarek; Institute of Plant Genetics, Polish Academy of Sciences, Poland

Authors’ Contributions

GRD: carried out all experiments, collected the data and wrote the first draft of manuscript; MRH: supervised and designed the research; DH: supervised the research, helped in the experiments, and wrote the final draft of manuscript; AM: analyzed and interpreted the data

Competing Interests

No competing interests have been declared.