Breeding protein-rich pea for semi-arid regions: insights from combining ability and heterosis

Abstract

Pea (Pisum sativum L.) is a protein-rich legume cultivated in semi-arid regions, where climatic variability constrains productivity. This study assessed combining ability, heterosis, and genetic parameters for yield- and protein-related traits under managed irrigation. Five cultivars (Serge, Boogie, Compana, Jofs, Ultrillo) were crossed in a full diallel, and F₁-F₂ generations were evaluated in Konya, Türkiye. Significant variation occurred for all traits. The F₁ population showed higher SCA effects and heterosis, particularly for grain and protein yield, whereas F₂ exhibited higher GCA effects, Baker ratios, and narrow-sense heritability. Because generations were evaluated in different years, these differences reflect both genetic segregation and contrasting environmental conditions. Jofs × Compana, Ultrillo × Serge, and Serge × Jofs performed consistently well. These superior combinations represent promising breeding materials for improving pea performance under managed irrigation in semi-arid environments, although their stability and breeding value should be validated across multiple environments.

Keywords:
Pea; diallel cross; combining ability; heterosis; semi-arid region

INTRODUCTION

Pea (Pisum sativum L.) is an important legume crop valued for its contribution to human nutrition due to its high protein content and digestibility when consumed both fresh and in dried forms. In addition, its nitrogen-fixing ability enables sustainable agricultural systems by improving soil fertility (Karadaş and Ceyhan 2023). High-yield and high-quality cultivars are being developed in response to increasing consumer demand, changing market conditions, and climate-related stresses. In arid and semi-arid regions, it is essential to cultivate stress-tolerant, lodging-resistant, and mechanically harvestable genotypes for sustainable production (Ceyhan and Karadaş 2023). In processing industries, especially in canning, seed size is an important market and processing-quality attribute, making large-seeded genotypes desirable. Therefore, 100-seed weight is frequently used as an indirect indicator of processing suitability in breeding programs.

The success of breeding programs relies on effectively utilizing genetic variation. Therefore, combining ability analyses are essential tools, as they reveal parental contributions and hybrid potential. General combining ability (GCA) measures additive gene effects, while specific combining ability (SCA) assesses non-additive gene effects. Research on peas has shown that non-additive gene effects play a dominant role in controlling traits such as seed yield, 100-seed weight, and seeds per pod (Sharma et al. 2023), highlighting the importance of heterosis breeding. F₁ hybrids often surpass their parents in yield, pod number, and plant height, with some combinations exceeding 200% of the parental heterosis values (Ceyhan et al. 2008). Despite their adaptation and commercial relevance, the cultivars Serge, Boogie, Compana, Jofs, and Utrillo have not previously been evaluated together in a comprehensive diallel system under semi-arid conditions. Consequently, information on their combining abilities, heterotic responses, and genetic control of yield and protein-related traits remains limited. This knowledge gap impedes the efficient identification of parental combinations suitable for developing high-yielding, protein-rich cultivars adapted to semi-arid environments.

A high yield is essential in pea breeding, but yield stability is equally important. The genotype × environment interaction directly affects agricultural performance, particularly in arid and semi-arid regions characterized by high environmental variability. Therefore, developing genotypes that remain stable across diverse environments ensures reliability for both producers and industry (Sharma et al. 2023). Principal component analysis (PCA) is an effective multivariate tool for evaluating relationships among genotypes and traits, as well as for identifying the major sources of variation within breeding populations (Matias et al. 2020, Pratap et al. 2021). These analyses help identify productive genotypes under both optimal and stressful conditions, thereby improving selection accuracy.

Genetic diversity plays a key role in pea breeding, as crosses between genetically distinct parents can generate a vast array of progeny populations (Souza et al. 2023, Gupta et al. 2024, Môro and Souza Júnior 2025, Ramalho et al. 2025). Moreover, traits with high heritability are especially valuable because they are less affected by environmental conditions, improving selection efficiency (Yadav et al. 2023). Recent legume studies indicate that protein concentration is generally associated with moderate-to-high additive genetic control and heritability, whereas protein yield is a more complex trait influenced by both protein concentration and grain yield. Consequently, protein yield may involve substantial non-additive gene action in addition to additive effects. Thus, understanding the relative importance of these genetic components is critical for designing effective breeding strategies targeting protein productivity.

Breeding new pea varieties that bring together high yield, coarse-grain structure, quality, and stress tolerance fosters the development of new cultivars that are significantly more valuable to both producers and industry. Integrating heterosis and combining ability analyses with classical selection methods, particularly in semi-arid regions, underpins successful and sustainable breeding strategies. This study aims to identify superior parental combinations and breeding strategies for developing high-yielding, protein-rich pea lines adapted to semi-arid environments. We hypothesized that (i) grain yield and protein yield would be predominantly controlled by non-additive gene effects, resulting in high SCA effects and heterosis in the F₁ generation; (ii) additive genetic effects and GCA would become more important after segregation in the F₂ generation; and (iii) genetically diverse parental combinations would exhibit superior heterotic performance for grain and protein yield under semi-arid conditions.

MATERIAL AND METHODS

The study was conducted under managed irrigation in a semi-arid agro-ecological environment during 2023-2024 at Selçuk University, Konya, Türkiye (lat 37° 52′ N, long 32° 30′ E, and alt 1016 m asl), to identify superior parents and hybrid combinations in pea. Climatic data showed notable interannual variability, with lower temperature and precipitation in 2023 (10.8 °C; 256.4 mm), which limited plant growth, while the 2024 conditions (11.9 °C; 341.2 mm) were closer to the long-term average and more favorable overall. The experimental soil was clay-loamy, alkaline, highly calcareous, low in organic matter, and deficient in phosphorus and zinc, typical of the calcareous soil conditions of the region (Ceyhan and Karadaş 2023, Karadaş and Ceyhan 2023).

The parental varieties used were the registered pea cultivars Serge, Boogie, Compana, Jofs, and Utrillo due to their high yield, large seed size, disease resistance, and adaptability. A full diallel crossing design was applied to utilize all the possible genetic variation (Ceyhan and Kahraman 2013). Crosses were conducted in a climate-controlled greenhouse with successive sowings to achieve synchronized flowering. The flowers were emasculated to prevent self-pollination, followed by manual pollination with microforceps, and covering the flowers to prevent contamination (Ceyhan et al. 2008). Each cross was labeled, and the greenhouses were maintained at 25 ± 2 °C with 50-60% humidity. At least 30 F₁ seeds were obtained from the 25 combinations.

The F₁ population was grown in 2023, followed by the F₂ population in 2024, using a Randomized block design with three replicates under field conditions. Planting was done in 2-meter-long rows with either 50 cm row spacing or 20 cm inter-row spacing. In the F₁ population, parents were placed in the side rows, and the F₁ individuals were placed in the middle row. In the F₂ population, each plot was arranged in three rows. In both years, 150 kg ha^-1 of 18-46 DAP fertilizer was applied during planting, and hoeing was carried out two to three times for weed control. The fields were irrigated using a sprinkler system whenever soil moisture in the 0-60 cm root zone declined below approximately 50% of field capacity. In total, three irrigations were applied during the 2023 growing season and four during the 2024 season. Each irrigation supplied approximately 40-50 mm of water, resulting in seasonal irrigation totals of approximately 120 mm and 150 mm in 2023 and 2024, respectively. Consequently, the total seasonal water availability was approximately 376.4 mm in 2023 (256.4 mm rainfall + 120 mm irrigation) and 491.2 mm in 2024 (341.2 mm rainfall + 150 mm irrigation). The irrigation regime was designed to reflect standard agronomic practices for pea production in semi-arid regions rather than to impose controlled drought stress. Thus, the experimental conditions were representative of commercial production systems in the region. Therefore, the results should be interpreted as representing pea performance under managed irrigation in a semi-arid agro-ecological zone rather than under rainfed drought-stress conditions. Harvesting was carried out in the first week of July in both years, when 90% of the plants had yellowed and dried (Ceyhan et al. 2008). Because the F₁ and F₂ generations were evaluated in different growing seasons (2023 and 2024), generation and year effects are inherently confounded in this experimental design. Consequently, the effects of genetic segregation and environmental variation cannot be statistically separated and should be considered when interpreting comparisons between generations.

Plant height (cm) was measured at the time of maturity from the soil surface to the tip of the main stem. The total mature pods on each plant were counted to determine pods per plant. The average number of seeds from mature pods was used to calculate the number of seeds per pod, and the total seed count per plant was used to determine the number of seeds per plant. Grain yield per plant (g) was recorded after harvesting and threshing individual plants, while the hundred-seed weight (g) was determined by weighing 100 randomly selected seeds. The protein content (%) was determined by the Kjeldahl method as per AOAC Official Method 979.09 (AOAC 2023). Nitrogen was converted to crude protein using the standard conversion factor of 6.25 (protein equals nitrogen multiplied by 6.25). Protein yield per plant (g) was determined by multiplying grain yield per plant by protein content.

ANOVA was conducted using a randomized complete block design, followed by diallel analysis based on Griffing’s Model I, Method 1, including parents and reciprocals (Griffing 1956). The estimation of variance components was done in the following manner:

Additive variance:

σ²A = 2σ²GCA

${\sigma }^{2}A=2{\sigma }^{2}GCA$

Variance of dominance:

${\sigma }^{2}D={\sigma }^{2}SCA$

Environmental variance:

${\sigma }^{2}E=MSE$

where ${\sigma }^{2}GCA$ is the general combining ability variance, ${\sigma }^{2}SCA$ is the specific combining ability variance, and MSE is the residual error variance obtained from the analysis of variance.

Narrow-sense heritability was estimated as:

$h^2={\sigma }^{2}A/({\sigma }^{2}A+{\sigma }^{2}D+{\sigma }^{2}E)$

Broad-sense heritability was estimated as:

$H^2=({\sigma }^{2}A+{\sigma }^{2}D)/({\sigma }^{2}A+{\sigma }^{2}D+{\sigma }^{2}E)$

The Baker ratio was estimated as:

$BakerRatio=\left(2{\sigma }^{2}GCA\right)/(2{\sigma }^{2}GCA+{\sigma }^{2}SCA)$

Mid-parent heterosis (MPH) was calculated according to the following formulas:

$MPH\left(\%\right)=\left[(F_1-MP)/MP\right]\times 100$

where,

$MP=(P_1+P_2)/2$

The significance of heterosis values was evaluated using t-tests based on the mean squared error from the analysis of variance, following standard procedures for diallel studies. Statistical analyses were conducted using the AGD-R package (Rodríguez et al. 2015).

Principal component analysis was conducted in R (R Core Team 2024) to assess relationships among genotypes and measured traits, as well as to identify the primary factors driving variation in the F₁ and F₂ populations. Prior to the analysis, trait data were standardized to a zero mean and unit variance to eliminate scale effects among the variables. Principal components with eigenvalues greater than 1.0 were retained and used for interpretation. PCA biplots were generated using genotype means to visualize the associations among traits and hybrid combinations.

RESULTS AND DISCUSSION

The diallel analysis revealed significant genetic variation for all the traits studied (Table 1). High and significant GCA variances for plant height, seeds per pod, 100-seed weight, and protein content indicated that additive gene action was important in the inheritance of these traits. In contrast, grain yield, pod number, and protein yield were more influenced by dominant and epistatic effects, highlighting their suitability for heterosis breeding. Significant reciprocal and maternal effects in some traits also emphasized the role of parental and cytoplasmic inheritance. Strong heterosis in F₁ and increased additive effects in F₂ suggest that hybrid vigor is more pronounced in early generations, whereas stable genetic improvements become more evident in later generations.

Table 1
Mean squares for traits examined in the full diallel analysis and variance analyses of combination abilities

Sources of variation	df	F1 Generation	F2 Generation	F1 Generation	F2 Generation
		Plant height		Number of pods per plant
Replication	2	59.453	67.480	1.213	2.773
Hybrids	24	444.447**	836.436**	12.861**	8.913**
GCA	4	808.410**	2694.327**	7.200	9.377**
SCA	10	313.275**	494.715**	6.270	6.358**
Reciprocal	10	430.033**	435.000**	21.717**	11.283**
Maternal	4	635.083**	96.433	31.933**	3.883
No Maternal	6	293.333*	660.711**	14.906**	16.217**
Error	48	97.662	85.786	3.449	1.523
Sources of variation	df	Number of seeds per pod		Number of seeds per plant
Replication	2	3.253	0.333	2226.013	609.760
Hybrids	24	3.174**	4.630**	3281.308**	2807.274**
GCA	4	4.043**	17.593**	3720.910**	8891.460**
SCA	10	1.668	2.075	1270.808	1574.258*
Reciprocal	10	4.333**	2.000	5115.967**	1606.617*
Maternal	4	6.717**	2.800	8463.217**	1872.467*
No Maternal	6	2.744*	1.467	2884.467**	1429.383
Error	48	0.948	1.292	829.291	637.579
Sources of variation	df	Grain yield per plant		Hundred seed weight
Replication	2	148.094	32.557	3.060	3.552
Hybrids	24	337.374**	263.813**	55.335**	77.959**
GCA	4	308.928**	1193.504**	91.111**	263.177**
SCA	10	340.634**	45.145	73.080**	46.128**
Reciprocal	10	345.492**	110.605**	23.280**	35.703**
Maternal	4	612.509**	122.176*	33.156**	18.312**
No Maternal	6	167.481**	102.890*	16.697**	47.297**
Error	48	49.846	37.814	4.796	2.654
Sources of variation	df	Protein content		Protein yield per plant
Replication	2	0.027	0.026	7.043	1.705
Hybrids	24	4.264**	4.260**	13.720**	14.107**
GCA	4	7.446**	7.440**	14.451**	62.263**
SCA	10	1.654**	1.651**	13.407**	2.689
Reciprocal	10	5.601**	5.597**	13.741**	6.263**
Maternal	4	6.414**	6.411**	22.969**	6.533*
No Maternal	6	5.059**	5.052**	7.589*	6.082*
Error	48	0.030	0.029	2.673	1.903

• * Significant at the 5% level; ** Significant at the 1% level

The full diallel analysis revealed significant differences among the parental, F₁, and F₂ generations (Table 2). In F₁, Jofs × Compana and Ultrillo × Serge produced the highest grain yield, while Ultrillo × Boggie showed superior seed size. Jofs × Compana and Boggie × Jofs were prominent for seed number and protein yield, whereas Serge × Jofs had the highest protein content. In F₂, segregation increased variation, with Serge × Jofs, Ultrillo × Jofs, and Jofs × Ultrillo achieving the highest grain yield, while Ultrillo × Boggie maintained its superior seed size. Overall, Jofs × Compana, Boggie × Jofs, Ultrillo × Serge, and Serge × Jofs consistently combined high yield, seed size, and protein performance, indicating strong heterosis and high breeding potential. These crosses will be valuable for developing heterotic hybrids in early generations and selecting superior lines for advanced breeding programs (Ceyhan et al. 2008, Gupta et al. 2024).

Table 2
Average values for traits examined in the full diallel cross set

Parents and hybrids	Plant height (cm)	Number of pods per plant	Number of seeds per pod	Number of seeds per plant	Grain yield per plant (g)	Hundred seed weight (g)	Protein content (%)	Protein yield per plant (g)
F1 Generation
Boggie	54.33gh	24.00b-e	5.67cde	136.00c-f	33.39g-h	20.80def	23.14abc	7.57d-g
Compana	90.67ab	24.67b-e	6.67a-d	164.67a-f	34.19fgh	18.03efg	20.11i	6.88efg
Serge	86.33a-d	24.33b-e	5.33cde	129.67c-f	27.73g-h	17.67efg	21.92ef	6.00f-g
Jofs	54.00gh	23.33b-e	7.33abc	171.33a-d	54.26ab	29.33ab	21.72ef	11.79ab
Ultrillo	70.67a-h	22.67cde	4.67de	107.33e-f	29.00g-h	22.33cde	22.09e	6.40f-g
Boggie x Compana	65.00d-h	25.00b-e	7.33abc	183.67abc	37.37d-h	17.67efg	19.83ij	7.34efg
Boggie x Serge	78.00a-f	26.33bc	6.33a-e	166.67a-f	38.06c-h	20.30d-g	23.23ab	8.84a-f
Boggie x Jofs	75.00a-f	26.00bcd	8.33a	216.33a	49.88a-e	21.00def	22.52d	11.23abc
Boggie X Ultrillo	67.33c-h	23.00cde	7.33abc	168.67a-e	52.56a-d	26.33bc	23.23ab	11.44abc
Compana X Boggie	67.33c-h	24.33b-e	4.33e	105.67ef	22.74h	17.30f-g	21.91e-f	4.98g
Compana x Serge	92.00a	24.00b-e	6.67a-d	160.00a-f	31.66g-h	16.57f-g	21.99e	6.81efg
Compana x Jofs	83.33a-d	22.00de	6.33a-e	139.67b-f	41.01b-g	26.33bc	19.51j	8.00c-g
Compana x Ultrillo	53.33h	22.67cde	5.33cde	121.67c-f	30.68g-h	21.00def	22.87bcd	6.87efg
Serge x Boggie	69.67b-h	23.67b-e	6.33a-e	150.00b-f	31.49g-h	18.03efg	23.22ab	7.31efg
Serge x Compana	59.00fgh	22.67cde	6.33a-e	143.67b-f	30.67g-h	18.23efg	22.06e	6.77efg
Serge x Jofs	82.00a-e	22.67cde	6.33a-e	145.33b-f	29.20g-h	16.00g	23.27a	6.48f-g
Serge x Ultrillo	59.00fgh	22.33cde	4.67de	104.67f	24.75h	18.33efg	23.26a	5.76f-g
Jofs x Boggie	54.33gh	24.00b-e	5.67cde	136.00c-f	32.72g-h	20.80def	21.57f	7.06efg
Jofs x Compana	85.00a-d	31.33a	7.00abc	217.67a	59.12a	24.67bcd	21.10g	12.20a
Jofs x Serge	76.67a-f	25.00b-e	8.00ab	201.67ab	49.14a-f	22.33cde	20.72h	10.18a-e
Jofs x Ultrillo	73.33a-h	24.67b-e	5.67cde	140.67b-f	35.05e-h	20.33dg	21.10g	7.40efg
Ultrillo x Boggie	61.67e-h	23.00cde	6.00b-e	137.67c-f	50.14a-e	33.33a	20.12i	10.08a-e
Ultrillo x Compana	71.67a-h	23.33b-e	6.00b-e	140.67b-f	40.75b-g	25.00bcd	20.76gh	8.32b-g
Ultrillo x Serge	88.33abc	27.33ab	7.33abc	202.67ab	53.09abc	24.00cd	20.78gh	11.02a-d
Ultrillo x Jofs	77.33a-f	21.00e	5.33cde	111.67def	24.34h	17.67efg	22.77cd	5.57f-g
F2 Generation
Boggie	54.33hij	19.33cd	5.67a-g	109.33b-e	27.23d-k	20.83d-g	24.11abc	6.56d-i
Compana	66.67e-j	18.33cde	5.00c-g	91.67cde	19.34h-k	16.20h-i	21.08i	4.08hij
Serge	114.67a	18.00cde	3.33g	60.00e	14.57k	16.87h-i	22.89ef	3.33j
Jofs	66.67e-j	18.33cde	7.33abc	134.00bc	40.39a-d	27.33b	22.69ef	9.16a-d
Ultrillo	71.67c-h	19.33cd	6.67a-e	129.33bc	33.28a-g	22.33cde	23.06e	7.67b-g
Boggie x Compana	51.33ij	17.67cde	5.00c-g	88.33cde	17.45j-k	14.70ı	20.80ıj	3.63ij
Boggie x Serge	91.33bc	19.67bc	4.33efg	85.67cde	21.10g-k	19.40e-h	24.20ab	5.10g-j
Boggie x Jofs	76.67c-f	18.67cde	7.33abc	137.00bc	33.64a-g	21.33def	23.49d	7.90b-g
Boggie X Ultrillo	51.00ij	17.33cde	6.00a-f	104.00b-e	29.85c-j	24.33bcd	24.20ab	7.23c-g
Compana X Boggie	66.00e-j	19.67bc	5.33b-g	105.00b-e	22.80f-k	17.50ghi	22.88ef	5.22g-j
Compana x Serge	70.67d-i	17.67cde	4.67d-g	82.33cde	18.12ıjk	16.60h-i	22.96e	4.16hij
Compana x Jofs	63.67f-j	18.00cde	6.67a-e	120.00bcd	30.31b-j	21.33def	20.48j	6.21d-j
Compana x Ultrillo	55.00g-j	16.33e	5.00c-g	82.00cde	25.24e-k	25.33bc	23.84bcd	6.02e-j
Serge x Boggie	70.33d-i	18.33cde	6.00a-f	109.67b-e	24.10e-k	17.93f-i	24.19ab	5.83e-j
Serge x Compana	46.67j	16.33e	4.00f-g	65.67d-e	16.09k	17.77f-i	23.03e	3.71ij
Serge x Jofs	105.33ab	24.33a	8.00a	198.00a	46.03a	21.00d-g	24.24a	11.15a
Serge x Ultrillo	84.67cde	16.67de	5.33b-g	88.67cde	24.62e-k	22.67cde	24.23a	7.77b-g
Jofs x Boggie	69.33d-i	19.67bc	7.67ab	151.67ab	36.72a-e	21.33def	22.54f	8.27a-f
Jofs x Compana	86.00b-e	22.33ab	5.67a-g	126.33bc	36.03a-f	25.00bc	22.07g	7.95b-g
Jofs x Serge	88.33bcd	18.33cde	7.00a-d	127.67bc	25.84e-k	16.67h-i	21.69h	5.60f-j
Jofs x Ultrillo	77.67c-f	19.67bc	6.00a-f	118.67bcd	43.77ab	33.00a	22.07g	9.67abc
Ultrillo x Boggie	77.67c-f	18.33cde	5.67a-g	104.33b-e	41.93abc	36.00a	21.09i	8.85a-e
Ultrillo x Compana	55.00g-j	18.67cde	6.00a-f	112.33b-e	31.11b-i	23.67cd	21.73gh	6.76c-h
Ultrillo x Serge	75.00c-g	19.00cde	7.00a-d	132.67bc	32.62a-h	21.33def	21.75gh	7.09c-h
Ultrillo x Jofs	82.33c-f	19.67bc	8.00a	157.67ab	44.84a	25.67bc	23.74cd	10.61ab

Our genetic analysis (Table 3) revealed differences in the relative contributions of additive and non-additive genetic effects among the traits and between the F₁ and F₂ populations. In the F₁ generation, grain yield and protein yield exhibited relatively higher SCA variances and stronger heterotic responses, whereas the F₂ generation showed comparatively higher GCA variances, Baker ratios, and narrow-sense heritability for several traits, including plant height, pod number, seed size, and protein yield. Seed number per plant appeared to be influenced by both additive and non-additive gene effects, while significant reciprocal variance suggested the involvement of maternal and cytoplasmic inheritance. However, because the F₁ and F₂ populations were exclusively evaluated in different growing seasons (2023 and 2024, respectively), generation and year effects were completely confounded in the present study. Therefore, these differences should not be interpreted solely as intrinsic generation-dependent shifts in gene action.

Table 3
The genetic components of traits were examined in a full diallel cross

Genetic component	Variance	Cont. %	Variance	Cont. %	Variance	Cont. %	Variance	Cont. %
	Plant height		Number of pods per plant		Number of seeds per pod		Number of seeds per plant
	F1 Generation
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}$	16.85	8.41	0.04	0.55	0.08	5.00	82.37	5.39
${\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	43.08	21.51	0.58	8.93	0.14	8.75	87.60	5.73
${\sigma }^2\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	11.39	5.69	0.57	8.78	0.13	8.13	185.96	12.17
${\sigma }^2\mathrm{N}\mathrm{o}\mathrm{}\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	32.87	16.41	1.92	29.76	0.30	18.75	342.53	22.42
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}/{\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	0.39		0.07		0.57		0.94
D	33.70		0.07		0.16		164.74
H	43.08		0.58		0.14		87.6
$\mathrm{H}/{\mathrm{D}}^{1/2}$	0.39		0.06		0.56		0.94
Baker ration	0.44		0.11		0.53		0.65
${\mathrm{h}}^2$	0.15		0.01		0.09		0.09
${\mathrm{H}}^2$	0.34		0.09		0.17		0.14
	F2 Generation
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}$	73.97	22.00	0.11	2.15	0.52	25.74	245.40	20.20
${\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	81.28	24.17	0.96	19.03	0.16	8.09	186.07	15.32
${\sigma }^2\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	0.01	0.00	0.00	0.00	0.04	2.21	14.77	1.22
${\sigma }^2\mathrm{N}\mathrm{o}\mathrm{}\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	95.94	28.53	2.45	48.59	0.04	1.76	132.15	10.88
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}/{\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	0.91		0.11		3.25		1.32
D	147.94		0.22		1.04		490.79
H	81.28		0.96		0.16		186.07
$\mathrm{H}/{\mathrm{D}}^{1/2}$	0.91		0.11		3.18		1.32
Baker ration	0.65		0.18		0.86		0.73
${\mathrm{h}}^2$	0.36		0.04		0.40		0.33
${\mathrm{H}}^2$	0.56		0.23		0.47		0.46
	Grain yield per plant		Hundred seed weight		Protein content		Protein yield per plant
	F1 Generation
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}$	0.01	0.00	0.71	3.29	0.52	7.87	0.01	0.01
${\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	57.70	40.64	13.56	62.92	1.95	29.50	2.46	39.81
${\sigma }^2\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	14.83	10.44	0.55	2.55	0.56	8.47	0.45	7.28
${\sigma }^2\mathrm{N}\mathrm{o}\mathrm{}\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	19.61	13.81	2.00	9.28	3.55	53.71	0.47	7.61
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}/{\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	0.01		0.05		0.27		0.01
D	0.01		1.42		1.03		0.01
H	57.70		13.56		1.95		2.46
$\mathrm{H}/{\mathrm{D}}^{1/2}$	0.01		0.05		0.26		0.01
Baker ration	0.01		0.09		0.35		0.00
${\mathrm{h}}^2$	0.01		0.06		0.13		0.00
${\mathrm{H}}^2$	0.37		0.66		0.39		0.37
	F2 Generation
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}$	38.29	43.07	7.30	28.06	0.52	7.87	1.98	40.82
${\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	1.50	1.69	8.63	33.17	1.95	29.50	0.20	4.12
${\sigma }^2\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	0.64	0.72	0.01	0.01	0.56	8.47	0.03	0.62
${\sigma }^2\mathrm{N}\mathrm{o}\mathrm{}\mathrm{M}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{l}$	10.88	12.24	7.44	28.59	3.55	53.71	0.69	14.23
${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}/{\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$	25.53		0.85		0.27		9.90
D	76.58		14.61		1.03		3.96
H	1.50		8.63		1.95		0.20
$\mathrm{H}/{\mathrm{D}}^{1/2}$	25.59		0.85		0.26		9.9
Baker ration	0.98		0.63		0.35		0.95
${\mathrm{h}}^2$	0.60		0.44		0.13		0.58
${\mathrm{H}}^2$	0.61		0.70		0.39		0.61

• ${\sigma }^2\mathrm{G}\mathrm{C}\mathrm{A}$

  : Variance of general combining ability;

  ${\sigma }^2\mathrm{S}\mathrm{C}\mathrm{A}$

  : Variance of specific combining ability; D: Additive variance; H: Variance of dominance;

  $\mathrm{H}/{\mathrm{D}}^{1/2}$

  : Average degree of dominance;

  ${\mathrm{h}}^2$

  : Narrow-sense heritability;

  ${\mathrm{H}}^2$

  : Broad-sense heritability.

The relatively low precipitation and greater thermal stress during the 2023 growing season may have intensified genotype × environment (G×E) interactions and enhanced the apparent non-additive gene action, SCA, and heterosis contributions by enlarging the differences among the hybrid combinations. Conversely, the more favorable climatic conditions in 2024 may have reduced environmental stresses and allowed the additive genetic differences among the segregated lines to be expressed more consistently, resulting in higher GCA effects, Baker ratios, and heritability estimates. Consequently, the observed contrasts between the F₁ and F₂ populations most likely reflect the combined effects of genetic segregation and contrasting environmental conditions rather than purely generation-dependent changes in gene action. It should also be noted that the experiment was conducted under managed irrigation in a semi-arid agro-ecological environment. Therefore, the observed differences between years reflect contrasting climatic conditions under supplemental irrigation rather than responses to severe rainfed drought stress. Similar findings regarding the importance of additive effects and high-GCA parents have been reported in pea and bean breeding studies (Ceyhan et al. 2014b, Şimşek and Ceyhan 2017, Tekin and Ceyhan 2023, Tamüksek and Ceyhan 2024, Môro and Souza Júnior 2025). However, such interpretations should likewise be considered within the environmental context of each individual study. Future studies evaluating both generations simultaneously across multiple environments and years are required to clearly separate genetic from environmental effects.

The moderate-to-high narrow-sense heritability estimates observed in the F₂ generation should likewise be interpreted with caution. Although increased additive genetic variance after segregation may contribute to these estimates, the more favorable climatic conditions in 2024 likely reduced environmental variance, thereby increasing the relative contribution of additive variance to the phenotypic variance. Therefore, the higher heritability values cannot be attributed exclusively to genetic segregation.

The relatively strong SCA effects observed for grain and protein yields in the F₁ generation suggest an important contribution from dominance and epistatic interactions. However, these estimates should be interpreted with caution because the F₁ population was evaluated under the relatively drier and more stressful environmental conditions of 2023. As stated above, such conditions may have intensified G×E interactions and enhanced the apparent expression of SCA and heterosis by increasing differences among the hybrid combinations. Consequently, the observed SCA effects likely reflect the combined influence of genetic factors and environmental conditions rather than purely intrinsic non-additive gene action. Nevertheless, the superior hybrid combinations identified in this study will be valuable breeding material for developing segregating populations with high yield and protein potential. Similar findings regarding the contribution of non-additive gene effects to yield and quality traits have been reported in previous pea studies (Ceyhan et al. 2008, Gupta et al. 2024). Likewise, integrating heterosis with selection has been shown to improve breeding efficiency and accelerate genetic gain, although the expression of heterosis may vary depending on environmental conditions (Ceyhan et al. 2014b, Yadav et al. 2023).

In the F₁ generation, Serge showed significant positive GCA for plant height, while Jofs exhibited the highest contributions to seed number per pod, seed yield per plant, hundred-seed weight, and protein yield. Ultrillo contributed positively to seed size but negatively to pod and seed number. In F₂, GCA effects became more pronounced, where Serge maintained a strong influence on plant height and Jofs demonstrated a superior combining ability across yield and quality traits, highlighting its key breeding value. Ultrillo remained important for improving seed size, whereas Compana showed negative GCA for seed number and grain yield, indicating a limited contribution to yield components (Table 4).

Table 4
General combining ability (GCA) effects of parental pea genotypes for the studied traits in the F₁ and F₂ generations

Parents	Plant height	Number of pods per plant	Number of seeds per pod	Number of seeds per plant	Grain yield per plant	Hundred seed weight	Protein content	Protein yield per plant
F1 Generation
Boggie	-7.113*	0.200	0.047	1.520	0.454	0.301	0.398	0.250
Compana	3.987	0.333	0.013	2.053	-1.481	-1.053	-0.766	-0.587
Serge	5.920*	0.133	0.013	1.253	-3.369*	-2.423**	0.445	-0.575
Jofs	-0.313	0.200	0.480*	13.020	5.180*	1.444*	-0.192	1.078*
Ultrillo	-2.480	-0.867*	-0.553*	-17.847	-0.783	1.731**	0.116	-0.166
F2 Generation
Boggie	-6.487*	0.013	-0.08	-2.447	-1.277	-0.425	0.398	-0.267
Compana	-9.953**	-0.453	-0.713*	-16.347*	-5.898**	-2.415**	-0.766	-1.600**
Serge	13.447**	-0.153	-0.647*	-11.847*	-5.715**	-3.135**	0.445	-1.073*
Jofs	5.547*	0.947**	1.153**	27.620**	8.315**	2.155**	-0.192	1.787**
Ultrillo	-2.553	-0.353	0.287	3.020	4.573*	3.821**	0.116	1.152*

The contrasting parental contributions suggest that trait-specific parental selection is essential in pea improvement. Jofs should be prioritized in breeding programs targeting high yield potential, enhanced protein yield, and balanced seed size, while Ultrillo is a valuable genetic resource for large-seeded cultivars. Serge, on the other hand, is particularly useful in breeding programs where plant height is a desired attribute, for example, in developing vigorous or lodging-resistant ideotypes when used with an afila plant type. These results confirm that GCA is a strong determinant of hybrid performance and selection strategies, echoing findings in peas and other legumes, where additive gene action plays a crucial role in stabilizing performance across generations (Ceyhan et al. 2008, Ceyhan et al. 2014a, Yadav et al. 2023, Gupta et al. 2024, Ceyhan et al. 2025, Dalgıç et al. 2025).

In the F₁ generation, Jofs × Compana and Ultrillo × Serge emerged as the most superior combinations, showing high and significant SCA effects for yield-related traits, while Boggie × Ultrillo and Ultrillo × Compana also contributed positively to grain yield and seed size. In contrast, Jofs × Boggie showed negative SCA effects across several traits. In the F₂ generation, segregation enhanced variability, with Serge × Jofs, Jofs × Compana, and Ultrillo × Boggie exhibiting strong SCA effects for yield, seed number, and plant height. Additionally, Ultrillo × Serge and Ultrillo × Jofs demonstrated high genetic potential. However, combinations such as Serge × Compana and Jofs × Serge showed negative SCA effects, indicating limited breeding value (Table 5).

Table 5
Specific combining ability (SCA) effects of pea hybrids for the studied traits in the F₁ and F₂ generations

Hybrids	Plant height	Number of pods per plant	Number of seeds per pod	Number of seeds per plant	Grain yield per plant	Hundred seed weight	Protein content	Protein yield per plant
F1 Generation
Boggie x Compana	-2.520	0.000	-0.480	-11.053	-6.639*	-3.101**	-0.557	-1.592*
Boggie x Serge	3.213	0.533	0.020	3.413	-0.031	-0.047	0.588	0.307
Boggie x Jofs	0.280	0.467	0.220	9.48	-2.05	-2.181*	0.047	-0.274
Boggie X Ultrillo	2.280	-0.467	0.920*	17.347	13.956**	6.466**	-0.628	2.585
Compana X Boggie	1.167	-0.333	-1.500**	-39.000**	-7.313*	-0.183	1.040	-1.180*
Compana x Serge	-6.22	-1.267	0.220	-3.620	-1.704	-0.461	0.553	-0.141
Compana x Jofs	8.680*	2.000**	-0.080	11.447	8.647**	3.773**	-0.527	1.516
Compana x Ultrillo	-10.820**	-0.600	-0.047	-5.187	0.262	0.986	0.677	0.255
Serge x Boggie	-4.167	-1.333*	0.000	-8.333	-3.283	-1.133*	-0.005	-0.762
Serge x Compana	-16.500**	-0.667	-0.167	-8.167	-0.498	0.833	0.037	-0.023
Serge x Jofs	1.913	-0.633	0.420	7.080	-0.362	-1.191	-0.048	-0.264
Serge x Ultrillo	-1.587	1.433*	0.287	18.113	5.352*	0.523	-0.332	1.038
Jofs x Boggie	-10.333**	-1.000*	-1.333**	-40.167**	-8.580**	-0.100	-0.472	-2.087**
Jofs x Compana	0.833	4.667**	0.333	39.000**	9.055**	-0.833	0.797	2.097**
Jofs x Serge	-2.667	1.167*	0.833**	28.167**	9.972**	3.167**	-1.273	1.848**
Jofs x Ultrillo	6.313	-0.633	-0.68	-21.153	-12.421**	-5.511**	0.219	-2.520**
Ultrillo x Boggie	-2.833	0.000	-0.667*	-15.500*	-1.210	3.500**	-1.557	-0.678
Ultrillo x Compana	9.167**	0.333	0.333	9.500	5.033*	2.000**	-1.055	0.728
Ultrillo x Serge	14.667**	2.500**	1.333**	49.000**	14.170**	2.833**	-1.243	2.633**
Ultrillo x Jofs	2.000	-1.833**	-0.167	-14.500*	-5.355*	-1.333*	0.832	-0.915
F2 Generation
Boggie x Compana	2.387	0.320	0.013	2.580	-2.184	-2.905	-0.557	-0.490
Boggie x Serge	1.153	0.353	-0.053	-0.920	0.111	0.382	0.588	0.024
Boggie x Jofs	1.220	-0.580	0.480	6.280	-1.339	-2.241	0.047	-0.217
Boggie X Ultrillo	0.653	-0.613	-0.320	-9.287	3.111	4.925	-0.628	0.371
Compana X Boggie	7.333**	1.000**	0.1670	8.333	2.675	1.400	1.040	0.792*
Compana x Serge	-17.547**	-1.180*	-0.253	-10.687	-0.765	0.889	0.553	-0.174
Compana x Jofs	6.520	0.887	-0.220	-0.987	1.274	1.582	-0.527	0.113
Compana x Ultrillo	-5.213	-0.480	-0.020	-2.387	0.017	1.249	0.677	0.056
Serge x Boggie	-10.500**	-0.667**	0.833*	12.000*	1.500	-0.733	-0.005	0.363
Serge x Compana	-12.000**	-0.667**	-0.333	-8.333	-1.017	0.583	0.037	-0.225
Serge x Jofs	5.120	1.753**	1.047*	34.180**	3.852	-2.031	-0.048	0.883
Serge x Ultrillo	-3.780	-0.447	0.580	6.613	0.281	-0.531	-0.332	0.570
Jofs x Boggie	-3.667	0.500	0.167	7.333	1.540	0.000	-0.472	0.185
Jofs x Compana	11.167**	2.167**	-0.500	3.167	2.862*	1.833	0.797	0.872*
Jofs x Serge	-8.500**	-3.000**	-0.500	-35.167**	-10.097**	-2.167	-1.273	-2.775**
Jofs x Ultrillo	4.287	0.287	-0.387	-5.353	1.936	1.512	0.219	0.416
Ultrillo x Boggie	13.333**	0.500	-0.167	0.167	6.038**	5.833	-1.557	0.812*
Ultrillo x Compana	0.000	1.167**	0.500	15.167*	2.937*	-0.833	-1.055	0.373
Ultrillo x Serge	4.833*	1.167**	0.833*	22.000**	4.003*	-0.667	-1.243	-0.338
Ultrillo x Jofs	-2.333	0.000	1.000**	19.500**	0.535	-3.667	0.832	0.470

In contrast, combinations such as Jofs × Boggie exhibited negative SCA effects and heterosis for several traits. This may indicate limited genetic complementation between the two parents, resulting in reduced dominance and epistatic interactions. Negative heterosis can occur when favorable alleles are already fixed in both parents or when parental genomes share similar genetic backgrounds, thereby limiting opportunities for heterotic expression. Such combinations are therefore less suitable for exploiting hybrid vigor in breeding programs.

The significant reciprocal and maternal effects observed for several traits indicate that parental orientation should also be considered during hybrid development. The superior performance of combinations such as Jofs × Compana compared with their reciprocal counterparts suggests that Jofs is more effective as the female parent when breeding crosses for improving yield and protein-related traits. Similarly, the favorable performance of Ultrillo × Serge indicates potential maternal contributions from Ultrillo to seed size and yield attributes. These findings suggest that breeding programs should focus not only on parental selection but also on cross-direction, particularly when developing populations for yield and quality improvement.

These findings align with previous studies showing that high SCA effects in pea hybrids enhance yield and quality traits (Ceyhan et al. 2008, Yadav et al. 2023, Gupta et al. 2024). Similarly, the predominance of non-additive gene effects in protein and mineral traits highlights the importance of superior SCA combinations (Ceyhan et al. 2025, Dalgıç et al. 2025). The frequent observation of ${\sigma }^{2}SCA>{\sigma }^{2}GCA$ further confirms the role of dominance in hybrid performance (Ceyhan et al. 2008, Sharma et al. 2023). Combinations such as Jofs × Compana, Ultrillo × Serge, Ultrillo × Boggie, and Serge × Jofs stand out as valuable candidates for exploiting heterosis and generating high-performing segregating populations in pea breeding (Ceyhan et al. 2008, Gupta et al. 2024). These combinations not only provide superior performance in early generations but also generate broad genetic variability for effective selection in later generations. Therefore, integrating heterosis breeding with selection-based strategies can significantly enhance breeding efficiency and accelerate genetic gain.

The results identified Jofs × Compana and Ultrillo × Serge in F₁, and Serge × Jofs, Jofs × Compana, and Ultrillo × Boggie in F₂ as the most promising combinations for breeding. In F₁, Boggie × Ultrillo, Ultrillo × Serge, and Jofs × Compana exhibited high heterosis for grain yield, seed number, and protein yield, whereas Jofs × Boggie showed negative heterosis. In F₂, segregation altered heterosis patterns, with Serge × Jofs emerging as the best-performing combination, followed by Ultrillo × Serge and Ultrillo × Boggie. Conversely, Serge × Compana and Compana × Ultrillo displayed negative heterosis and limited breeding value (Table 6).

Table 6
Heterosis values of pea hybrids for the studied traits in the F₁ and F₂ generations

Hybrids	Plant height	Number of pods per plant	Number of seeds per pod	Number of seeds per plant	Grain yield per plant	Hundred seed weight	Protein content	Protein yield per plant
F1 Generation
Boggie x Compana	-10.34	2.74	18.92*	22.17*	10.58*	-9.01	-8.32**	1.66
Boggie x Serge	10.90	8.97*	15.15	25.47*	24.54*	5.55	3.10**	30.24*
Boggie x Jofs	38.46**	9.86*	28.21**	40.78**	13.82	-16.22**	0.40	16.07
Boggie X Ultrillo	7.73	-1.43	41.94**	38.63**	68.47**	22.10**	2.75**	63.78**
Compana X Boggie	-7.13	0.00	-29.73**	-29.71**	-32.70**	-10.90	1.30**	-31.01*
Compana x Serge	3.95	-2.04	11.11	8.72	2.28	-7.19	4.62**	5.77
Compana x Jofs	15.21*	-8.33*	-9.52	-16.87	-7.27	11.19*	-6.72**	-14.27
Compana x Ultrillo	-33.88**	-4.23	-5.88	-10.54	-2.89	4.05	8.40**	3.39
Serge x Boggie	-0.95	-2.07	15.15	12.92	3.05	-6.24	3.06**	7.79
Serge x Compana	-33.33**	-7.48	5.56	-2.38	-0.94	2.15	4.96**	5.05
Serge x Jofs	16.86*	-4.90	0.00	-3.43	-28.78**	-31.91**	6.65**	-27.13*
Serge x Ultrillo	-24.84**	-4.96	-6.67	-11.67	-12.74	-8.33	5.73**	-7.20
Jofs x Boggie	0.31	1.41	-12.82	-11.5	-25.33*	-17.02**	-3.80**	-27.05*
Jofs x Compana	17.51*	30.56**	0.00	29.56**	33.68**	4.15	0.90*	30.66*
Jofs x Serge	9.26	4.90	26.32**	34.00**	19.87*	-4.96	-5.02**	14.42
Jofs x Ultrillo	17.65*	7.25	-5.56	0.96	-15.81	-21.29**	-3.65**	-18.65*
Ultrillo x Boggie	-1.33	-1.43	16.13	13.15	60.71**	54.56**	-11.02**	44.36**
Ultrillo x Compana	-11.16	-1.41	5.88	3.43	28.97**	23.86**	-1.60**	25.32
Ultrillo x Serge	12.53	16.31**	46.67**	71.03**	87.17**	20.00**	-5.57**	77.70**
Ultrillo x Jofs	24.06**	-8.70*	-11.11	-19.86	-41.54	-31.61**	3.95**	-38.77**
F2 Generation
Boggie x Compana	-15.15**	-6.19	-6.25	-12.11	-25.08*	-20.61**	-7.96**	-31.68*
Boggie x Serge	8.09*	5.36	-3.70	1.18	0.95	2.92	2.97**	3.17
Boggie x Jofs	26.72**	-0.88	12.82	12.60	-0.50	-11.42**	0.38	0.51
Boggie X Ultrillo	-19.05**	-10.34**	-2.70	-12.85	-1.34	12.74**	2.64**	1.55
Compana X Boggie	9.09*	4.42	0.00	4.48	-2.10	-5.49	1.25**	-1.91
Compana x Serge	-22.06**	-2.75	12.00	8.57	6.85	0.40	4.41**	12.28
Compana x Jofs	-4.50	-1.82	8.11	6.35	1.48	-1.99	-6.42**	-6.17
Compana x Ultrillo	-20.48**	-13.27**	-14.29	-25.79*	-4.09	31.49**	8.04**	2.41
Serge x Boggie	-16.77**	-1.79	33.33*	29.53*	15.30	-4.86	2.93**	17.86
Serge x Compana	-48.53**	-10.09*	-4.00	-13.41	-5.14	7.46	4.75**	0.13
Serge x Jofs	16.18**	33.94**	50.00**	104.12**	67.49**	-4.98	6.37**	78.55**
Serge x Ultrillo	-9.12***	-10.71**	6.67	-6.34	2.88	15.65**	5.48**	41.19**
Jofs x Boggie	14.60**	4.42	17.95	24.66*	8.61	-11.42**	-3.65**	5.22
Jofs x Compana	29.00**	21.82**	-8.11	11.96	20.65	14.85**	0.86*	20.17
Jofs x Serge	-2.57	0.92	31.25*	31.62*	-5.99	-24.59**	-4.81**	-10.30
Jofs x Ultrillo	12.29**	4.42	-14.29	-9.87	18.82*	32.89**	-3.49**	14.85
Ultrillo x Boggie	23.28**	-5.17	-8.11	-12.57	38.57**	66.80**	-10.57**	24.36*
Ultrillo x Compana	-20.48**	-0.88	2.86	1.66	18.23	22.84**	-1.53**	15.12
Ultrillo x Serge	-19.50**	1.79	40.00**	40.14**	36.34**	8.84*	-5.34**	28.89**
Ultrillo x Jofs	19.04**	4.42	14.29	19.75	21.73*	3.36	3.78**	26.02**

On the other hand, the combinations Ultrillo × Serge, Boggie × Ultrillo, Serge × Jofs, and Jofs × Compana exhibited high heterotic performance in both the F₁ and F₂ generations (Table 6). These combinations particularly stand out for their grain yield, seeds per plant, and protein yield, offering strategic potential for developing high-yielding, high-quality hybrid candidates. These results are consistent with previous studies (Ceyhan et al. 2008, Sharma et al. 2023, Yadav et al. 2023, Jou-Nteufa and Ceyhan 2024, Gupta et al. 2024), which emphasize the importance of high heterosis in developing heterotic hybrids in early generations.

PCA analysis of the F₁ generation showed that principal component 1 and 2 (PC1 and PC2, respectively) explained 55.37% and 19.46% of the total variation, respectively, accounting for 74.83% cumulatively. The positive positions of Ultrillo × Boggie, Jofs, Boggie × Ultrillo, and Jofs × Compana on PC1 indicated strong associations with grain yield, hundred-seed weight, and protein yield. Conversely, Serge, Compana, and Compana × Serge were characterized by lower yields and seed-size performance. Among all combinations, Ultrillo × Boggie exhibited the highest PC2 score, highlighting its superiority for combining high grain yield with large seed size (Figure 1a).

Figure 1
Principal component analysis (PCA) biplot of pea hybrids showing relationships among yield and quality traits. (a) F₁, (b) F₂.

PCA analysis of the F₂ generation showed that PC1 and PC2 explained 55.02% and 17.16% of the total variation, respectively, accounting for 72.18% cumulatively. Positive PC1 scores of Serge × Jofs, Jofs × Ultrillo, Ultrillo × Boggie, and Jofs × Compana indicated strong associations with grain yield, hundred-seed weight, and protein yield, whereas Serge, Compana, and Boggie × Compana were linked to lower yield performance. Separation along PC2, particularly between Serge and Boggie × Serge versus Ultrillo × Boggie and Jofs × Ultrillo, reflected substantial trait variation among genotypes (Figure 1b).

The PCA results were largely consistent with the combining ability analyses. Hybrids positioned on the positive side of PC1, including Jofs × Compana, Ultrillo × Serge, Serge × Jofs, and Ultrillo × Boggie, were also identified as superior combinations based on SCA effects and heterosis estimates. Likewise, the strong positive GCA effects of Jofs for grain yield, seed number, and protein yield were reflected in the clustering of Jofs-derived hybrids near vectors associated with these traits. Therefore, the agreement among PCA, GCA, SCA, and heterosis analyses strengthens the evidence that these combinations represent the most promising breeding materials for semi-arid environments (Tables 4, 5, Figure 1a, b).

These results confirm PCA as an effective tool for summarizing relationships among multiple traits. Similarly, the effectiveness of PCA (Pratap et al. 2021) and its ability to identify links between large seeds and high yield, as well as guiding the selection of superior combinations in legumes, has been previously demonstrated (Matias et al. 2020, Sharma et al. 2023, Gupta et al. 2024).

CONCLUSIONS

The present study demonstrated that several parental combinations have substantial potential to improve grain yield and protein productivity under semi-arid conditions. Although strong heterosis was observed in the F₁ generation, the practical exploitation of hybrid vigor in pea is constrained by the self-pollinating nature of the crop and the difficulty of large-scale hybrid seed production. Consequently, the principal value of superior F₁ combinations lies in generating highly variable segregating populations for pure-line selection.

Based on the combined evidence from GCA, SCA, heterosis, and PCA analyses, the most promising breeding combinations were ranked as follows: Jofs × Compana, Serge × Jofs, and Ultrillo × Serge. These crosses consistently exhibited superior grain and protein yields, as well as overall breeding potential, across generations.

The F₁ population exhibited stronger SCA effects, while the F₂ population showed higher GCA effects and heritability estimates. However, these differences cannot be attributed exclusively to generation-dependent changes in gene action because generation and environmental effects were confounded in the present study. The relatively stressful conditions of 2023 may have enhanced heterosis and SCA effects through G×E interactions, whereas the more favorable conditions in 2024 likely facilitated the expression of additive genetic effects. Therefore, these findings should be interpreted as representing the combined effects of genetic segregation and contrasting environmental conditions. Confirmation through multi-environment and multi-year evaluations, in which both generations are grown simultaneously, will be necessary before drawing definitive conclusions about changes in gene action across generations.

In short, these breeding strategies provide a practical pathway to developing stable, high-yielding, protein-rich pea cultivars in semi-arid agriculture that maintain productivity under variable environmental conditions.

ACKNOWLEDGEMENTS

This study was supported by the Scientific Research Projects Coordination Unit (BAP) of Selçuk University under Project No. 23211020. This article was derived from the PhD thesis of Dr. Gökhan Akkaya.

Data Availability Statement

The datasets generated and/or analyzed during the current research are available from the corresponding author upon reasonable request.

REFERENCES

• AOAC - Association of Official Analytical Chemists2023 Official method 979.09. In Latimer GW (ed) Official methods of analysis of AOAC international. AOAC International, Gaithersburg, Chapter 32.2.03.
• Ceyhan E, Kahraman A2013 Genetic analysis of yield and some characters in peasLegume Research 36:273-279
• Ceyhan E, Karadaş S2023 Forage pea pure lines: winter hardiness, high seed and biological yieldSelcuk Journal of Agriculture and Food Sciences 37:487-495
• Ceyhan E, Avci MA, Karadaş S2008 Line × tester analysis in pea (Pisum sativum L.): identification of superior parents for seed yield and its componentsAfrican Journal of Biotechnology 7:2810-2817
• Ceyhan E, Harmankaya M, Kahraman A2014a Combining ability and heterosis for concentration of mineral elements and protein in common bean (Phaseolus vulgaris L.)Turkish Journal of Agriculture and Forestry 38:581-590
• Ceyhan E, Kahraman A, Avci MA, Dalgıç H2014b Combining ability of bean genotypes estimated by line × tester analysis under highly calcareous soilsJournal of Animal and Plant Sciences 24:579-584
• Ceyhan E, Korkmaz A, Ali A, Karadaş S, Harmankaya M, Şimşek D, Karaköy T2025 Exploring genetic diversity: the inheritance of protein and mineral contents in dwarf common beansTurkish Journal of Agriculture and Forestry 49:24-36
• Dalgıç H, Ali A, Karadaş S, Korkmaz A, Ayran Çolak İ, Harmankaya M, Karaköy T, Ceyhan E, Türkoğlu A2025 Diallel analysis of cowpea (Vigna unguiculata L. Walp) for protein and mineral composition: genetic variability and heritabilityEuphytica 221:168
• Griffing B1956 Concept of general and specific combining ability in relation to diallel crossing systemsAustralian Journal of Biological Sciences 9:463-493
• Gupta A, Singh B, Kumar M, Sharma V, Chauhan C, Rout S, Yadav S2024 Combining ability and heterosis analysis for seed yield and yield-related traits in table pea (Pisum sativum L. var. hortense)Legume Research 47:1858-1866
• Jou-Nteufa C, Ceyhan E2024 Genetic analysis of seed yield and some traits in cowpea using diallel analysisTurkish Journal of Agriculture and Forestry 48:278-286
• Karadaş S, Ceyhan E2023 Determination of the effects of organic and chemical fertilization on grain yield and some agricultural characteristics of peaSelcuk Journal of Agriculture and Food Sciences 37:419-426
• Matias FI, Valle CB, Gouveia BT, Moro GV, Barrios SCL2020 Using additive indices and principal components to select sexual genitors and hybrids of Urochloa decumbensCrop Breeding and Applied Biotechnology 20:e28082022
• Môro GV, Souza Júnior CL2025 Genetic control of efficiency in association of tropical maize with Azospirillum brasilense for grain yieldCrop Breeding and Applied Biotechnology 25:e51622534
• Pratap V, Sharma V, Kumar H, Kamaluddin Kamaluddin, Shukla G, Kumar M2021 Multivariate analysis of quantitative traits in field pea (Pisum sativum var. arvense)Legume Research 47:917-922
• R Core Team2024 R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. Available at <Available at https://www.R-project.org/ >. Accessed on June 12, 2026,
  » https://www.R-project.org/
• Ramalho MAP, Gonçalves FMA, Techio VH, Souza LC2025 Advances in genetics and plant breeding 160 years after the publication of Mendel’s researchCrop Breeding and Applied Biotechnology 25:e521725212
• Rodríguez F, Alvarado G, Pacheco A, Crossa J, Burgueno J2015 AGD-R: analysis of genetic designs with R for Windows Version 4.0. Centro Internacional de Mejoramiento de Maíz y Trigo (CIMMYT). Available at <Available at https://hdl.handle.net/11529/10202 >. Accessed on June 11, 2026.
  » https://hdl.handle.net/11529/10202
• Sharma A, Yadav R, Sheoran R, Kaushik D, Mohanta T, Sharma K, Yadav A, Dhanda P, Kaushik P2023 Estimation of heterosis and combining ability effects for yield and its attributes in field pea using PCA and GGE biplotsHorticulturae 9:256
• Şimşek D and Ceyhan E2017 Inheritance of some agronomic characters in peaJournal of Agricultural Sciences 23:34-41
• Souza BM, Abreu LM, Pappas MCR, Azevedo V, Santos PET, Sousa VA, Santos RF, Lopes MTG, Aguiar AV2023 Using of relatedness and heritability in a Eucalyptus benthamii trial for conservation and breedingCrop Breeding and Applied Biotechnology 23:e46452348
• Tamüksek Ş and Ceyhan E2024 Genetic analysis of grain yield and its components in green bean for soils with high lime contentBrazilian Archives of Biology and Technology 67:e24231090
• Tekin NB, Ceyhan E2023 Diallel analysis of dry bean varieties for seed yield and important traits for calcareous soilsBrazilian Archives of Biology and Technology 66:e23230170
• Yadav I, Sharma V, Kumar M, Yadav L, Mishra A, Singh V, Dhandha P, Yadav A, Yadav M, Singh S, Kamaluuddin P, Kaushik P, Gangwar V2023 Assessment of gene action and identification of heterotic hybrids for enhancing yield in field peaHorticulturae 9:997

Edited by

• SCIENTIFIC EDITOR:
  Luiz Antônio dos Santos Dias

Publication Dates

• Publication in this collection
  28 Sept 2026
• Date of issue
  2026

History

• Received
  22 May 2026
• Accepted
  31 July 2026
• Published
  20 Aug 2026