Issue 3 (203)/2025

ALSE and ACS Style
Bankole, F.A.; Aboderin, O.S.; Kolawole, A.O.; Olajide, O. Genetic diversity and trait associations in sorghum (Sorghum bicolor L. Moench) germplasm under rainfed conditions. Journal of Applied Life Sciences and Environment 2025, 58 (3), 387-410.
https://doi.org/10.46909/alse-583182

AMA Style
Bankole FA, Aboderin OS, Kolawole AO, Olajide O. Genetic diversity and trait associations in sorghum (Sorghum bicolor L. Moench) germplasm under rainfed conditions. Journal of Applied Life Sciences and Environment. 2025; 58 (3): 387-410.
https://doi.org/10.46909/alse-583182

Chicago/Turabian Style
Bankole, Folusho Anuoluwapo, Olawale Serifdeen Aboderin, Adesike Oladoyin Kolawole, and Olasunkanmi OlajidE. 2025. “Genetic diversity and trait associations in sorghum (Sorghum bicolor L. Moench) germplasm under rainfed conditions.” Journal of Applied Life Sciences and Environment 58, no. 3: 387-410.
https://doi.org/10.46909/alse-583182

Genetic diversity and trait associations in sorghum (Sorghum bicolor L. Moench) germplasm under rainfed conditions

Folusho Anuoluwapo BANKOLE¹, Olawale Serifdeen ABODERIN^1*, Adesike Oladoyin KOLAWOLE² and Olasunkanmi OLAJIDE¹

¹Department of Agronomy, Faculty of Agriculture, University of Ilorin, Ilorin, Kwara State, Nigeria; email: bankole.fa@unilorin.edu.ng; olajide.sunkanmiolushola@gmail.com

²Department of Crop Production and Soil Science, Faculty of Agricultural Sciences, Ladoke Akintola University of Technology, Ogbomoso, Oyo State, Nigeria; email: aokolawole@lautech.edu.ng

*Correspondence: olawaleaboderin@yahoo.com 

Received: Apr. 16, 2025. Revised: Jul. 20, 2025. Accepted: Aug. 18, 2025. Published online: Sep. 19, 2025

ABSTRACT. Sorghum is a key cereal crop in Sub-Saharan Africa that is used as food, fodder, and biofuel. This study evaluated the genetic diversity and trait associations among sorghum accessions collected from the Nigeria Southern Guinea Savanna (SGS). A total of 160 accessions were initially collected from farmers’ fields and preliminarily screened based on distinct morphological traits. Of these, 20 distinct accessions were selected and evaluated in 6 SGS environments with 2 improved varieties under rainfed conditions in the 2021 and 2022 growing seasons. The results revealed moderate to high genotypic and phenotypic coefficients of variation and high heritability estimates (66.83–97.75%) for most traits, indicating that the accessions harbour valuable genetic diversity and that selection among them would be effective for breeding purposes. Accessions LR55 (1323 kg ha⁻¹), DR25 (1293 kg ha⁻¹), and LR2 (1226 kg ha⁻¹) were selected for their high yield and stability, and LR2, DR18, and DR15 were selected through a Genotype by Yield*Trait biplot approach for their optimal combination of yield and nutritional quality traits. These accessions are promising candidates for both direct cultivation and as elite genetic resources in sorghum improvement programs. Significant positive and negative correlations were observed among agronomic and nutritional traits, but most yield-related traits showed non-significant correlations with nutritional quality traits. Leaf width, percentage of grain coverage, number of panicles harvested, 100-seed weight, and 1000-seed weight were identified as reliable selection indices for yield improvement based on their high heritability and strong positive correlation with grain yield.

Keywords: genotype × environment interaction; nutritional and antinutritional traits; sorghum accessions; sorghum grain quality; trait correlations; yield-related traits.

 

INTRODUCTION

Sorghum (Sorghum bicolor L. Moench), a versatile crop with food and feed applications, plays an important role in global agriculture (Kavithamani et al., 2019; Tasie and Gebreyes, 2020). Despite its lesser recognition compared to other major cereal crops, such as maize (Zea mays L.) and wheat (Triticum aestivum L.), sorghum is immensely important, especially in regions with harsh growing conditions, due to its resilience and adaptability to diverse environments (Hegde et al., 2023). Considering its rich nutritional profile boasting high levels of proteins, vitamins, minerals, dietary fibre, and bioactive compounds, such as phenolic, sorghum provides valuable nutrition for both humans and animals (Nemukondeni et al., 2022; Tasie and Gebreyes, 2020).

Genetic diversity is a fundamental prerequisite for effective crop improvement, providing a foundation for selection, adaptation, and long-term genetic gain (Shariatipour et al., 2022) and reflecting a species’ capacity to survive and thrive under unpredictable environmental conditions (Govindaraj et al., 2015). Landraces, in particular, are important reservoirs of genetic variability. These traditional cultivars have been selected and maintained by farmers over generations and have been adapted to specific agroecological conditions (Elangovan et al., 2012). Significant efforts have been made to collect sorghum landraces in Nigeria, particularly in the core northern states. Despite the widespread cultivation of diverse landraces by smallholder farmers in the Southern Guinea Savanna (SGS), comprehensive evaluations of their genetic diversity, particularly in relation to agronomic and nutritional traits, remain limited.

Understanding trait associations is important for interpreting genetic variation and guiding effective selection in breeding programs. Relationships between traits have significant implications for the development of selection strategies. Trait associations typically arise due to genetic linkage, pleiotropy, or both and can manifest in three general forms, each with different breeding consequences (Assefa et al., 2020). When two traits have a weak or no correlation, improvements in one trait are unlikely to affect the other, enabling independent selection. A positive correlation between traits indicates that selection for one trait can simultaneously improve the other trait, facilitating efficient indirect selection. Negative correlations between traits often pose a challenge in breeding, as enhancing one trait may compromise the other trait, resulting in potential trade-offs. To address such complexities, the Genotype by Yield*Traits (GYT) biplot approach proposed by Yan and Frégeau-Reid (2018) has emerged as a valuable selection tool. This method enables breeders to simultaneously select for multiple traits by optimising their combined performance, even in the presence of negative correlations. GYT has been effectively applied in crops, including oat, sesame, wheat, cotton, barley, and sunflower, to identify genotypes that achieve a desirable balance between high yield and favourable agronomic or nutritional traits (Merrick et al., 2020; Peixoto et al., 2022; Yan et al., 2019). Studies by Feyzbakhsh et al. (2020) and Welderufael et al. (2024) demonstrated the utility of GYT in identifying sorghum genotypes with superior yield–trait combinations across diverse genetic backgrounds and environments Genotype × environment interactions (GEIs) are another important consideration in plant breeding, as they significantly influence the phenotypic expression, adaptability, and stability of genotypes across diverse environments (Aboderin et al., 2023; Shahriari et al., 2018). Genotype expression may vary considerably under different environmental conditions, leading to inconsistencies in performance that complicate superior cultivar selection. Multi-environment trials (METs) are widely used to evaluate GEI, offering insights into genotype adaptability and environmental influences (Bankole and Aboderin, 2024). Statistical tools, such as the Additive Main Effects and Multiplicative Interaction (AMMI) and Genotype plus Genotype-by-Environment (GGE) biplots, have become indispensable in this process. These methods combine analysis of variance and principal component analysis to dissect GEIs. Although AMMI quantifies the contributions of genotypes and environments to GEI, GGE biplots provide visual interpretations for identifying stable, high-performing genotypes, discriminating among environments, and defining mega-environments (Yan and Tinker, 2006). Therefore, the objectives of this study were to (i) determine the genetic diversity among sorghum landraces in the Nigerian SGS, (ii) investigate the relationships among agronomic and nutritional traits, and (iii) identify superior accessions suitable for use as parents in breeding programs.

 

MATERIALS AND METHODS

Genetic materials

The genetic materials for this study comprised 20 sorghum accessions and 2 improved varieties obtained from the International Crop Research Institute for the Semi-Arid Tropics (ICRISAT), Kano, Nigeria. The accessions were derived through a systematic five-stage sampling procedure. In the first stage, four local government areas (LGAs) in Kwara State, namely Ilorin East, Ilorin South, Ifelodun, and Irepodun, were purposively selected based on the known diversity in their sorghum morphotypes. The second stage involved a stratified random sampling of five towns from each selected LGA to ensure broad and representative coverage of sorghum’s genetic variation in the region. The selected towns were Oke-Ose, Oke-Oyi, Apata-Yakuba, Ile-Apa, and Sentu in Ilorin East, Agbabiaka, Tanke, Unilorin Environs, Fufu, and Akanbi in Ilorin South, Babanla, Sare, Omupo, Ganmu-Alheri, and Oke-Ode in Ifelodun, and Ajase, Omu Aran, Gaa Federal, and Ara Orin in Irepodun. In the third stage, collection areas within these towns were targeted for their morphologically diverse sorghum cultivation and purposively selected based on local knowledge and expert recommendations. The fourth stage involved the random collection of 160 sorghum accessions between 10 December 2019 and 10 January 2020, either during harvest or from farmers’ stored produce. In the fifth stage, 20 accessions were selected from the initial pool based on observable and distinct morphological traits of agronomic and taxonomic relevance, including panicle structure (compact vs. loose), peduncle exertion (exserted vs. enclosed), grain shape and size, glume and grain coloration, presence of awns, and leaf orientation (Bankole et al., 2025). The selected accessions were screened for uniformity and seed viability prior to inclusion in the multi-environment trials (METs).

Experimental sites

Field trials were carried out over 2 successive years (2021 and 2022) under rainfed conditions at the University of Ilorin Teaching and Research Farm (Unilorin T&R) Station I, Unilorin T&R Station II, and the Institute for Agricultural research (IAR), Mokwa, Niger State. Both Unilorin stations are within the southern Guinea Savanna (SGS) ecology of Nigeria, with Station I located at 8°30’N, 4°32’E and an altitude of 289 m and receiving a mean annual rainfall of 1318 mm. Station II shares similar agroecological features but is located near a riverside typically utilised for irrigation with distinct soil characteristics compared to Station I, predominantly featuring sandy loam soil. IAR Mokwa, which is also situated in the SGS ecology, is at an elevation of 457 m, a longitude of 5°E, latitude 9°18’N, and receives an annual rainfall of approximately 1100 mm.

The location–year combination was treated as a distinct environment, resulting in six environments: Environment 1: Unilorin T&R Station I, 2021; Environment 2: Unilorin T&R Station I, 2022; Environment 3: Unilorin T&R Station II, 2021; Environment 4: Unilorin T&R Station II, 2022; Environment 5: IAR Mokwa, 2021; and Environment 6: IAR Mokwa, 2022.

Experimental design and agronomic practices

The experiment was arranged in a randomised complete block design with two replications. Two rows, each measuring 5 m in length, were utilised, with inter and intra-row spacings set at 0.75 and 0.5 m, respectively. Planting was carried out at the onset of rainfall in July of each year, with a 2-week interval maintained between planting dates across the locations to facilitate effective data collection and monitoring. Five seeds were initially sown per hill and later thinned to two plants per hill at two weeks after planting. Urea fertiliser was applied six weeks after planting at a rate of 50 kg N ha⁻¹. Weed control was performed manually and chemically as required to maintain a weed-free environment throughout the growing season.

Data collection

Data were collected on a plot and sample basis. Agronomic data, including plant height (PHT, cm), number of leaves (NL), leaf length (LL, cm), leaf width (LW, cm), inflorescence length (IL, cm), inflorescence width (IW, cm), and percentage grain coverage (GC, %), were recorded from five randomly selected plants per plot in each replication. The number of days to 50% flowering (DF) was recorded as the number of days for 50% of the plants per plot to flower. The weight at harvest (WH, g) and grain yield (GY, g) were determined in each plot and converted to kg ha⁻¹. The 100- (HSW) and 1000-seed weights (TSW, g) were also recorded.

Representative samples of each sorghum accession were collected for detailed nutritional and antinutritional analysis in the laboratory. The analysed nutritional components included the carbohydrate content (CHO), caloric content, lipid content, dietary fibre, and protein content. CHO (%) was determined using the phenol-sulfuric acid method, and the caloric content (kcal 100 g⁻¹) was calculated using the Atwater factor. The lipid content (%) was measured using the Soxhlet extraction method, and the dietary fibre content (%) was assessed using the enzymatic-gravimetric method. The protein content (%) was determined using the Kjeldahl method, with nitrogen converted to protein using a factor of 6.25 (AOAC, 2019). The analysis of antinutritional factors was conducted using specific protocols. The oxalate content (mg g⁻¹) was determined following the method described by Day and Underwood (1986) and Kayode et al. (2013). The phytate content (mg g⁻¹) was quantified using the colorimetric method (McKie and McCleary, 2016), and the hydrocyanic acid (mg g⁻¹) levels were analysed using the alkaline titration method. The phenolic content (mg g⁻¹) was determined spectrophotometrically using Folin–Ciocalteu reagent (Makkar et al., 2009), and the tannin content (mg g⁻¹) was measured using the vanillin-HCl method (Maxson and Rooney, 1972). To ensure the accuracy of the results, all measurements were performed in triplicate, and standard reference materials were included where appropriate for validation.

Data analysis

The collected data were subjected to combined analysis of variance (ANOVA) across the six environments. Prior to ANOVA, the homogeneity of error variances across environments was assessed using Levene’s test for each trait. Levene’s test returned non-significant results (p > 0.05) for all traits, indicating that the assumption of homogeneity was satisfied and that a combined ANOVA was appropriate. The combined ANOVA was performed using the General Linear Model (GLM) procedure in the Statistical Analysis System (SAS Institute Inc., 2008) version 9.2. In this model, the genotype was treated as a fixed effect, and environment and their interactions were considered random effects. The following genetic parameters were calculated from the variance components derived from ANOVA (Equations 1-7):

where is the mean square due to genotype; is the mean square due to the GEI; is the mean square due to error (residual); is the number of replications; is the number of environments; is the mean; and is the selection differential (2.06 at 5% selection intensity). Genotypic coefficient of variation (GCV) and Phenotypic coefficient of variation (PCV) were categorised as low (<10%), moderate (10–20%), and high (>20%). Broad-sense heritability (h²) was classified following Johnson et al. (1955): low (0–30%), medium (31–60%), and high (>60%). Following Johnson et al. (1955), genetic gain was categorised as low (<10%), moderate (10–20%), and high (>20%).

GGE biplot analysis

The 20 accessions were further subjected to GGE biplot analysis to evaluate genotype performance and stability across environments using GEA-R software (version 2.0) (Pacheco et al., 2016). The GGE model (Yan and Tinker, 2006) was calculated as follows (Equation 8):

 

where  is the mean yield of accession i in environment j;  is the mean yield of all accessions in environment j;  are singular values for the first and second principal components (PC1 and PC2);  are the scores of accession i for PC1 and PC2;  are the scores of environment j for PC1 and PC2; and  is the residual associated with accession i in environment j not explained by PC1 or PC2.

Correlation analysis

Pearson correlation coefficients (r) were computed using SAS software (version 9.2) to determine the strength and direction of the linear relationships among agronomic, nutritional, and antinutritional traits. To further explore trait interrelationships, Genotype × Trait (GT) biplot analysis was conducted using GEA-R software (Pacheco et al., 2016). As the traits were measured on different scales, all variables were standardised prior to GT biplot analysis.

Genotype by Yield*Trait (GYT) analysis

GYT biplot analysis was performed to identify accessions with optimal combinations of GY and nutritional quality. The GY of each accession was multiplied by its respective nutritional, and antinutritional trait values to generate GYT values following the procedure outlined by Yan and Frégeau-Reid (2018). These values were then standardised and subjected to biplot analysis using GEA-R Windows software.

 

RESULTS

Genetic variability, heritability and genetic gain

The combined ANOVA revealed highly significant (p < 0.01) environmental (E) and accession (A) effects for all evaluated agro-morphological traits, except for the number of panicles harvested (PH), where E effects were not significant. Significant (p < 0.01) A × E interactions were observed for DF, WH, and GY (Table 1). For nutritional and antinutritional traits, highly significant differences (p ≤ 0.01) were observed among the sorghum accessions (Table 1).

Table 2 shows the genetic parameters for morphological traits among the sorghum accessions. PCV exceeded GCV for all evaluated agro-morphological traits. GY had the highest GCV and PCV, while LL had the lowest values for both parameters. Eight traits, namely DF, NL, LW, PHT, IL, IW, HSW, and LL exhibited a low GCV (<10%), whereas the remaining traits had a moderate GCV (10–20%).

Similarly, WH and GY recorded high PCV values, and seven traits (DF, NL, LL, LW, IL, IW, and HSW) had a low PCV. The other traits, namely GC, PHT, PH, and TSW had a moderate PCV. Broad-sense heritability (h²) ranged from 66.83% for DF to 97.75% for LL (Table 2). Five traits, namely LL, LW, TSW, PH, and PHT showed heritability estimates above 90%. The remaining traits had moderate heritability (70–90%), except for DF and IW, which had values below 70%. A high genetic gain (>20%) was observed for GY, WH, PH, GC, and TSW. A moderate genetic gain (10–20%) was recorded for PHT, LW, DF, HSW, and NL, and the remaining traits had low genetic gain estimates (<10%).

Agro-morphological performances of the sorghum accessions

Table 3 shows the agronomic mean performances of the improved sorghum accessions and landraces. Improved varieties exhibited a higher PH compared to local varieties. Additionally, improved varieties recorded a higher WH, GY, HSW, and TSW. Improved varieties also flowered earlier, taking only 62 days compared to 84 days for local varieties. They had a higher total NL and a greater LW. In contrast, the local accessions outperformed the improved varieties in terms of PHT, LL, IL, and IW.

For the landraces, DR1 had the highest (30) PH, while LR24 and LR6 had the lowest (Table 2). WH varied from 1051 to 1374 kg ha⁻¹, and GY varied from 954 to 1323 kg ha⁻¹. Accession LR55 had the highest GY, followed by DR25 and LR2, and W60 recorded the lowest. Among the local accessions, DF varied from 78.5 to 90 days, with DR38 flowering the earliest. For leaf morphology, 14 varieties had 7 leaves, while the others had 8. DR18, R57, LR11, R20, and R73 had the highest LL, while LR55 recorded the lowest. LR46 had the widest LW, while LR2 had the narrowest. PHT varied from 192.5 to 245.8 cm, with DR25 being the tallest and LR46 the shortest. IL ranged from 8.9 to 9.6 cm, and IW ranged from 5.7 to 6.8. R20 and R79 had the longest panicle length, and DR49 had the highest panicle width. Most varieties had over 50% GC, with DR1 recording the best GC.

Nutritional and antinutritional composition of the sorghum accessions

Among the nutritional qualities, the phenolic content of the sorghum landraces ranged from 9.6 (W60) to 32.3 mg g⁻¹ (R73), with an average of 20.9 mg g⁻¹ (Table 4).

The moisture content ranged from 5.2 (LR11) to 7.9% (DR18), with an average of 6.0%. LR2 had the highest ash content, while W86 had the lowest. CHO varied from 69.81 to 81.04%, with LR55, DR38, and DR25 having the highest CHO.

Accession LR2 had the highest lipid content, while W60 and W86 had the lowest. The fiber content ranged from 2.26 (LR55) to 4.1% (R57), with an average of 3.25%. The highest protein content was observed in LR24, LR2, and DR18, while DR25 and DR38 had the lowest. Caloric content varied from 355.34 to 376.18 kcal 100 g⁻¹, with an average of 364.3 kcal 100 g⁻¹. The top three accessions with the highest caloric value were LR2, LR11, and LR24, while R73 had the lowest.

Among the antinutritional factors, oxalate levels ranged from 2.9 (DR25) to 14.5 mg g⁻¹ (W86), with an average of 8.8 mg g⁻¹. Of the 20 accessions, 15 had a phytate content of 0.4 mg g⁻¹, 4 had a phytate content of 0.3 mg g⁻¹, and only 1 (DR15) had a phytate content of 0.2 mg g⁻¹. The lowest hydrogen cyanide (HCN) levels were observed in W86, and the highest was measured in R65. The tannin content varied from 1.0 (LR2) to 4.4 mg g⁻¹ (DR15), with an average of 2.7 mg g⁻¹.

Correlation among sorghum accession traits

The angle between two traits indicates their correlation in the biplot (Figures 1-3). An angle less than 90º denotes a positive correlation, with smaller angles indicating stronger positive correlations. Angles greater than 90º, but less than 180º signify a negative correlation, while angles closer to 180º indicate stronger negative correlations. Traits with shorter vector lengths are typically considered not significant.

Correlation among agronomic traits

Table 5 and Figure 1 present the correlation results among the agronomic traits of the sorghum accessions. Strong positive correlations were observed among the yield-related traits, including GC, PH, WH, GY, HSW, and TSW (r > 0.45, p ≤ 0.05). DF was significantly negatively correlated with all yield-related traits (r < −0.7, p ≤ 0.01), except for HSW, which exhibited a non-significant negative correlation. Leaf morphology traits showed distinct correlation patterns. NL had a significant negative correlation with LL, and LL and LW had a strong negative correlation (r = −0.90, p ≤ 0.01). LL also exhibited strong positive correlations with DF and PHT (r > 0.8, p ≤ 0.01) but showed strong negative correlations with all yield-related traits (r > 0.6, p ≤ 0.01). In contrast, LW had strong positive correlations with all yield-related traits and strong negative correlations with DF and PHT. Inflorescence traits showed variable relationships. IL was significantly positively correlated with DF, LL, and PHT (r > 0.75, p ≤ 0.01) but was strongly negatively correlated with all yield-related traits. IW exhibited a significant positive correlation with DF (r = 0.45, p ≤ 0.05) and non-significant correlations with all other traits. PHT had a strong positive correlation with DF but strong negative correlations with all yield-related traits (r < −0.50, p ≤ 0.01).

Correlation among nutritional and antinutritional traits

The correlations among the nutritional and antinutritional traits of 20 sorghum landraces are presented in Table 6 and Figure 2.

The lipid content exhibited a significant positive correlation with the phenolic (r = 0.42, p ≤ 0.05) and ash contents (r = 0.77, p ≤ 0.01).

 

 

 

 

The oxalate content was negatively correlated with the ash content (r = −0.40, p ≤ 0.05). The phytate content was negatively correlated with the moisture content (r = −0.73, p ≤ 0.01), and the caloric content had a strong negative correlation with the fibre content (r = −0.99, p ≤ 0.01). CHO exhibited a negative correlation with the fibre (r = −0.91, ρ ≤ 0.01) and protein contents (r = −0.49, p ≤ 0.05) but a positive correlation with the caloric content (r = 0.89, p ≤ 0.01).

Correlation between agronomic and nutritional traits

The correlations among the agronomic and nutritional traits of the sorghum accessions are presented in Table 7 and Figure 3. Several significant correlations, both positive and negative, were identified. DF was significantly positively correlated with the protein content (r = 0.54, p ≤ 0.05) and significantly negatively correlated with CHO (r = −0.45, p ≤ 0.05). NL showed a significant negative correlation with CHO (r = −0.46, p ≤ 0.05). Among the yield-related traits, WH displayed a significant negative correlation with the oxalate content (r = −0.44, p ≤ 0.05). PH was significantly positively correlated with the moisture (r = 0.51, p ≤ 0.05) and fibre contents (r = 0.65, ρ ≤ 0.05) but significantly negatively correlated with the caloric content (r = −0.69, p ≤ 0.05). HSW exhibited a significant negative correlation with the caloric content (r = −0.50, p ≤ 0.05) and a significant positive correlation with the fibre content (r = 0.59, p ≤ 0.05).

In the biplot (Figure 3), DF and HSW were strongly positively correlated with the fibre and protein contents and moderately positively correlated with the moisture content.

 

Both traits also displayed a perfect linear relationship with the caloric content but were strongly negatively correlated with the carbohydrate, ash, and lipid contents. GY and WH demonstrated strong positive correlations with the caloric, ash, and lipid contents and strong negative correlations with the fibre and phytate contents. The vector lengths of the other agronomic and nutritional traits on the biplot were relatively short, indicating weaker relationships and minimal contributions to the observed variation; thus, they were not considered for further interpretation.

Performance and stability of sorghum accessions across environments

The performance and stability of 20 sorghum accessions were evaluated using GGE biplot analysis (Figure 4 and Figure 5). The two principal components, PC1 and PC2, explained 66.83% of the total variation in GY across 6 environments in the Nigerian SGS.

The “Which-Won-Where” polygon view (Figure 4) identified the top-performing accessions in specific environments. Accessions at the vertices of the polygon closest to the environment markers represented the best-performing genotypes in those environments. Accession LR55 demonstrated superior GY in all test environments, except E3, where LR6 was the top-performing genotype. The “Mean vs. Stability” view (Figure 5) assessed the mean GY and stability of the accessions across environments. The single-arrowed line, representing the average environment coordinate (AEC) axis, showed the direction of the higher mean GY. Accessions positioned further along this axis achieved higher average yields. Seven accessions recorded GYs above the overall mean, with LR55 exhibiting the highest yield, followed by DR25, LR2, DR1, LR11, and LR6. Stability was evaluated based on the projection of each accession onto the axis perpendicular to the AEC. Accessions with shorter projecttions were more stable across environments, and LR55 was the most stable genotype with consistently high yields. DR25 and LR2 had moderate stability, and other high-yielding ace-ssions showed relatively lower stability.

Genotype by Yield*Traits (GYT)

Figure 6 displays the polygon view of the GYT biplot, showing the combinations of yield and nutritional traits among the evaluated sorghum accessions. The biplot explained 48.87% of the total variation, with PC1 and PC2 accounting for 25.04 and 23.83%, respectively. The genotype positioned at the vertex closest to the point representing a yield–nutritional trait combination indicates superior performance in those traits. LR2 emerged as the vertex genotype for the combinations of GY with the caloric, ash, lipid, and phenolic contents. DR15 and DR18 showed superior performance for the combinations of GY with the protein, moisture, and HCN contents. R57 was the leading genotype for the GY–oxalate combination, and LR55 was top performing for the GY–CHO and GY–phytate combinations. DR1 exhibited superior performance for the GY–fibre combination.

Figure 7 illustrates the “mean vs. stability” view, highlighting the strengths, weaknesses, and overall superiority of the accessions across yield–trait combinations.

 

 

 

 

 

The absolute length of each accession’s projection on the average tester axis represents its mean performance across all combinations, with double-arrow lines distinguishing accessions with below-average means from those with above-average means. Among the accessions, LR2, DR18, and DR15 had the highest mean performance across combinations of GY and nutritional traits.

 

DISCUSSION

The observed significant genotype variance for GY and most measured traits in the combined ANOVA indicates that the sorghum accessions have useful inherent genetic potential. Future breeding work should be built upon the existing genetic variability present in the population, and as such, the observed genetic variation among the collected accessions in the Nigeria SGS is highly desirable. The significant environment variance observed from the ANOVA for most traits indicate that the growing conditions at each location have a great influence on accession performance. The variance due to the GEI was also significant for most traits, indicating that the performance of the accessions for these traits was inconsistent, with some showing good performance at one location but failing to replicate such performance at another location. This indicates the need for METs to identify accessions with stable performance for these traits, especially GY, which is the primary target. Similar findings have been reported by Girma et al. (2020) and Welderufael et al. (2024).

In the present study, highly significant differences were observed among the accessions for GY, nutritional traits, and most agronomic traits, suggesting that the accessions performed differently for these traits, with some performing well in one trait and others performing well in multiple traits. The best performing accessions for each trait can thus be selected for used in future breeding program. As such, accessions LR55, DR25, and LR2 could be selected for their high yields, LR55, DR38, and DR25 for their high CHOs, and LR24, LR2, and DR18 for their high protein contents. Based on their high caloric content, accessions LR2, LR11, and LR24 could also be useful as sources of important dietary supplements for the small households commonly found in this region. In line with our findings, Girma et al. (2020), and Welderufael et al. (2024) also reported high genetic differences for GY and nutritional traits among sorghum accessions.

Heritability was also estimated in this study using the variance component from ANOVA. Heritability estimates are important for identifying the major driver of the observed variation among the accessions, whether genetic or environmental, and is used to guide future breeding programs. According to Patil and Lokesha (2018), heritability estimates should be used in conjunction with genetic advancements to achieve high genetic gain from selection. A high heritability estimate (broad-sense) and GCV for a trait in a population indicates that the performance of genotypes in the population for that trait will be based on their inherent genetic potential with minimal environmental effects; therefore, traits exhibiting such a performance can be improved through simple selection (Holland et al., 2002; Mukondwa et al., 2020; Sobhaninan et al., 2019). Based on this, the high heritability estimates for all traits in this study indicate that the accessions identified as best performing for the studied traits can be reliably selected for use in future breeding programs. These accessions should be able to replicate the high performance and transmit the genes responsible for this performance to their progeny.

The yield performance of a genotype remains the primary focus of both breeders and farmers. For a genotype to be acceptable for cultivation by farmers, it must display both adaptability and stability. Adaptability indicates that it must be high yielding, and stability indicates that it must be able to replicate the high performance across seasons and under different growing conditions (Aboderin et al., 2023; Mafouasson et al., 2018). In this study, accessions LR55, DR25, and LR2 were identified on the GGE biplot as having high yield stability and adaptability. Thus, these accessions can be recommended for continuous cultivation by farmers in the Nigerian SGS due to their high yield, moderate to high yield stability, and adaptability.

Some accessions recorded a high yield performance but poor nutritional values and vice versa. The GYT biplot was used to select accessions with superior performance based on their nutritional values combined with yield. The “Which-Won-Where” view of the biplot identified accession LR2 as the best for combining GY with the caloric, ash, lipid, and phenolic contents. DR15 and DR18 recorded the best values for GY and protein, moisture, and HCN contents, while LR55 was the best accession combining GY with CHO and the phytate content. However, the “mean vs. stability” view of the GYT biplot revealed the overall superiority of the accessions in different GY and nutritional traits combinations. LR2, DR18, and DR15 recorded the best values among the accessions with high values for yield and nutritional traits combinations. These accessions represent valuable resources for both cultivation and as elite genetic materials in future breeding programs.

The relationships among agronomic traits and nutritional traits were also investigated in this study. From the results of both Pearson correlation coefficients and the GT biplot, LW, GC, PH, HSW, and TSW were positively associated with GY, indicating that improvement in any of these traits leads to an improvement in GY. Additionally, these traits can also be used as selection indices for the early identification of sorghum accessions with genetic potential for a high GY. Similar findings have been reported by Mukondwa et al. (2020) and Welderufael et al. (2023) in their studies on trait relationships using different sorghum germplasm.

Relationships observed between GY and some agronomic traits in this study deviate from those previously reported in studies involving cereal crops. For example, the negative correlation observed between DF and yield-related traits is surprising, as it is widely known that cereal varieties that take a longer time to flower and attain maturity are usually high yielding compared to early maturing varieties (Bankole and Aboderin, 2024; Bello et al., 2012; Oluwaranti et al., 2008). Similarly, the negative correlations of GY with LL, PHT, and IL also contrast with previous studies by Bello et al. (2012) and Pedersen et al. (2022), who reported positive associations between GY and increased vegetative growth traits in cereals. The implication of these results is that these traits cannot be reliably used as selection indices for high yield performance in sorghum.

Significant negative and positive associations were observed among the nutritional and antinutritional traits. The positive correlation observed between the lipid content and ash content, phenolic content and lipid content, and caloric content and CHO indicate the possibility of improving these traits concurrently in breeding programs. However, the negative correlations observed between CHO and the fibre content and between CHO and the protein content indicate potential trade-offs in which the improvement in one leads to a negative effect on the other. Breeders must carefully balance these trade-offs depending on their target consumers. The correlations between most yield-related traits and nutritional qualities were largely not significant, suggesting that breeding for improved GY in these accessions will not necessarily affect (increase or decrease) their nutritional quality. Thus, both traits can be improved independently since they are not strongly linked.

 

CONCLUSIONS

This study shows that the collected sorghum accessions harbour useful genetic diversity. The high heritability estimates for all traits in this study indicate that the accessions identified as best performing for the traits can be reliably selected for use in future breeding programs. Based on their high yield performance, stability, and adaptability, accessions LR55, DR25, and LR2 are recommended for continued cultivation by farmers in the Nigerian SGS. LR2, DR18, and DR15 are recommended for either direct cultivation or as genetic donors in breeding programs due to their high nutritional value and high yield. LW, GC, PH, HSW, and TSW, which were positively associated with GY and had high heritability estimates, are recommended as selection indices for the early identification of high-yielding sorghum accessions.

 

Author contributions: Conceptualization and methodology: FAB, OSA, AOK, OO; Investigation: FAB, OO; Data analysis: FAB, OSA; Data curation: FAB, OSA; Writing – original draft preparation: FAB, OSA; Writing – review and editing: FAB, OSA, AOK; Supervision: FAB. All authors declare that they have read and approved the publication of the manuscript in this present form.

Acknowledgments: We sincerely thank the staff and members of the University of Ilorin Teaching and Research Farm for their invaluable assistance throughout the study.

Data availability statement: The data presented in this study are available on request from the corresponding authors.

Conflicts of interest: The authors report there are no competing interests to declare.

 

REFERENCES

Aboderin, O.S.; Bankole, F.A.; Oyekunle, M.; Olaoye, G. Yield stability and inter-trait relationships of maize hybrids under low- and optimum-nitrogen conditions. Agriculture (Poľnohospodárstvo) 2023, 69(4), 171-185. https://doi.org/10.2478/agri-2023-0015

AOAC. Official Methods of Analysis of AOAC INTERNATIONAL; 21st ed.; AOAC International: Washington DC, USA, 2019.

Assefa, A.; Bezabih, A.; Girmay, G.; Alemayehu, T.; Lakew, A. Evaluation of sorghum (Sorghum bicolor (L.) Moench) variety performance in the lowlands area of Wag Lasta, north eastern Ethiopia. Cogent Food & Agriculture 2020, 6(1), 1778603. https://doi.org/10.1080/23311932.2020.1778603

Bankole, F.A.; Aboderin, O.S. Yield performance and stability of maize hybrids of different maturity groups in multiple environments in Nigeria. Thailand Journal of Agricultural Science 2024, 57(3), 164-180.

Bello, O.B.; Abdulmaliq, S.Y.; Ige, S.A.; Mahamood, J.; Oluleye, F.; Azeez, M.A.; Afolabi, M.S. Evaluation of early and late/intermediate maize varieties for grain yield potential and adaptation to a Southern Guinea Savanna agro-ecology of Nigeria. International Journal of Plant Research 2012, 2(2), 14-21. https://doi.org/10.5923/j.plant.20120202.03

Day, R.A.; Underwood, A.L. Quantitative Analysis. 5th Edition, Prentice Hall Publication, Upper Saddle River, 1986, 701.

Elangovan, M.; Ganesamurthy, K.; Rajaram, S.; Sankarapandian, K.; Babu, K.P. Collection and conservation of sorghum landraces from Tamil Nadu. Electronic Journal of Plant Breeding 2012, 3, 753-762.

Feyzbakhsh, M.T.; Payghamzadeh, K.; Khazari, A. Genotype selection of Sorghum Bicolor L. based on interrelationship of quantitative traits using genotype × trait and genotype by yield × trait biplots. Iranian Journal of Genetics and Plant Breeding 2020, 9(2), 83-93.

Girma, G.; Nida, H.; Tirfessa, A.; Lule, D.; Bejiga, T.; Seyoum, A.; Mekonen, M.; Nega, A.; Dessalegn, K.; Birhanu, C. A comprehensive phenotypic and genomic characterization of Ethiopian sorghum germplasm defines core collection and reveals rich genetic potential in adaptive traits. The Plant Genome 2020, 13(3), e20055. https://doi.org/10.1002/tpg2.20055

Govindaraj, M.; Vetriventhan, M.; Srinivasan, M. Importance of genetic diversity assessment in crop plants and its recent advances: An overview of its analytical perspectives. Genetics Research International 2015, 431487. https://doi.org/10.1155/2015/431487

Hegde, S.; Thangalakshmi, S.; Singh, R. A review of gluten and sorghum as a gluten-free substitute. Trends in Horticulture 2023, 6 (2), 1-17. https://doi.org/10.24294/th.v6i2.2840

Holland, J.B.; Nyquist, W.E.; Cervantes-Martínez, C.T. Estimating and interpreting heritability for plant breeding: an update. In Plant Breeding Reviews. John Wiley & Sons, Inc. 2002, 22, 9-112. https://doi.org/10.1002/9780470650202.ch2

Johnson, W.W.; Robinson, H.F.; Comstock, R.E. Genotypic and phenotypic correlation in soybeans and their implications in selection. Agronomy Journal 1955, 47, 477-482. https://doi.org/10.2134/agronj1955.00021962004700100008x

Kavithamani, D.; Yuvaraja, A.; Selvi, B. Principal component analysis and grouping of sorghum (Sorghum bicolor L. Moench) gene pool for genetic diversity. Electronic Journal of Plant Breeding 2019, 10(4), 1426-1434. https://doi.org/10.5958/0975-928X.2019.00182.0

Kayode, R.M.; Sanni, A.; Oladoye, C.; Annongu, A.A.; Arekemase, M.O.; Obalowu, M.A.; Abdulsalam, K.O. Physico-chemical and anti-nutritional characterization of the kernels of some mango (Mangifera indica) cultivars grown in western parts of Nigeria. Food Science and Quality Management 2013, 22, 1-8.

Mafouasson, H.N.A.; Gracen, V.; Yeboah, M.A.; Ntsomboh-Ntsefong, G.; Tandzi, L.N.; Mutengwa, C.S. Genotype-by-environment interaction and yield stability of maize single cross hybrids developed from tropical inbred lines. Agronomy 2018, 8(5), 62. https://doi.org/10.3390/agronomy8050062

Makkar, H.P.S.; Norsambuu, T.; Lkhavatsere, S.; Becker, K. Plant secondary metabolites in some medical plants of Mongolia used for enhancing animal health and production. Tropicultura 2009, 27 (3), 159-167.

Maxson, E.D.; Rooney, L.W. Evaluation of methods for tannin analysis in sorghum grain. Cereal Chemistry 1972, 49(6), 719-728.

McKie, V.; McCleary, B. A novel and rapid colorimetric method for measuring total phosphorus and phytic acid in foods and animal feeds. Journal of AOAC International 2016, 99(3), 738-743. https://doi.org/10.5740/jaoacint.16-0029

Merrick, L.F.; Glover, K.D.; Yabwalo, D.; Byamukama, E. Use of Genotype by Yield*Trait (GYT) analysis to select hard red spring wheat with elevated performance for agronomic and disease resistance traits. Crop Breeding, Genetics and Genomics 2020, 2(1), e200009. https://doi.org/10.20900/cbgg20200009

Mukondwa, O.; Manjeru, P.; Ngirazi, S.; Mavankeni, B.O.; Kutywayo, D.; Kamutando, C.N. Genotype-by-trait association of sorghum (Sorghum bicolor L. Moench) advanced lines grown under arid and semi-arid regions of Zimbabwe. Journal of Crop Science and Biotechnology 2020, 24, 760. https://doi.org/10.1007/s12892-020-00060-7

Nemukondeni, N.; Mbajiorgu, C.A.; Hassan, Z.M.; Sebola, N.A.; Manyelo, T.G.; Bodede, O.; Mabelebele, M. Physical characteristics, nutritional composition and phenolic compounds of some of the sorghum landraces obtained in South Africa. Food Research 2022, 6(4), 312-328. https://doi.org/10.26656/fr.2017.6(4).555

Oluwaranti, A.; Fakorede, M.; Badu-Apraku, B. Grain yield of maize varieties of different maturity groups under marginal rainfall conditions. Journal of Agricultural Sciences Belgrade 2008, 53(3), 183-191. http://dx.doi.org/10.2298/JAS0803183O

Pacheco, A.; Vargas, M.; Alvarado, G.; Rodríguez, F.; Crossa, J.; Burgueño, J. GEA-R (Genotype × Environment Analysis with R for Windows), Version 2.0; CIMMYT: 2016.

Patil, M.; Lokesha, R. Estimation of genetic variability, heritability, genetic advance, correlations and path analysis in advanced mutant breeding lines of Sesame (Sesamum indicum L.). Journal of Pharmacognosy and Natural Products 2018, 4, 151. http://dx.doi.org/10.4172/2472-0992.1000151

Pedersen, I.F.; Christensen, J.T.; Sørensen, P.; Christensen, B.T.; Rubæk, G.H. Early plant height: A defining factor for yields of silage maize with contrasting phosphorus supply. Soil Use and Management 2022, 38, 537-548. https://doi.org/10.1111/sum.12697

Peixoto, M.A.; Evangelista, J.S.P.C.; Coelho, I.F.; Carvalho, L.P.; Farias, F.J.C.; Teodoro, P.E.; Bhering, L.L. Genotype selection based on multiple traits in cotton crops: The application of Genotype by Yield*Trait biplot. Acta Scientiarum. Agronomy 2022, 44(1), e54136. https://doi.org/10.4025/actasciagron.v44i1.54136

SAS Institute Inc. SAS/STAT Software 9.2; SAS Institute: Cary, NC, USA, 2008.

Shahriari, Z.; Heidari, B.; Dadkhodaie, A. Dissection of genotype × environment interactions for mucilage and seed yield in Plantago species: Application of AMMI and GGE biplot analyses. PLoS ONE 2018, 13(5), e0196095. https://doi.org/10.1371/journal.pone.0196095

Shariatipour, N.; Heidari, B.; Shams, Z.; Richards, C. Assessing the potential of native ecotypes of Poa pratensis L. for forage yield and phytochemical compositions under water deficit conditions. Scientific Reports 2022, 12(1), 1121. https://doi.org/10.1038/s41598-022-05024-1

Sobhaninan, N.; Heidari, B.; Tahmasebi, S.; Dadkhodaie, A.; McIntyre, C.L. Response of quantitative and physiological traits to drought stress in the SeriM82/Babax wheat population. Euphytica 2019, 215, 32. https://doi.org/10.1007/s10681-019-2357-x

Tasie, M.M.; Gebreyes, B.G. Characterization of nutritional, antinutritional, and mineral contents of thirty-five sorghum varieties grown in Ethiopia. International Journal of Food Science 2020, 1-11. https://doi.org/10.1155/2020/8243617

Welderufael, S.; Abay, F.; Ayana, A.; Amede, T. Genotype by Trait (GT) and Genotype by Yield × Traits (GYT) Analysis of Sorghum Landraces in Tigray, Northern Ethiopia. Crop Breeding, Genetics and Genomics 2023, 5(2), e230002. https://doi.org/10.20900/cbgg20230002

Welderufael, S.; Abay, F.; Ayana, A.; Amede, T. Genetic diversity, correlation and Genotype × Yield × Trait (GYT) analysis of grain yield and nutritional quality traits in sorghum (Sorghum bicolor [L.] Moench) genotypes in Tigray, Northern Ethiopia. Discover Agriculture 2024, 2(1). https://doi.org/10.1007/s44279-024-00016-3

Yan, W.; Tinker, N.A. Biplot analysis of multi-environment trial data: Principles and applications. Canadian Journal of Plant Science 2006, 86, 623-645. http://dx.doi.org/10.4141/P05-169

Yan, W.; Frégeau-Reid, J. Genotype by Yield*Trait (GYT) biplot: A novel approach for genotype selection based on multiple traits. Scientific Reports 2018, 8(1), 1-10. https://doi.org/10.1038/s41598-018-26688-8

Yan, W.; Frégeau-Reid, J.; Mountain, N.; Kobler, J. Genotype and management evaluation based on Genotype by Yield*Trait (GYT) analysis. Crop Breeding, Genetics and Genomics 2019, 1, e190002. https://doi.org/10.20900/cbgg20190002

 

Academic Editor: Dr. Iulian GABUR

Publisher Note: Regarding jurisdictional assertions in published maps and institutional affiliations ALSE maintain neutrality.