Common bean (Phaseolus vulgaris L.) is one of the most important legume crops worldwide, playing a vital role in human nutrition and food security due to its high protein content and ability to fix atmospheric nitrogen in the soil (Broughton, 2003; Gepts et al., 2008). It is a diploid 2n = 22 (Broughton, 2003) and a predominantly self-pollinating crop with a low frequency of crossing (Burle et al., 2011). It has two distinct gene pools, namely the Mesoamerican and Andean gene pools (Musango et al., 2016). The Andean gene pool genotypes are uniformly large-seeded (> 40 g per 100 seeds), whereas the Middle American ones are characterized as small-seeded (< 25 g) or medium-seeded (25-40 g) ( Singh et al., 1991). Common bean is an extremely diverse crop in terms of cultivation methods, uses, the range of environments to which they have been adapted and morphological variability (Antonio et al., 2025; Basavaraja et al., 2021). P. vulgaris exhibits wide adaptability and is cultivated across diverse agro-ecological zones, ranging from tropical to temperate regions (Blair et al., 2010). Genetic diversity in common bean genotypes is a key factor determining the potential for improving yield and yield related traits for bean genotypes (Kwak and Gepts, 2009; Mamidi et al., 2013; Singh, 1999).
       
Analysis of genetic variation is essential for under-standing the basic genetics of common bean (Zargar et al., 2014). It would be useful for determining the morphological variation among the gene pool (Basavaraja et al., 2021), thereby supporting the selection of appropriate genotypes for breeding programs (Kwak and Gepts, 2009). However, the number of studies on genetic variability and agromorphological traits of common bean genotypes in specific climatic regions such as Siberia remains limited. The Priobia steppe region of Siberia is characterized by a harsh climate with a short growing season and low temperatures, requiring common bean genotypes to possess special adaptive traits to ensure productivity (Ivanov and Smolenskaya, 2017). Therefore, evaluation of genetic variation and morphological traits of local common bean genotypes in this region is essential for developing cultivars adapted to these specific environmental conditions. Therefore, this study aims to: (i) assess the genetic variability, heritability and genetic advance of key agronomic traits of common bean under these conditions; (ii) evaluate phenotypic and genotypic correlations among traits; and (iii) apply path analysis to clarify the direct and indirect effects of traits on grain yield. This study aimed to estimate the genetic variability and association of quantitative traits among genotypes.

Treatments and experimental design
 
The present investigation involving 26 common bean genotypes (Table 1) was sown from May to September at the experimental farm of the Novosibirsk State Agrarian University over two seasons (2023 and 2024). The soil in the experimental field is gray forest heavy loam. It presents the following physical, chemical features: humus content 4.5%, mobile phosphorus 9.8-2.8 mg/100 g, mobile potassium 6.2-6.4 mg/100 g, nitrate nitrogen 6-10 mg/kg and ammonia nitrogen 14.2-15.9 mg/kg. The sum of absorbed bases 30.8-52.0 mg-eq. per 100 g of soil, reaction pH 6.0-6.5 (Yakubenko and Parkina, 2019). The experiment was conducted under rainfed conditions, without irrigation or fertilization throughout the growing season. Weather conditions varied greatly during the different growing seasons. During 2023, the weather was rather dry compared to the long-term average, with low average precipitation (38 to 47 mm) and average temperatures ranging from 12 to 24^oC, whereas 2024 was wetter (66 mm to 113 mm) and average temperatures ranging from 10 to 22^oC. During the experiment, the average temperature tended to gradually increase from May to July and gradually decrease from August to September (Fig 1). The experiment was conducted using a randomized block design with three replications. The experimental field was 80 m in length and 2 m in width. Plots were 200 cm in length, 70 cm in width and 70 cm apart. Each genotype was sown in three rows of 200 cm long, with an inter-plant spacing of 6 cm and inter-row spacing of 70 cm. Thirdty five seeds were planted in each row.




 
Data collection and meaasurements
 
Plant parameters recorded were days to 50% flowering (DF), days to maturity (PM), plant height (PH), number of nodes per stem (NNS), number of pods per plant (NPP), pod length (PL), pod width (PW), number of seeds per pod (NSP), grain yield per plant (SYP) and hundred seed weight (HSW) was recorded on five plants of each genotype which were chosen randomly, with border plants excluded. Pest and disease monitoring was conducted throughout the growing season. Weeds were removed manually and plants showing symptoms of pest or disease infestation were excluded from further measurements.
       
The average mean for each trait over three replications was computed for each genotype and analysed statistically to determine analysis of variance (ANOVA) was performed using IRRISTAT software version 5.0 (IRRI). The phenotypic coefficient of variation (PCV) and genotypic coefficient of variation (GCV) were estimated using the approach given by Burton (1952).
 

Where
σ²g = Genotypic variance.
σ²p = Phenotypic variance.
x = Sample mean.
       
Genetic advance as per cent of mean (GAM%) were calculated using the formula given by Johnson and Comstock (1955).
 

 

Where
K = 2.06 at 5% selection intensity.
H = Heritability.
δp = Phenotypic standard deviation.
x = Sample mean.
       
Broad sense heritability (H²) was assessed as per the procedure given by Falconer (1989).

 

Genotypic (r_g) and phenotypic correlation coefficient (r_p) were computed using procedure recommended by Miller et al., (1983). Path coefficient analysis was done according to Dewey and Lu (1959). Standard path coefficients which are the standardized partial regression coefficients were obtained using statistical software packages OPSTAT (Pal et al., 2017).

Genetic variation of common beans
 
Analysis of variance (ANOVA) showed significant differences between genotypes for all traits under investigation (Table 2). This variation in genotypes may be a result of genotypic diversity, environmental effect and their interaction (Sonali et al., 2025). Estimates of heritability and genetic advances are important preliminary steps in the breeding program as they provide information needed in designing the effective breeding program and the relative practicability of selection (Sadeghi et al., 2011). The results indicated that the phenotypic coefficient of variability (PCV) were higher than the genotypic coefficient of variability (GCV) for all the traits (Basavaraja et al., 2021; Madakbas and Ergin, 2011), which reflect the influence of environment on the expression of all traits.


       
According to Johnson et al., (1955), estimated genotypic coefficient of variation (GCV) and phenotypic coefficient of variation (PCV) are categorised as: low (<10%); medium (10-20%) and high (>20%). Both high genotypic coefficient of variation (GCV) and phenotypic coefficient of variation (PCV) were observed for NNS (27.19% and 27.74%), NPP (57.85% and 61.42%), SYP (22.32% and 27.03%), HSW (36.41% and 36.44%) and PL (21.98% and 22.09%). The remaining traits recorded moderate to low GCV estimates. 

The broad sense heritability is an important genetic parameter that reflects the relationship between genotype and phenotype of a trait. According to Dabholkar (1992), the broad sense heritability values are classified as high (>60%), medium (30-60%) and low (<30%). Except for DF and PH, which showed a moderate heritability (53.29% and 53.90%), all other traits exhibited high heritability. This result is consistent with previous studies, which showed that most traits of common bean possess high heritability values (Eyuel et al., 2022; Nigussie et al., 2020).
       
Genetic advance (GAM) serves as an important indicator for predicting the potential improvement of traits through selection, enabling breeders to identify and prioritize traits with the greatest potential for genetic gain. Genetic advance as a percent of the mean ranged from 9.44% for PH to 112.25% for NPP (Table 3). This result indicated that selecting the top 5% of the genotypes could result in an advance of 9.44-112.25% over the respective population mean.


       
According to Johnson et al., (1955), the genetic advance values are classified as high (>20%), medium (10-20%) and low (<10%). High GAM values indicate additive gene effect, whereas low GAM values indicate non-additive gene effect (Singh and Narayanan, 1993). Except for DF, PM and PH, which showed low to moderate genetic advance, all other traits exhibited high genetic advance, ranging from 28.47 to 112.25%.
       
The results indicated relatively high broad-sense heritability and genetic advance for traits such as NNS, NPP, NSP, SYP, HSW and PL. These findings suggest that a considerable proportion of the phenotypic variation in these traits is genetically controlled, possibly with some contribution from additive effects and that selection could be effective for improving these traits. The present results concur with the findings of Simon, Gobeze and Mebede (2020).
       
The high heritability (68.07) associated with moderate genetic advance (11.94) for PM and PW. This suggests that both additive and non-additive gene effects contribute to the genetic control of these traits. The moderate broad-sense heritability (53.29) associated with low genetic advance (9.44) for DF. This indicates the predominance of non-additive gene effects in the genetic control of this trait. Out of all the traits under study, five traits viz., NNS, NPP, SYP, HSW and PL recorded maximum values for heritability (h²), GCV and GAM% thus, depicting the effect of additive gene action on these traits and therefore, may be helpful for efficient selection.
 
Correlation coefficients
 
In plant genetics and breeding studies, correlated traits are important because of genetic causes of correlations through pleiotropic action or developmental interactions of genes and changes can be brought in correlated traits either by natural or artificial selection (Belay et al., 2024; Sharma, 1998). Crop improvement programs can be more effective through understanding the relationships between yield and its component traits. Indirect improvement of a target trait may occur through the enhancement of another trait if they are positively correlated (Remzeena and Anitha, 2021).
       
DF (rg = 0.34), NPP (rg = 0.64), NSP (rg = 0.51), HSW (rg = 0.62) and PL (rg = 0.35) exhibited significant and positive genotypic association with grain yield where the highest genotypic correlation coefficient was recorded for NPP. This indicates that the genotypes with high number of pods per plant, number of seeds per pod, hundred seed weight, pod length and days to 50% flowering were producing higher grain yield. The results are in line with previous research, in which Sadeghi et al., (2011) reported highly significant correlations of seed yield with number of seeds per pod, number of pods per plant, days to flowering and hundred seed weight. Number of pods per plant showed a significant positive association with plant height (rg = 0.5), number of nodes per stem (rg = 0.52), hundred-seed weight (rg = 0.71) and pod length (rg = 0.47). Pod length with DF (rg = 0.67), PH (rg = 0.61), NPP (rg = 0.47) and NSP (rg = 0.54) had significant positive association.
       
Phenotypic correlation coefficients for agronomic traits ranged from -0.5 to 0.82 (Table 4). Seed yield per plant had significant positive phenotypic correlation with DF (rp = 0.34), PM (rp = 0.39), PH (rp = 0.38), NPP (0.75), HSW (rp = 0.7) and PL (rp = 0.8). Number of pods per plant showed a significant positive association with DF (rp = 0.53), PM (rp = 0.42), PH (rp = 0.42), NNS (rp = 0.65) and HSW (rp = 0.63). Pod length with DF (rp = 0.41), PM (rp = 0.52), PH (0.34) and HSW (rp = 0.8) had significant positive association.


 
Path co-efficient analysis
 
Path co-efficient analysis is an efficient approach to separate correlation coefficients into direct and indirect component effects since it assesses the direct impact of one variable on the other (Sonali et al., 2025). Correlation studies give a greater understanding of the causes and effects of relationships between different pairs of component traits and the main trait when paired with path coefficient analysis (Verma et al., 2021). In our study, we considered seed yield per plant (SYP) as the dependent variable and the other traits as independent ones. Multicollinearity among independent traits was tested using the Variance Inflation Factor (VIF). All VIF values were below 5, indicating no multicollinearity problem among the independent variables in the path analysis model (Table 5).


       
The results revealed that NPP had maximum positive (0.58) and direct effect on SYP followed by HSW (0.54), DF (0.27), NNS (0.09), PH (0.07), NSP (0.06) and PL (0.04) suggesting that they are the major contributors to seed yield per plant. Daniel (2015) had reported that path coefficient at genotypic level showed that number of pods per plant had positive direct influence on grain yield.               

NPP showed a significant positive correlation with SYP (r = 0.64) and had a direct effect coefficient of 0.58; this result indicates that the number of pods per plant directly influences seed yield of common bean genotypes. Similarly, HSW and DF showed significant positive correlations with SYP (r = 0.62 and r = 0.34) and had direct effect coefficients of 0.54 and 0.27, respectively. Therefore, seed yield per plant is directly influenced by hundred seed weight and days to 50% flowering.
       
Pod length showed a significant positive correlation with SYP (r = 0.35) but had a low direct effect coefficient (0.04) and it indirectly influenced seed yield per plant via DF (0.11), NPP (0.12) and NSP (0.15). Similarly, plant height influenced seed yield mainly through indirect effects via the traits DF (0.14), NPP (0.16) and HSW (0.11). The residual effect shows how much the explanatory variables represent the variability of the dependent variable (Singh and Chaudhary, 1985). The residual effect was 0.151, indicating that the independent traits accounted for 97.7% (R² = 0.977) of the total variation in seed yield, while a small proportion (2.3%) remained unexplained by the model.

This study analyzed the genetic variation and evaluated the correlations among quantitative traits related to yield of 26 common bean genotypes collected from the Priobia steppe, Siberia, by estimation of genetic parameters and path coefficient analysis. The results revealed significant genetic variation among genotypes. The high broad-sense heritability and genetic advance observed for traits such as NNS, NPP, NSP, SYP, HSW and PL indicate that these traits are mainly controlled by additive gene effects, offering opportunities for effective selection and genetic improvement.
       
Seed yield per plant showed a significant positive correlation with DF (rg = 0.34 and rp = 0.34), NPP (rg = 0.64 and rp = 0.75), HSW (rg = 0.62 and rp = 0.70) and PL (rg = 0.35 and rp = 0.40). The number of pods per plant, 100-seed weight and days to flowering had a direct influence on seed yield per plant. Pod length and plant height influenced seed yield mainly through indirect effects via days to flowering, number of pods per plant, number of seeds per pod and 100-seed weight. These results may help enhance the effectiveness of conservation and improvement of common bean genetic resources in the Priobia steppe, Siberia.
 

The authors sincerely thank the Ministry of Education and Training of Vietnam and the Ministry of Science and Higher Education of the Russian Federation for granting scholarships for postgraduate studies in the Russian Federation. The authors are also grateful to the Institute of Fundamental and Applied Agrobiotechnologies, Novosibirsk State Agrarian University, for their encouragement and support.
 
Disclaimers
 
The views and conclusions expressed in this article are solely those of the authors and do not necessarily represent the views of their affiliated institutions. The authors are responsible for the accuracy and completeness of the information provided, but do not accept any liability for any direct or indirect losses resulting from the use of this content.

The authors declare that there are no conflicts of interest regarding the publication of this article. No funding or sponsor- ship influenced the design of the study, data collection, analysis, decision to publish, or preparation of the manuscript.