Glutinous rice (Oryza sativa L.) holds a paramount position in Vietnam’s agricultural landscape, deeply intertwined with traditional culture and regional food security. Over generations, indigenous landraces have been selectively preserved by local farming communities, evolving distinctive adaptations to specific agro-ecological niches (Rana et al., 2009). However, the genetic diversity of these specialty landraces faces severe erosion due to the widespread adoption of high-yielding improved cultivars (Schiller et al., 2006). Characterizing and conserving remaining indigenous germplasm is therefore imperative to prevent permanent genetic loss and to secure valuable foundational material for future crop improvement (Begum et al., 2018; Acevedo-Siaca et al., 2020; Ivin and Anbuselvam, 2025).
       
Historically, the evaluation of crop diversity relied predominantly on agro-morphological traits, which remain a standard, cost-effective method for preliminary identification (Watson and Eyzaguirre, 2002; Guru et al., 2017). Nevertheless, phenotypic expressions are frequently confounded by environmental interactions. To overcome this limitation, Simple Sequence Repeat (SSR) markers have become an indispensable tool in modern plant breeding. Characterized by high polymorphism, co-dominance and high reproducibility, SSR markers enable precise assessment of genetic structures and evolutionary relationships in rice (Virk et al., 2000; Lee et al., 2011; Singh et al., 2016). Recent studies during the 2018-2024 period have further highlighted that genomic profiling of indigenous landraces provides irreplaceable insights into local adaptation, drought/salinity resilience and unique grain quality traits that modern cultivars frequently lack (Begum et al., 2018; Hue et al., 2018; Lei et al., 2021; Tran et al., 2021). Integrating both phenotypic characterization and molecular genotyping has therefore become the benchmark approach for identifying genuine accession identities, eliminating duplicate germplasm and establishing core conservation collections (Acevedo-Siaca et al., 2020; Lang et al., 2023; Singh et al., 2025).
       
In this study, we evaluate the genetic diversity and agro-morphological characteristics of the local “Cút” glutinous rice (historically designated as King’s glutinous rice) from Quang Ngai province, Vietnam. This unique landrace is historically tied to the ancient Tu Nghia archaeological region, representing a valuable component of local agricultural heritage. By integrating comprehensive field morphometrics with polymorphic SSR marker analysis, this research aims to establish a scientific foundation for in-situ and ex-situ conservation strategies. Furthermore, the insights gained will assist plant breeders in identifying elite initial materials to optimize yield, improve grain quality and support sustainable agricultural development in Central Vietnam.

Plant materials and field trials
 
The plant materials comprised 104 local glutinous rice (Oryza sativa L.) accessions, with 100 samples collected directly from traditional on-farm fields of local farming households across Nghia Ky and Nghia Thuan communes (Tu Nghia district, Quang Ngai province) and 4 samples collected from indigenous ethnic farming valleys in Vinh Thanh district (Binh Dinh province), Vietnam (Table 1). The overall research timeframe spanned 2022 to 2025: germplasm exploration, on-site farmer interviews and panicle collection were performed during 2022–2023; subsequent multi-season field evaluations were executed from 2023 to 2025 at a single experimental field in Nghia Ky commune, Tu Nghia district, Quang Ngai province (15°07'42.8''N, 108°44'14.1''E), in collaboration with Le Khiet High School for the Gifted and local agricultural stations.


       
The field trials were laid out in a randomized complete block design (RCBD) with three replications. Each experimental plot measured 5 m², with a hill spacing of 20 cm × 20 cm (transplanting 12-15-day-old single seedlings per hill). Crop management, including basal and top dressing fertilization (5 tons manure + 90 kg N + 50 kg P₂O₅ + 60 kg K₂O per hectare) and standard water management, strictly adhered to the Bioversity International and IRRI descriptor standards (Bioversity International et al., 2007). Molecular genotypic characterization and SSR-PCR analyses were performed at the Crop Science, Biochemistry and Biotechnology Laboratory (Quy Nhon University) and the Institute of Biotechnology (Hue University, Vietnam).
 
Agro-morphological characterization
 
At the stage of physiological maturity, phenotypic characteristics were documented in accordance with the standard evaluation system for rice (IRRI, 2014). Measurements were taken for eleven quantitative traits: growth duration (GT, days), plant height (PH, cm), effective tiller count per plant (ET/P), panicle length (PL, cm), panicle count per plant, filled spikelet count per panicle (NFS/P), empty spikelet percentage (ESR, %), weight of 1000 grains (1000 W, g), biomass (B, g), yield (Y, g/hill) and harvest index (HI, %).
       
Grain yield per hill was calculated at 14% moisture content using the formula:

Where the moisture factor (MF) is derived from the grain moisture content (MC, %) as:

 
The harvest index was calculated as follows:

 

The economic yield represents the total grain weight of 10 sampled plants and the biological yield includes both grain weight and straw biomass.
 
DNA extraction and SSR-PCR analysis
 
Total genomic DNA was isolated from fresh, frozen young leaf tissues following the modified cetyltrimethylammonium bromide (CTAB) protocol reported by Oliveira et al., (2015). The integrity and concentration of the extracted DNA were subsequently evaluated through 1% agarose gel electrophoresis.
       
A total of 108 simple sequence repeat (SSR) loci spanning the 12 rice chromosomes were chosen using the Gramene database. Polymerase chain reactions (PCR) were set up in a 10 μL volume consisting of 20 ng genomic DNA, 0.2 mM forward/reverse primer pairs and 0.5 U Taq DNA Polymerase (Merck KGaA, Germany). Amplification conditions were programmed as follows: initial denaturation at 94°C for 5 min; 35 cycles of  (30 s),  (30 s) and  (30 s) and a final 10-min extension step at 72°C. The resulting PCR products were separated on 3% polyacrylamide gels, ethidium bromide-stained and observed under ultraviolet light.  
 
Data analysis
 
The collected agro-morphological dataset was subjected to descriptive analysis, ANOVA and Pearson correlation using Statistix 10.0 software. Morphological relationships among accessions were quantified by constructing a distance matrix calculated via Euclidean distance coefficients according to Sneath and Sokal (1973).

 

 

In this equation,
X_ki and X_kj= The standardized phenotypic values of the k^th trait recorded for the i^th and j^th accessions, respectively.
       
For the molecular evaluation, SSR alleles were recorded as binary data, with 1 indicating presence and 0 indicating absence. The polymorphic information content (PIC) per locus was subsequently computed following the formula proposed by Weir (1996).

 

 

In this context,
P_ij= The frequency of allele j identified at locus i.
       
To construct the genetic dendrogram, cluster analysis was executed via the Unweighted Pair-Group Method with Arithmetic Mean (UPGMA) utilizing the SAHN module embedded in NTSYS-pc (version 2.11a) software.

Polymorphism of SSR markers and genetic diversity
 
Screening 104 glutinous rice accessions with 108 SSR loci identified 52 highly polymorphic markers, corresponding to a polymorphism rate of 48.15%. Locus profiling produced 214 total alleles, averaging 4.12 per locus (spanning 2 to 7 alleles; Table 2). This average allele number is higher than those reported by Tran et al. (2011) and Doan et al., (2016), but lower than the diversity levels observed in larger regional collections (Hue et al., 2018; Lei et al., 2021). These variations in allelic richness can be attributed primarily to differences in the geographical scale and sample size of the germplasm evaluated. While our collection comprises 104 accessions concentrated within a specialized traditional agro-ecological zone of Central Vietnam, broader collections encompass distinct eco-strains across multiple macro-regions, naturally harboring broader genetic backgrounds. Furthermore, the variation in allele counts among studies may stem from the specific panel and repeat types of SSR markers employed, as well as the resolution capacity of the genotyping platforms utilized. Marker informativeness was confirmed by PIC values ranging from 0.33 (RM333) to 0.86 (RM514) with a mean of 0.60 (Table 2). This elevated mean PIC value reflects significant local genetic variation, which can be attributed to long-term traditional selection pressures by farmers in the specific microclimate of the Tu Nghia region. Empirical studies on on-farm genetic conservation have shown that decentralized seed saving, informal intra-village seed exchanges and diverse micro-ecological niches (such as valley lowlands and mineral thermal springs) act synergistically to maintain high allelic heterogeneity within landrace populations while preventing complete genetic fixation (Watson and Eyzaguirre, 2002; Rana et al., 2009; Tran et al., 2021). Field investigations in this area further confirmed that local growers have maintained heterogeneous seed stocks across households and continuously selected for culinary preferences and stress resilience, thereby driving micro-evolutionary divergence. These polymorphism estimates are comparable to the SSR profiling reported by Singh et al., (2025), who recorded PIC values up to 0.86 across diverse rice accessions, underscoring the discriminatory power of microsatellite loci in landrace differentiation. Consequently, these SSR markers prove highly effective for in-situ conservation tracking and geographical brand protection.


       
At a genetic similarity coefficient threshold of 0.41, UPGMA-based cluster analysis partitioned the 104 accessions into two primary groups, designated as Cluster A and Cluster B (Fig 1). Cluster A represented the predominant share of the germplasm (82.69%), whereas Cluster B encompassed 17.31%. Upon shifting the similarity threshold to 0.55, Cluster B was partitioned into three clear sub-clusters (B1, B2 and B3). This sub-grouping highlights marked genomic differentiation among the local accessions, offering useful options to counteract genetic erosion.


 
Variance in agro-morphological characteristics
 
Descriptive statistics for the 11 quantitative traits demonstrated substantial phenotypic variability, with all parameters exhibiting a normal distribution (p<00.5, Table 3), confirming substantial genetic variation across yield-contributing traits as observed in comparable rice germplasm evaluations (Kumar et al., 2026).


       
The frequency distribution of growth duration (GT) highlights a highly skewed population toward late maturity, as visualized in Fig 2. The GT ranged widely from 129 to 181 days, with a predominant majority (97.12%) classified into the late-maturing group (>130 days) (Table 4). This late-maturing profile contrasts with typical short-to-medium duration cultivars in Northern Vietnam (Doan, 2013). This pronounced late maturity is intimately linked to photoperiod sensitivity, a distinctive physiological hallmark of traditional Asian rice landraces. As a short-day plant, the Cút glutinous rice accessions typically require day lengths falling below a critical photoperiod threshold to trigger panicle initiation and heading. Such strong photoperiod sensitivity reflects evolutionary adaptation to the local agro-ecological rhythms of Central Vietnam, preventing flowering during the severe rainy or flooding period while maximizing vegetative biomass accumulation. However, because strong photoperiod responsiveness limits flexible multi-cropping rotations, future breeding initiatives should select for moderate photoperiod sensitivity and reduced growth durations without compromising grain quality.




       
For plant architecture, Fig 3a demonstrates that 95.19% of the accessions exhibit a tall stature (>120 cm), with plant height (PH) stretching up to 166 cm. While tall statures are advantageous for weed competition and light interception under traditional extensive farming, they pose a high lodging risk, particularly under heavy rain and strong fertilization (Catón et al., 2003; Lang et al., 2014; Guru et al., 2017). Conversely, the effective tiller number per plant (ET/P) follows a classic bell-shaped normal distribution curve centered at 7-8 tillers (Fig 3b), confirming a highly stable medium tillering capacity across the core population (Table 5).




       
Yield-attributing structures displayed extensive phenotypic spreads, all well-documented by their respective frequency distributions in Fig 4. Panicle length (PL) distribution peaked strongly between 24 and 30 cm (Fig 4a), representing a long-to-extra-long panicle profile. This structural advantage directly correlates with the high density of filled grains, where the number of filled spikelets per panicle (NFS/P) predominantly clustered within the 70-130 grain range (Fig 4c). Concurrently, the population maintained an exceptional fertility rate, with the empty spikelet ratio (ESR) heavily skewed toward ultra-low thresholds of 2-10% (Fig 4b). Additionally, the panicle density predominantly ranged between 100 and 120 panicles/m², with a population mean of 101.52 panicles/m² (Fig 4d; Table 3).


       
In terms of productivity, the 1000-grain weight (1000W) followed a tight distribution concentrated at 24-26 g (Fig 4e). This contributed to an individual grain yield (Y) distribution peaking at 36-44 g/hill (Fig 4f), with an average performance (39.62 g/hill) outperforming the “Nep Ngu” landrace in Binh Dinh (36.33 g/hill) (Lang et al., 2023). This strong performance is backed by a robust vegetative biomass (B) centering at 62-68 g (Fig 4g) and a balanced harvest index (HI) shifting around 0.40-0.60 (Fig 4h).
 
Congruence and differentiation between molecular clusters and agro-morphological traits
 
A comparative assessment between the UPGMA dendrogram (Fig 1) and phenotypic clustering revealed both notable correspondences and distinct divergences across the 104 accessions. Accessions assigned to major Cluster A (accounting for 82.69% of the collection) shared a remarkably uniform agronomic profile characterized by tall plant height (mean >145 cm), late heading (growth duration >155 days), robust vegetative biomass (60-68 g) and high 1000-grain weight (24.5-26.2 g). This phenotypic stability reflects long-term farmer selection targeting high vegetative competitiveness and specific eating quality within the Tu Nghia valley lowlands.
       
Conversely, Cluster B (comprising 17.31% of the germplasm, further resolved into sub-clusters B1, B2 and B3) exhibited significant phenotypic divergence from Cluster A. Specifically, the accessions originating from Binh Dinh (BD-101 to BD-104) and distinct peripheral strains clustered within sub-cluster B1, displaying shorter panicle lengths (20-23 cm), reduced 1000-grain weights (22.0-23.5 g) and earlier maturity (129-140 days) compared to the core Nghia Ky and Nghia Thuan strains in Cluster A. Sub-clusters B2 and B3 gathered lines exhibiting intermediate tillering capacity (5-6 tillers/plant) but higher grain harvest indices.
       
Nevertheless, minor non-congruence observed between molecular grouping and certain agro-morphological parameters is expected. While quantitative phenotypic traits are polygenically inherited and strongly subject to environmental and agronomic pressures, SSR markers probe hypervariable, non-coding genomic loci that are selectively neutral and primarily reflect ancestral lineage and historical gene flow via local seed exchanges. Therefore, coupling molecular SSR genotyping with morphological evaluation provides complementary power: SSRs reliably distinguish authentic landrace identities and geographical provenance, whereas agronomic trait matrices pinpoint elite donor accessions for targeted breeding.
 
Correlation among agro-morphological traits
 
Significant phenotypic associations suitable for indirect trait selection were identified through Pearson correlation analysis (Table 6). Specifically, plant height was positively associated with filled spikelets per panicle (r = 0.319*) and panicle length (r = 0.468*), aligning with plant architecture principles in native rice varieties (Lang et al., 2014; Sujata et al., 2025). A notable positive correlation emerged between panicle length and filled spikelet count (r = 0.656), while filled spikelet count was strongly inversely related to the empty spikelet percentage (r = -0.514).


       
Most importantly, single-hill grain yield demonstrated significant positive associations with both filled spikelet count per panicle (r = 0.374*) and 1000-grain weight (r = 0.348*). Yield further showed positive trends with tillering capacity and panicle length, whereas it was inversely associated with growth duration and plant height. These trait relationships suggest an evolutionary trade-off between vegetative growth and reproductive potential in local landraces. Consequently, crop breeding programs aiming to maximize productivity should emphasize selecting lines with optimized filled spikelets per panicle, extended panicles and heavy grain weight.

The phenotypic diversity and genetic structure of 104 local ‘Cút’ glutinous rice accessions were successfully elucidated using polymorphic SSR markers alongside morphological characterization. Molecular profiling underscored high informativeness, resolving the collection into two primary genotypic groups. Agronomic evaluations indicated that despite inheriting traditional landrace features-specifically late maturity and tall stature-the germplasm exhibits promising yield potential underpinned by large 1000-grain weights and dense filled spikelets.
       
These findings provide a vital scientific foundation for establishing targeted in-situ and ex-situ conservation frameworks to mitigate genetic erosion. Furthermore, this diverse gene pool serves as elite foundational material for future crop improvement programs. Future breeding strategies should focus on selecting high-yielding lines optimized for shorter growth durations and lodging resistance to promote sustainable agricultural systems and secure the regional geographic brand of this valuable agricultural heritage in Central Vietnam.

This paper presents findings from “Investigation, collection, evaluation and conservation of genetic recourses of Sa Huynh’s Ngu glutinous rice and local Cut glutinous rice in Quang Ngai province” project. We thank the Quang Ngai People’s Committees and Department of Science and Technology for supporting this project under grant number QGT35.ÐT.01/2022.
 
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 sponsorship influenced the design of the study, data collection, analysis, decision to publish, or preparation of the manuscript.