Tay Nguyen Journal of Sciences

Selection and cultivation of endophytic bacteria from durian trees antagonistic to Colletotrichum gloeosporioides in Dak Lak Province

Thi Ha Trang Tran — 2026-08-31

Colletotrichum gloeosporioides is the causal agent of anthracnose in durian and is responsible for significant yield and quality losses, particularly under warm and humid conditions during the rainy season. This study aimed to isolate and screen indigenous endophytic bacteria from durian roots for their antagonistic activity against the pathogen. From 30 durian root samples, 20 bacterial strains were isolated. Antifungal activity was evaluated using the dual culture assay, and the data were analyzed by one-way ANOVA followed by Duncan's multiple range test (p Bacillus amyloliquefaciens based on 16S rRNA gene sequencing. Under flask culture conditions (36 °C, pH 7, 200 rpm), strain EB.KN18 reached a cell density of 95.0×10⁸ CFU/mL after 48 h. When cultivated in a 14 L bioreactor, the bacterial density reached 7.08×10¹¹ CFU/mL after 10 h. Bacillus amyloliquefaciens EB.KN18 is a promising candidate for the development of biological products, contributing to safe and sustainable durian cultivation.

Development of an automated grading system based on Office Open XML, rubric JSONs, and AI-assisted feedback for a general informatics course

Phan Thi Dai Trang — 2026-08-31

This study presents the development and evaluation of an Automated Grading System (AGS) for Microsoft Word, Excel, and PowerPoint practical assignments in a General Informatics course. The system analyzes the Office Open XML structures of .docx, .xlsx, and .pptx files, normalizes the extracted data into a JSON-based abstract syntax tree (JSON AST), locates assessment objects using an Anchor Locator, and executes criteria represented as JSON Rubrics through a Rule Engine. The workflow is orchestrated on the n8n platform, while a large language model is used only to interpret results and generate feedback rather than to determine scores directly. The initial dataset comprised 1,374 submissions. After invalid and duplicate files were removed, 1,257 submissions were processed, of which 1,247 complete instructor–AGS score pairs were included in the analysis. The overall mean absolute error was 0.618 points on a 10-point scale; 84.4% of the automated scores differed from the instructor’s scores by no more than one point, and the mean bias was −0.051 points. Excel produced the lowest absolute discrepancy. Analysis of 204 cases requiring review showed that discrepancies were mainly associated with the system’s failure to recognize valid alternative implementations or with grading rules that did not fully verify the required techniques. The findings demonstrate the feasibility and traceability of a unified grading architecture while also highlighting the continued importance of instructor oversight for exceptional cases.

A mobile-based course grade management system with a multi-layer security architecture integrating hybrid OCR

Truong Hai — 2026-08-31

Digital transformation in higher education requires course grade management systems that provide efficient, secure, and mobile-based data processing. This study proposes a mobile course grade management system with a multi-layer security architecture integrating Hybrid OCR to support the automatic recognition of handwritten scores from paper-based grade sheets. The proposed system consists of a Gateway API Server, Database Server, AI Server, and a mobile application. Hybrid OCR combines EasyOCR and a Convolutional Neural Network (CNN) for handwritten score recognition, while Human Verification and Feedback Learning enable instructors to validate recognition results and accumulate verified samples for future model retraining. The multi-layer security architecture strengthens device authentication, session authentication, and secure data communication. Experimental results on real course grade sheets collected at Tay Nguyen University show that the CNN achieved an accuracy of 95.17% on the independent test set. The multiprocessing architecture reduced the average processing time to 4.19 seconds per grade sheet, representing an improvement of approximately 40.3% compared with sequential processing. The system operated reliably while preserving lecturer verification before official grade submission (Tay Nguyen University, 2025). The proposed system demonstrates the feasibility of deploying secure mobile-based course grade management in credit-based education environments (Tay Nguyen University, 2021) and provides a foundation for continuous model improvement through Feedback Learning.

Propagation of black ginger (Distichochlamys orlowii K.Larsen & M.F.Newman) via thin cell layer (TCL) technology

Nguyen Van Tinh¹ — 2026-08-31

Black ginger (Distichochlamys orlowii K.Larsen & M.F.Newman) is a rare and endemic medicinal plant of Vietnam, possessing various valuable biological activities such as anti-inflammatory, antioxidant, and anticancer properties. This study applied thin cell layer technology to establish an efficient in vitro micropropagation protocol. Species identification using DNA barcoding with the matK gene confirmed that the experimental specimens belonged to D. orlowii. Transverse thin cell layers (tTCLs) derived from stem nodes (0.5 mm in thickness) exhibited the highest callus induction frequency (100%) on Murashige and Skoog (MS) medium supplemented with 1.0 mg/L 2,4-dichlorophenoxyacetic acid (2,4-D) and 0.7 mg/L benzyladenine (BA). Optimal shoot regeneration from callus (95%) was achieved on MS medium containing 1.5 mg/L BA and 0.5 mg/L Knetin (Kn). The highest shoot multiplication coefficient (10.0 shoots per cluster) was recorded upon adding 0.5 mg/L alpha-naphthaleneacetic acid (NAA) to the regeneration medium. Complete in vitro rooting (100%) was successfully induced on MS medium supplemented with 1.0 mg/L NAA. Plantlets successfully acclimatized to an ex vitro substrate mixture (Soil: Husk Ash: Coconut Coir = 1:1:1 v/v) with a survival rate of 85%. This study successfully established an in vitro propagation protocol utilizing the TCL technique, laying the groundwork for conserving and developing the precious medicinal plant D. orlowii.

Intercropping intensity and crop structure in coffee-based systems among ethnic minority farmers in Dak Lak Province, Vietnam

Phan Thi Thuy — 2026-08-31

This study analyzes intercropping systems in coffee farms of ethnic minority households in Dak Lak province, Vietnam. Using cross-sectional survey data from 150 ethnic minority coffee households in Dak Lak province by using a purposive sampling approach. .Intercropping intensity is measured by the Density Equivalent Ratio (DER), whilecrop structure is displayed by dominance index. The results indicate that the average DERplot reaches 1.26 and most farming systems operate at a moderate level of intensification. Coffee remains the dominant crop, accounting for over 70% of total system density, while other crops are integrated at relatively low densities. These findings provide useful empirical evidence for policymakers to design context-specific strategies that promote efficient and sustainable intercropping in coffee-based farming systems among ethnic minority households.

The current status of implementing support policies for full-time undergraduate students at Tay Nguyen University

Phan Xuan Tho — 2026-08-31

This paper focuses on analyzing the current status of implementing support policies for full-time undergraduate students at Tay Nguyen University. The research findings indicate that Tay Nguyen University has established procedures for receiving, reviewing, and disbursing support policies. The study examines the implementation of support policies across three main groups: tuition exemption and reduction along with social allowances, scholarship policies, and support policies for pedagogy students. Based on these findings, the paper proposes several solutions to improve the effectiveness of implementing support policies for full-time undergraduate students at Tay Nguyen University.

Effective master’s thesis and project supervision: Implications for improving the quality of application-oriented projects at Tay Nguyen University

Do Thi Nga — 2026-08-31

Application-oriented master’s training necessitates a stronger integration of knowledge and practice. However, Tay Nguyen University currently faces core challenges: barriers in infrastructure and digital competency; a disconnect between curricula and practical demands; and the lack of an effective output quality assessment mechanism. This paper aims to systematize a theoretical framework for effective supervision based on the Input-Process-Output model, thereby proposing solutions to optimize the quality of graduation projects at the university. Using a systematic review combined with internal policy analysis and practical benchmarking, the study establishes a foundation for policy implications. The findings suggest a synchronized solution system across three pillars: (1) standardizing inputs; (2) optimizing processes; and (3) controlling output quality within the context of digital transformation. These implications not only enhance the applicability of graduation projects but also affirm the university’s role in addressing local practical challenges.