Science & Technology Asia

The Development a Job Position Recommendation System Using Collaborative Filtering Technique

Natratanon Kanraweekultana — 2026-09-30

Unemployed individuals and recent graduates often lack the experience or skills necessary to effectively search and filter job positions on job search websites, hindering their ability to access job listings and requirements. To address this issue, a research process was conducted involving extracting job information from the top job search website in Thailand. The data was then prepared, transformed, and subjected to Matrix Factorization. The data was randomly split into 80% for training and 20% for testing, out of a total of 30,000 records. This data was used to test the performance of a recommendation system utilizing the Collaborative Filtering technique. The Learning Curve, which measures the efficiency of model learning through Training Loss and Validation Loss, showed a continuous decrease in value until it reached its lowest validation loss of 0.0442, indicating a well-fitted model. This research suggests that the model can be effectively used for predicting and creating a prototype of a job recommendation system that connects job opportunities via API to efficiently meet the needs of job seekers and recent graduates.

A Spatial Modulation-Based Time and Bandwidth Sharing NOMA Framework for Enhanced IoT Connectivity at Cell Edges

V. Siva Bhaskara Rao — 2026-09-30

The Non-Orthogonal Multiple Access (NOMA) has attracted every wireless communication network based on its flexibility and scalability. However, handling or functioning the NOMA system for a specific wireless communication system, such as cellular or sensor networks, became a complicated task. In this present study, the User Pairing Spatial Modulation NOMA (UPSM-NOMA) has been considered, and its efficiency was improved by implementing the novel decision tree Fossa Resource Allocation (DTFRA). Initially, as per the need, UPSM-NOMA was designed in the MATLAB environment, and then a novel optimized resource allocation process was executed with the help of Fossa's optimal best solution. The considered resources for this present study are the spectral sources. Finally, the performance of the designed optimal UPSM-NOMA was compared with other traditional NOMA in terms of Bit Error Rate (BER), sum rate band throughput. In every aspect, the proposed NOMA has gained the finest solution.

Valorization of Geothermal Silica Waste in the Fabrication of PVA-Based Nanofiber via Electrospinning: Morphological, Mechanical, and Surface Property Evaluation

Kusno Isnugroho — 2026-09-30

This study explores the valorization of geothermal silica waste as a functional filler for the fabrication of polyvinyl alcohol (PVA)-based nanofibers via electrospinning. Geothermal silica was processed through a sol–gel route and incorporated at varying weight fractions (0–7 wt%) into PVA solutions before electrospinning. The resulting nanofiber mats were characterized in terms of morphology, elemental composition, chemical structure, mechanical properties, and surface wettability using FESEM-EDS, FTIR, Vickers hardness testing, and water contact angle measurements. Silica incorporation influenced the nanofiber morphology, reducing the average fiber diameter to 170 nm at 5 wt% silica loading; however, excessive silica content promoted bead formation and fiber fusion due to limited silica dispersion within the PVA matrix. EDS and FTIR analyses confirmed the incorporation of geothermal-derived silica within the PVA nanofiber matrix. Mechanical evaluation revealed that 1 wt% silica loading exhibited the highest hardness value (20.36 HV), while higher silica concentrations reduced mechanical performance due to morphological defects and silica aggregation. Surface wettability analysis showed improved hydrophilic behavior with increasing silica loading up to 5 wt%, while excessive silica addition resulted in reduced wettability due to morphological heterogeneity. These findings demonstrate that geothermal silica waste can be successfully converted into functional PVA/silica nanofibers fcomposites, highlighting the importance of silica loading optimization to achieve balanced mechanical and surface properties for potential membrane-related applications and circular economy strategies.

Computation of SAR and Temperature Increase in Human Eye Exposed to 5G Radiation

Deepshikha Bhargava — 2026-09-30

The rapid deployment of fifth-generation (5G) wireless communication systems has raised questions regarding the potential biological impacts of prolonged exposure to radiofre-quency electromagnetic (RF-EM) radiation. Due to the sensitivity of ocular tissues, this study investigates the interaction between 5G electromagnetic waves and the human eye. A two-dimensional computational model of the human eye was developed and subjected to EM exposure at various 5G operating frequencies. The resulting electric field distribution, specific absorption rate (SAR), and temperature increase caused by RF energy absorption were evaluated. A coupled multiphysics approach was implemented using Maxwell’s equations to simulate EM wave propagation through the eye tissues and the Pennes bioheat transfer equation to estimate temperature variations within different ocular layers. The coupled EM–thermal analysis was performed using the finite element method (FEM). The results indicate that, under certain exposure conditions, the resultant SAR and temperature values exceeded the threshold values.

Empirical and Simulation Based Safety Evaluation of Forced Gap Behaviour at an Urban Unsignalised Intersection under Mixed Traffic

Abhishek Ramesh — 2026-09-30

Unsignalised intersections present significant safety risks due to uncontrolled conflicts between vehicles on major and minor roads, especially during right-turn manoeuvres. This study examines the impact of speed bumps in reducing forced gap behaviour at an
unsignalised intersection using a variety of surrogate safety indicators. Forced gap behaviour occurs when vehicles on a major road reduce their speed in response to aggressive right-turn manoeuvres from minor roads. Data was collected using a videographic technique at a Tintersection in the Thiruvananthapuram district of Kerala, India. The study analyses changes in the forced gap behaviour of vehicles at unsignalised intersections using various surrogate safety indicators, including Post Encroachment Time (PET), Occupancy Time at the conflict area (OT) and Speed Differential Percentage (SDP). The study found that when no speed bump was present, PET values ranging from 0 to -5 seconds accounted for 36.57% of the total sample size. However, this percentage increased to 55.72% when a speed bump was implemented indicating a reduced forced gap behaviour. A VISSIM simulation-based analysis indicated that the presence of higher proportion of motorised two-wheelers increases the forced gap behaviour (PET ∈ [-5, 0]) even under the presence of a speed bump.

Impact of a High-Pressure Condensate Return System Upgrade on Boiler Fuel and Utility Consumption: A Case Study in a Corrugated Board Manufacturing Plant

Wirote Ritthong — 2026-09-30

This study evaluates the technical and economic performance of a high-pressure condensate return system installed on an 8-ton industrial steam boiler at a corrugated board manufacturing plant in Samut Sakhon, Thailand. Before installation, surface temperatures at ten steam-trap points ranged from 129◦C to 166◦C, while the feedwater holding tank temperature was 50–60◦C. A 15-kW high-pressure booster pump and 300-liter tank were installed to improve condensate recovery. Following installation, annual heavy fuel oil (HFO) consumption decreased by 58.5%, from 1,202,000 liters in 2022 to 498,500 liters in 2023, with comparable production output. Annual water consumption decreased by 28.9% (5,896,000 liters), and electricity consumption decreased by 10.3%. The feedwater holding tank temperature increased to approximately 150◦C. Net annual monetary savings, after additional pump electricity costs, reached 21,197,558 THB. Over five years at a 6.25% discount rate, the project achieved an NPV of 86,187,810 THB, IRR of approximately 848%, and simple payback of approximately 43 days. However, fuel savings reflect both the pump upgrade and concurrent steam-trap repairs. High-pressure condensate recovery can improve resource efficiency, but site-specific assessment and further validation are recommended.

A Review of Non-Linear Vibrations in Heavy Trucks: Challenges, Control Strategies, and Advances in Suspension Technologies

Wutthipath Tinsuwan — 2026-09-30

Non-linear vibrations in heavy trucks affect transportation efficiency, vehicle durability, and operational safety. They arise from interactions among suspension components, tire–road dynamics, and varying payloads, increasing fuel consumption, component wear, and cargo damage. This study reviews advances in the characterization, modeling, and control of non-linear vibration in heavy truck suspensions. Studies published between 2000 and 2024 were retrieved from Scopus, IEEE Xplore, ScienceDirect and SpringerLink and screened against predefined criteria ([N] studies included). Existing analytical and computational models capture fundamental dynamics, but their accuracy is limited by dynamic loading, geometric effects, and stochastic road inputs. Semi-active and active suspensions improve vibration suppression yet need better adaptability in unpredictable environments. Opposite-phase (counter- phase) force control, in which an actuator applies a force opposing the measured body motion as in active noise control, is examined and compared with passive, semi-active and active suspensions in terms of control principle, reported vibration reduction and limitations. Its potential for improving ride comfort and reducing structural loads is assessed against the reported evidence. Key research gaps include adaptive suspension design, real-time non-linear vibration prediction, and intelligent control. Addressing them is essential for sustainable, reliable, and cost-effective freight transportation.

Blockchain-Based Quantum-Secure Federated Learning with Hybrid Qkd-Pqc Encryption and Multi-Party Secure Aggregation for Healthcare Systems

Sandeep H. — 2026-09-30

Predictive modelling of brain and critical care conditions requires accuracy, speed, and privacy, yet sensitive clinical data and fragmented healthcare systems hinder AI implementation. Existing Federated Learning frameworks struggle to balance performance, scalability, security, and privacy in real-world medical environments and deployment. This work proposes a fully encrypted federated learning architecture integrating locally trained models, enhanced by Adaptive Differential Privacy (ADP), with MIMIC-III/IV clinical datasets, ensuring hospital-level data locality while maintaining model accuracy. Post-Quantum Secure Communication is enabled through a Hybrid Obfuscated Quantum Key Distribution–Post-Quantum Cryptography (QKD-PQC) system, providing secure key exchange, dynamic obfuscation, and quantum-resistant encryption. AI-driven smart contracts on a Permissioned Blockchain verify PQC signatures, monitor node reputation, coordinate aggregation, and reward reliable participants. During global model synthesis, a Secure Aggregation Layer using Multi-Party Computation (MPC) and Shamir’s Secret Sharing protects encrypted model updates, ensuring privacy and immutability.

Machine Learning-Based Time-to-Failure Prediction for a Beverage Packaging Labeling Machine

Sarocha Kaewpradab — 2026-09-30

Predictive maintenance applies machine learning (ML) to forecast industrial time-to-failure (TTF), but reliable evaluation is challenged by changing operating conditions, repeated failure events, and information leakage. This study develops a leakage-aware TTF prediction framework using four years of operational and maintenance data (2021–2024) from an industrial beverage-packaging labeling machine, including 1,560 operational records and 479 unplanned failures. Six ML algorithms—Random Forest (RF), Gradient Boosting, XGBoost, Support Vector Regression, Multi-Layer Perceptron, and CatBoost—were com-pared with five historical, temporal, persistence, and linear baselines. CatBoost-Cox, using a Cox survival objective, was also evaluated for relative failure-risk discrimination. Nested walk-forward validation with event-grouped partitioning and fold-specific preprocessing, feature selection, and hyperparameter tuning was employed. All ML models outperformed baselines in RMSE, MAE, and MedAE. RF achieved the lowest mean RMSE (1.8813 days), MAE (1.0121 days), and MedAE (0.2552 days), while CatBoost achieved the highest mean R² (0.6280). However, no significant RMSE differences remained after Holm correction. CatBoost-Cox achieved a mean concordance index of 0.9155 (SD = 0.0336). Residual, temporal, computational, and sensitivity analyses revealed model-specific trade-offs, emphasizing leakage-aware temporal validation and application-specific model selection.

Bezier Curve in Modeling Bonds Yield Curve

Emy Siswanah — 2026-09-30

Before buying bonds, investors must ensure that the bonds chosen by investors have good prospects and can provide profits as expected, so investors must analyze the yield curve. The results of yield curve analysis will help investors to model the yield curve and help make decisions about investing in certain bond products. Bond yield curves can be modeled using Bezier curves. The Bezir curve was chosen because the Bezier curve has a control value which functions to direct the points on the curve to move from the starting point to the ending point. The more control values, the more complex the bezier curve formed will be and the bezier curve can be adjusted to real data. This study compares high order Bezier curves (6^th order) with low order Bezier curves (5^th order). The results of the research show that the Bezier order 6 curve has a smaller error compared to the Bezier order 5 curve. More control values on the order 6 Bezier curve make the curve move more complexly and adapt to real bond yield data.

Parameter Estimation for Brownian Motion with Trend Mixed with Two Fractional Brownian Motions

Phisitphong Ketrat — 2026-09-30

This work studies parameter estimation for a stochastic process that combines a Brownian motion with trend and two fractional Brownian motions with distinct Hurst indices. The process is given by , where is a standard Brownian motion and are fractional Brownian motions with Hurst indices such that , with , and being mutually independent. An ergodic theorem is applied to derive estimators for all unknown parameters and examine the effect of different Borel function selections on the estimator formulations. Numerical simulations are performed to confirm the asymptotic properties of the estimators and to assess their accuracy under various configurations. The estimators are then applied to cryptocurrency data to demonstrate their practical use.

New φ-fixed point results in metric spaces and well-posedness property

Binayak S. Choudhury — 2026-09-30

In this paper, we introduce two new φ-fixed point problems in the setting of complete metric spaces using α-admissible mappings. The motivation of the work is derived from the efforts for combining several trends of research in fixed point theory to create new results. Existence and uniqueness results are established under appropriate conditions. The main results are supported by two separate illustrative examples. The well-posedness property of the proposed problems are investigated. One of the theorems is effectively applied to study the solvability of a problem involving functional equations.

Pivot-Based Parameter Estimation for the Gumbel Distribution with Progressive Type-II Censoring Scheme

Chitrapa Sangarun — 2026-09-30

In this article, we propose a parameter estimation method based on pivotal quantities for the Gumbel distribution following the progressive Type-II censoring scheme. We also conduct a simulation study to assess the efficiency of the proposed method compared to the classical maximum likelihood estimation method. Simulation results indicate that pivot-based method outperforms the likelihood method in terms of both bias and mean squared error, particularly for the case of small sample sizes. Finally, we apply the proposed method to two real applications and evaluate its performance through different goodness-of-fit criteria. The results demonstrate strong performance of the method across both applications.

The Marshall-Olkin Extended Circular Uniform (MOECU) Distribution: Its Properties and Application

Phani Yedlapalli — 2026-09-30

Circular data arise naturally in numerous scientific fields, including meteorology, biology, geology, environmental sciences, and directional statistics. The development of flexible circular distributions capable of modeling asymmetric angular observations remains an active area of research. In this article, we introduce the Marshall-Olkin Extended Circular Uniform (MOECU) distribution, a new one-parameter circular distribution on the interval , obtained through the Marshall-Olkin transformation [1] to the circular uniform baseline distribution. The proposed model possesses a simple closed-form probability density function, cumulative distribution function, and quantile function, making it particularly attractive for theoretical investigations and practical utility. We derive the characteristic function and trigonometric moments of the distribution, and obtain its population characteristics. The hazard rate, mode and median are also examined. The parameter is estimated by the method of maximum likelihood and the finite sample performance of the estimator is investigated through an extensive Monte Carlo simulation study, together with several rival estimation techniques(least squares, weighted least squares, Cramer-Von Mises, Anderson-Darling and its variants: maximum product of spacing’s, and percentile method) The practical applicability of the MOECU distribution is illustrated using the classical angular data on orientations of starhead topminnows[2](FisherB4), where it is compared with several competing circular models through standard goodness-of-fit measures. The results demonstrate that the proposed distribution provides an effective and parsimonious alternative for modeling asymmetric circular data and yields competitive performance in practical applications.

A Comparative Study of Lot-Sizing Techniques under Fluctuating Demand

Kesinee Sansee — 2026-09-30

This research aims to compare the efficiency of different lot-sizing techniques under fluctuating demand patterns. Demand was simulated using a normal distribution and categorized into two levels based on the variability coefficient (VC): low variability (VC < 0.25) and high variability (VC ≥ 0.25), across two planning horizons: 12 periods (monthly) and 52 periods (weekly). The study evaluates both the exact optimization approach using the Wagner–Whitin algorithm (WW) and various heuristic methods, including Lot-for-Lot (LFL), Economic Order Quantity (EOQ), Period Order Quantity (POQ), and Part-Period Balancing (PPB). Each scenario was tested across two levels of ordering cost and five levels of holding cost. The results indicate that the WW algorithm outperformed all other techniques across all simulated scenarios by achieving the lowest total inventory management cost. Among the heuristic methods, LFL and POQ emerged as viable alternatives; their efficiency improved significantly as holding costs increased because they align order quantities closely with actual demand, thereby reducing average inventory levels and carrying costs.

A Reconstruction Approach for Noisy Data by using Techniques Based on Hankel Matrix and Least-Squares Approximation

Saifon Chaturantabut — 2026-09-30

Estimating missing data in volatile time series with oscillatory and noisy patterns is a challenging issue in many applications. This work introduces a technique for missing data imputation that combines data denoising and least-squares approximation. Our method first applies the Hankel Singular Value Decomposition (HSVD) to the dataset to filter out noise and capture the underlying data trend. The cleaned, trend-focused data then serves as the basis for a least-squares approximation to accurately estimate the missing values. This approach is specifically applicable for noisy datasets where related information is available, such as financial data from a stock market. The efficiency of this combined approach is validated through a numerical case study on the stock prices of companies within Thailand's healthcare sector. The results confirm that preprocessing the data with HSVD before applying the least-squares estimation leads to a more accurate and reliable reconstruction of missing entries, highlighting the method's potential for financial time series analysis.

An Analytical Study of Fixed Point Theorems in Intuitionistic Fuzzy b-Metric Spaces

Kshama Pandey — 2026-09-30

In this paper, the study is devoted to investigating fixed point theorems in intuitionistic
fuzzy b-metric spaces (IFb-MS) and extends classical results from fuzzy b-metric spaces
(Fb-MS). It introduces a broader class of sufficient conditions ensuring the Cauchy nature
of sequences, thereby generalizing and simplifying contraction mappings and their proofs.
Theunified approach accommodates diverse contractive conditions and offers novel insights
into the establishment of existence and uniqueness results for fixed points in the proposed
generalized topology

A Bi-Objective Two-Level Capacitated Facility Location Model with Primary–Backup Assignments under Depot Disruptions

Aua-aree Boonperm — 2026-09-30

Supply chain disruptions affect the economic and delivery performance of distribution networks. This study develops a bi-objective two-level capacitated facility location model under depot disruptions, assigning customers to primary and backup depots through the same satellite to ensure service continuity. The model minimizes expected system cost and delivery time under capacity restrictions. The initial nonlinear formulation is reformulated as an equivalent mixed-integer linear programming model, and a normalized weighted-sum approach examines the objective trade-off. Experiments on 20–70-node instances reveal a clear cost–delivery-time trade-off. Both single-objective models reach proven optimality for up to 35 nodes, while larger instances exhibit increasing solution times and optimality gaps. The results demonstrate the value of balancing economic and delivery performance in disruption aware distribution network design.

A Proteomic Dataset of HepG2 Cellular Responses to Cycloheximide-Induced Translational Arrest

Thanawan Soimanee — 2026-09-30

Cycloheximide (CHX), a potent translational inhibitor, is widely used in cellular assays requiring translational blockade. Due to limited characterization of CHX-exposed cellular models, this study investigated the molecular and cellular effects of CHX in human hepatoma HepG2 cells. Cells were treated with various CHX concentrations for early (1 and 3 hours) and prolonged (18 and 24 hours) periods and evaluated for cellular responses. MTT assays showed no significant viability changes during early exposure, whereas prolonged treatment with 25 and 50 uM CHX reduced viability by approximately 30%. Sustained CHX exposure induced Sphase cell cycle arrest with minimal apoptosis (approximately 10%). LC-MS/MS analysis identified 19,482 differentially expressed proteins (DEPs), while PLS-DA revealed clear separation between untreated and CHX-treated HepG2 proteomes across time points. Volcano plot analysis identified the five most altered proteins at each time point, representing diverse cellular compartments and biological functions. STITCH network analysis further identified three interactome clusters associated with CHX responses: calcium-mediated survival signaling, xenobiotic detoxification, and ubiquitin-proteasome degradation. This characterized CHX-exposed HepG2 model provides a useful methodological reference for bioassays requiring translational stress.

In vitro Induction of Axillary Turion Production in Hydrilla Verticillata with the Potential for Short-Term Storage via Alginate Encapsulation

Pinida Joradon — 2026-09-30

Hydrilla verticillata is an aquatic macrophyte that has attracted increasing attention due to its potential applications in ecological management and pharmaceutical research. To advance these applications, cultivating this plant using efficient large-scale operations under controlled conditions and germplasm storage technology, such as plant tissue culturing, are much more desirable than harvesting it from nature, which also has the potential to cause environmental damage. However, information regarding the tissue culture of this submerged vascular plant remains limited. Therefore, the present study aimed to establish an efficient in vitro stock culture system that could subsequently be utilized to investigate the effects of in vitro culture conditions, as well as the potential for synthetic seed production and short-term storage. In vitro cultivation of H. verticillata was successfully achieved using basal liquid Murashige and Skoog (MS) medium. The addition of either benzyladenine or paclobutrazol to the medium resulted in reduced shoot elongation and decreased lateral branching. Notably, when explants cultured in basal liquid MS basal medium were transferred to semi-solid MS medium, the development of axillary turions was observed in 55% of the jars, compared to 0% in those that remained submerged in the liquid medium, suggesting that water stress, reduced direct availability of nutrients and improved aeration conditions may stimulate axillary turion formation in vitro. Furthermore, these axillary turions were found to be a suitable explant source for synthetic seed production and enhanced plant survival during low-temperature storage compared with synthetic seeds derived from shoot tissues. The findings of this study provide a useful approach for H. verticillata germplasm conservation and facilitate the transportation of plant materials.

Cytogenetic Studies of Wrinkle-lipped Bat, Chaerephon plicatus, from Tham Pra Thun Non-hunting Area, Uthai Thani Province, Thailand

Kanya Anukulthanakorn — 2026-09-30

Chromosome and morphometric analyses of the wrinkle-lipped bat, Chaerephon plicatus (Buchannan, 1800), from Khao Khong Chai cave, Tham Pra Thun Non-hunting area in Uthai Thani province, Thailand were studied. Bat chromosomes in both males and females were determined by the squash technique using bone marrow and testis following conventional staining and Ag-NOR banding technique. The results showed that the diploid chromosome number of C. plicatus was 2n = 48, and the fundamental number (NF) was 56 in both male and female. The types of autosomes were 4 large metacentric, 2 medium metacentric, 8 medium telocentric, 4 small metacentric, and 28 small telocentric chromosomes. The X chromosome was medium metacentric, and the Y chromosome was small telocentric. The NOR chromosome markers were found on chromosome pair 6 telocentric and 9 metacentric. The karyotype of C. plicatus was as follows:

2n (48) = L^m₄+ M^m₂ + M^t₈ + S^m₄ + S^t₂₈ + sex chromosomes

These findings provide additional cytogenetic information for C. plicatus and contribute to the understanding of chromosomal variation within the family Molossidae.

A Network Pharmacology Approach to Explore Flavonol Compounds from the KNApSAcK Database as Therapeutic Candidates for COVID-19

Neni Frimayanti — 2026-09-30

COVID-19 is an infectious disease caused by SARS-CoV-2. The identification of new therapeutic candidates targeting multiple proteins and biological pathways may provide an alternative strategy for COVID-19 treatment. This study aimed to investigate the potential of flavonol compounds as multitarget antiviral candidates against COVID-19 using network pharmacology and molecular docking approaches. Flavonol compounds were obtained from the KNApSAcK database and evaluated for drug-likeness based on Lipinski’s Rule of Five using SwissADME. Flavonol-associated protein targets were predicted using SwissTargetPrediction, while COVID-19-related proteins were identified using GeneCards. The overlapping proteins were analyzed using a Venn diagram, followed by protein-protein interaction analysis using STRINGDB to identify potential hub proteins. Of the 25 flavonol compounds evaluated, 21 fulfilled the Lipinski criteria and were retained for further analysis. Network analysis identified 170 overlapping proteins and 16 hub proteins, comprising ACE2, ACE, AGT, REN, TMPRSS2, FURIN, CD4, CD8A, IL6, IL1B, IL10, TNF, IFNG, CRP, CXCL8, and CXCL10. ACE2 showed the highest connectivity and was therefore selected as a representative hub target for molecular docking. Based on docking binding free energy, the four flavonols with the most favorable predicted binding affinities toward ACE2 were Broussonol E, 5,7,3′,4′-tetrahydroxy-6- geranylflavonol, Broussoflavonol A, and Broussoflavonol B, all stronger than baicalein and sharing several interaction residues with the reference compound. These findings support the potential of the investigated flavonols as multitarget antiviral candidates at the network level, involving proteins associated with viral entry, the renin–angiotensin system, and immune and inflammatory responses. However, the present docking analysis provides target-specific structural evidence only for ACE2. Further computational and experimental studies involving the remaining hub proteins are required to validate the proposed multitarget mechanism and antiviral activity.

In Silico Studies of Carotenoid Antioxidants in Sand Lobster (Panulirus homarus) as Candidate Anti-Breast Cancer Compounds

Violanda Aldila Rahayu — 2026-09-30

Breast cancer is the most diagnosed and deadliest type of cancer among women worldwide. In Indonesia, breast cancer ranks as the most prevalent and deadly cancers, with 58,256 cases and 22,692 deaths. This research aims to analyze the potential of carotenoid antioxidants found in the sand lobster (Panulirus homarus) as breast cancer treatment candidates through an in silico approach. In this study, six carotenoid compounds were evaluated: dinoxanthin, diadinoxanthin, zeaxanthin, lutein, astaxanthin, and violaxanthin, using 4-hydroxytamoxifen, genistein, and exemestane as standard ligands, and tamoxifen, metformin, and letrozole as positive controls. The 3D crystal protein structures used for molecular docking were obtained from the Protein Data Bank (PDB), including estrogen receptor alpha (PDB ID: 3ERT), estrogen receptor beta (PDB ID: 1X7J), and aromatase (PDB ID: 3S7S) as target receptors. Subsequently, bioactivity and bioavailability analyses of the test compounds were performed. Proteins and test compounds were prepared, followed by molecular docking against the estrogen receptor alpha, estrogen receptor beta, and aromatase, with visualization performed using Discovery Studio. Molecular dynamics simulations were performed to evaluate the stability of binding interactions. Toxicity prediction was carried out using the Protox Web Server (https://tox-new.charite.de/protox_II/). The in silico results indicate that two of the six tested compounds, lutein and violaxanthin, showed favorable predicted binding interactions with estrogen receptor alpha, estrogen receptor beta, and aromatase. Additionally, in silico toxicity predictions indicated favorable toxicity profiles for these compounds, suggesting their potential as candidates for further investigation against breast cancer.

Thalassemia Screening Based on Hematological Data Using Random Forest and Support Vector Machine Models: A Preliminary Study

Mifta Mardiah — 2026-09-30

Thalassemia is a hereditary blood disorder caused by impaired globin chain synthesis and represents a public health concern due to increasing incidence and mortality. Screening comprises three levels: hematological examination (Level 1), hemoglobin analysis (Level 2), and molecular genetic testing (Level 3). When Levels 1 and 2 cannot conclusively determine thalassemia status, resource-intensive Level 3 testing is required. This study applied a hybrid exploratory approach combining unsupervised clustering and supervised classification to support Level 1 screening using hematological data from the Thalassemia Research Group. After data preparation, K-means and Gaussian Mixture Model (GMM) clustering were used to identify patterns and outliers, while Random Forest (RF) and Support Vector Machine (SVM) were used for classification. K-means produced six groups, with normal and 𝛽-thalassemia showing relatively clear clustering, although overlap remained. Overlap was also observed between hemoglobin E (HbE) and normal groups and among other categories. GMM achieved 79.11% accuracy. RF achieved the highest classification accuracy of 91%, followed by SVM-linear at 86%. However, SVM-linear correctly classified 40 of 52 test samples, slightly exceeding RF performance. These findings demonstrate the potential of combining unsupervised and supervised learning for Level 1 screening. Larger datasets and additional parameters may improve model performance and generalization.

Macrofungi Diversity in a Community Forest, Tha Mai District, Chanthaburi Province, Thailand

Saowapha Surawut — 2026-09-30

This study aimed to survey macrofungi diversity in a community forest in Tha Mai District, Chanthaburi Province, Thailand, and to evaluate their trophic mode and edibility; the knowledge gained from this study will help to establish a macrofungi database as part of a biodiversity learning center for the community. Macrofungi sampling was performed in the area, with identification being done by morphological analysis and internal transcribed spacer (ITS) sequencing. The results showed that 18 unique macrofungi samples with diverse fruiting body characteristics were examined. These could be classified into 9 groups: (i) Ascomycetes fungi, (ii) Gilled fungi, (iii) Polypores and bracket fungi, (iv) Leather-bracket fungi, (v) Jelly fungi, (vi) Earthstar fungi, (vii) Stinkhorn fungi, (viii) Coral fungi, and (ix) Bird’s nest fungi. The gilled fungi were the most common (27.78%). Subsequently, molecular identification by ITS region analysis was able to classify them into 2 phyla, 2 classes, 8 orders, 14 families, and 17 genera. The phylum Basidiomycota (83.33%) and the family Polyporaceae (22.22%) were the most predominant taxon and family, respectively, in this area. There were 6 samples (33.33%), the genus classification was reliable (95 to 98.5% similarity), but the species are yet to be confirmed. Most fungi acted as saprotrophs (94.44%) in the ecosystem of the community forest, and only 1 species (5.56%), Termitomyces sp. (KN20), functioned as a symbiotroph. There was however no pathotrophy documented in this study. Additionally, the edible macrofungi (16.67%) reported in this study include Auricularia novozealandica (KN40), Phallus lutescens (KN38), and Termitomyces sp. (KN20), with no report of any poisonous macrofungi in this area. However, the remaining macrofungi species (83.33%) have no information on edibility or toxicity.