International Journal of Bioinformatics Research and Applications (31 papers in press)

Regular Issues

• Deep Learning Approach using Modified DarkNet-53 for Renal Cell Carcinoma Grading  
  by G. Sathish Kumar, G. Uma Maheshwari, C. Selvan, M. Nagasuresh, Rasi D, Swathypriyadharsini Palaniswamy, Sathish Kumar Danasegaran 
  Abstract: An accurate and effective diagnostic procedures are required for appropriate treatment planning for renal cell carcinoma, the most frequent form of kidney cancer. Using fusion module a network dubbed Modified Darknet (MDNet) was developed for image-based small-target detection. We built MDNet on top of a modified version of Darknet53, which itself a scale matching approach, to increase its speed and accuracy. By combining the results of several convolutional neural network (CNN) models, the ensemble structure improves classification accuracy. The effectiveness of a classification algorithm using kidney histopathology pictures dataset is measured in accuracy, precision, recall, sensitivity, specificity and f1-score. The results show that the ensemble deep learning method outperforms both standalone CNN models and more conventional machine learning techniques in RCC classification. Overall grade classification accuracy of 98.9%, a sensitivity of 98.2%, and a high classification specificity of 98.7%, in distinguishing tissues.
  Keywords: Modified Darknet; Convolutional Neural Network; Ensemble Deep Learning; Kidney Cancer; Renal Cell Carcinoma; Whole Slide Images.
  DOI: 10.1504/IJBRA.2025.10064488
   
  
• HDAC Inhibitors and their Potential towards Cancer Treatment  
  by Sanjay Kumar Choubey, Sachin Kumar, Medha Kumari 
  Abstract: Histone deacetylases (HDACs) play a key role in chromatin structure modulation through deacetylation of histones leading to formation of highly compact DNA-histone complex. HDACs have been reported to be implicated in multiple types of cancers. Blocking the activities of histone deacetylases will help to overcome gene repression pressure and it would be possible to check the incessant growth of cells in tumour. Therefore the interest has been developed to design the HDAC inhibitors and their analogues and histone deacetylases are now considered as potential targets for their wide distribution in various forms of cancer. HDAC inhibitors display their role by regulating cyclin dependent kinases (cdK), inducing p21 and various preapoptotic genes like Bax, Bak, repressing the activities of growth factors like VEGF, repressing transcription factor HIF-1 facilitating arrest of cell cycle, modulating various signalling pathways like STAT signalling, AMPK signalling, inducing cell adhesion molecule E-cadherin.
  Keywords: cyclin dependent kinase; histone deacetylase; carcinogenesis; HDAC inhibitor.
  DOI: 10.1504/IJBRA.2025.10065088
   
  
• Computational Analysis of Alkoxy-Azoxybenzene Liquid Crystals: A Comparative Investigation with Experimental Data for Bioinformatics Applications  
  by Sushma M, Mahadev J, Manju V. V, Nandaprakash M. B, Somashekar R 
  Abstract: Through computational modelling, we have gained valuable insights into the homologous series of liquid crystalline materials. Our study involved comparing the computational results with reported experimental values for several members of the series. We focused on various parameters, including lattice energy, orientational order parameter, moduli, stress-strain behaviour, Helmholtz free energy, orientational distribution function, zero-point energy, and molecular polarisabilities. The primary motivation behind this study was to unravel the intricate inter- and intra-molecular interactions that govern the range and nature of mesophases observed in these compounds. We are excited to report that our results align with this objective, highlighting the significance of our findings in this direction. Knowledge of these compounds finds applications in sensitive nucleic acid detection, label-free protein analysis, and the development of biocompatible sensors for real-time cellular monitoring.
  Keywords: liquid crystal; odd-even effect; elastic moduli.
  DOI: 10.1504/IJBRA.2025.10065091
   
  
• A Novel Linear Discriminant Analysis Based Classification of R-peaks in Different ECG Signal Datasets  
  by Varun Gupta 
  Abstract: In the current scenario, there is a need to develop efficient pre-processing and classification techniques which can form the basis of an automated health monitoring system. In this paper, independent component analysis (ICA) is proposed to be used for electrocardiogram (ECG) signal processing as reported by the same authors, who found it to yield better results that time for limited datasets. Here, it has been applied on a variety of datasets, viz., real and standard and the obtained results are compared with those obtained using another widely used and reported technique, viz., adaptive notch filter (ANF) in the literature. For classification, linear discriminant analysis (LDA) is proposed to be used as it performs multi-class classification tasks better. The obtained results demonstrate the utility of the proposed methodology for bioinformatics community, especially during critical heart surgeries and designing of evolving healthcare systems in future.
  Keywords: electrocardiogram; ECG; adaptive notch filter; ANF; independent component analysis; ICA; linear discriminant analysis; LDA; signal-to-noise ratio; SNR.
  DOI: 10.1504/IJBRA.2025.10065193
   
  
• A Novel Approach for Early Detection and Grading of Diabetic Retinopathy by using Ensemble Model  
  by Riddhi Parasnaik, Anvita Agarkar, Raashi Jatakia, Gajanan Nagare 
  Abstract: This study investigates the factors driving HR professionals' intention to adopt AI in talent acquisition in the Indian IT industry by adopting a mixed technology-organization-environment (TOE) and task-technology fit (TTF) model. We administered a survey instrument on 459 HR professionals including talent acquisition executives randomly selected from various IT firms located in major Indian cities. The PLS-SEM results revealed that the perception of cost effectiveness, relative advantage, HR readiness, top management support and competitive pressure significantly influence the adoption intention of HR professionals of the Indian IT companies. The findings of the study would help understand the factors that influence HR managers' decisions to adopt AI in talent acquisition process. Further, the study contributes to the existing adoption theories by integrating TOE and TTF models to HR contexts and offers actionable insights for practicing managers of the organisations aiming to adopt AI in the recruitment process.
  Keywords: digital transformation; artificial intelligence; AI; talent acquisition; technology-organisation-environment model; task technology fit model; India.
  DOI: 10.1504/IJBRA.2025.10065195
   
  
• Characterising the Cardioprotective Potential of Sida Rhombifolia, Polygonum Chinense and Phyla Nodiflora Aqueous Extracts: Investigating its Effect on Foam Cell Formation  
  by Xiao Wei Lee, Wei Sheng Siew, Siau Hui Mah, Wei Hsum Yap 
  Abstract: Cardiovascular diseases represent one of the leading causes of mortality. Studies have shown that medicinal plants with anti-inflammatory and antioxidant activities are potential cardioprotective agents. This study aimed to determine cardioprotective potential of Sida rhombifolia, Polygonum chinense and Phyla nodiflora in inhibiting macrophage foam cells formation and its regulatory mechanisms. The findings showed that S. rhombifolia and P. nodiflora have minimal cytotoxicity effect on THP-1 macrophages, however P. chinense exhibited cytotoxic effect with an IC50 of 11.83
  Keywords: atherosclerosis; foam cell; network pharmacology.
  DOI: 10.1504/IJBRA.2025.10065324
   
  
• A Comparative Study on the Classification of SARS-CoV-2 Variants from Biosequence Images using Pre-Trained Deep Learning Models  
  by Shahina K, Biji C. L, Achuthsankar S. Nair 
  Abstract: Coronavirus disease has raised serious health concern across the globe. Identification of severe acute respiratory syndrome corona virus 2 (SARS-CoV-2) variants are indeed a concern in controlling its spread. SARS-CoV-2 variants are classified based on the variation in its genomic sequences. Alpha, beta, delta, gamma and omicron were reported as the most deleterious variants. Genome sequence can be represented uniquely using chaos game representation (CGR) images. A large-scale genome sequence dataset, belonging to the five categories of these variant were retrieved from GISAID. An attempt was made to compile benchmark CGR images of 25,000 SARS-CoV-2 variants genomic sequences. The present study aims to compare the performance of different pre-trained deep learning models in classifying SARS-CoV-2 variants from its CGR images. VGG16, VGG19, ResNet50, InceptionV3, Xception, InceptionResNetV2 and MobileNetV2 were the models used for the study. SARS-CoV-2 variant detection was found effective with VGG19 with an accuracy of 94%. Data augmentation techniques were also applied on the CGR images of biosequences and it was found that data augmentation methods decreased the accuracy of different transfer learning models.
  Keywords: genome sequence; deep learning; SARS-CoV-2 variants; chaos game representation; transfer learning; classification; COVID-19.
  DOI: 10.1504/IJBRA.2025.10065325
   
  
• Skin Cancer Classification using Ensemble Classification Model with Improved Deep Joint Segmentation  
  by Jinu P. Sainudeen, Sathyalakshmi S 
  Abstract: We present a six-phase skin cancer classification model based on Improved Deep Joint Segmentation (IDJS) in this work. The pre-processed image is segmented using IDJS in the second phase, after contrast enhancement with assistance from Contrast Limited Adaptive Histogram Equalization (CLAHE) in the first phase. The features of GLCM, CCF, LGIP, and Median Ternary Pattern (MTP) are retrieved in the third phase. Data augmentation for the extracted features is carried out in the fourth phase. The fifth phase is ensemble classification using the Deep Maxout, LSTM, and CNN based on the enhanced data. To determine the final classified label, the enhanced score level fusion receives the output scores from these classifiers. While the RF is 0.9171, Deep Maxout is 0.9382, LSTM is 0.9362, Bi-GRU is 0.8150, RNN is 0.8687, CNN is 0.9382, TL-GOOGLENET is 0.9134, and KNN is 0.9328, respectively, the accuracy of the Ensemble approach is 0.9689.
  Keywords: DL; Skin cancer; segmentation; Classification; Recommendation.
  DOI: 10.1504/IJBRA.2025.10065333
   
  
• Shri-Anna (Millets), Super Food for Present Epoch: A Thoughtful Study in Diverse Dimensions  
  by Bhavna Singh, Rohit Rastogi, Bhupinder Singh, Harpreet Kaur, Richa Singh, Ratik Dubey, Jagriti K, Tanya Tyagi 
  Abstract: The 21st century, the era of science, the era of development, when humankind is excelling in all fronts, is also the time when we have to face the gravest problems of all time. Global warming, climate change, overpopulation, chronic hunger, pandemics, wars are to name a few of them. We are fortunate to have the treasure of knowledge ensnared in the ancient Indian Vedic texts and scriptures, which have the potential solution for all the grave problems encircling the world in today's time. Millets are one such solution, these are traditional grains used from the past of five thousand years. Nowadays millets are being popular as nutri cereals, nutri millets and as superfood. Millets are gluten free hence does not lead to celiac disease, unlike wheat. These are very excellent food for a diabetic person as it takes longer time to get digested and hence can provide energy for longer tenure.
  Keywords: millets; biofortification; Kshudra Danya; Ayurveda; gluten; chronic hunger; nutrition.
  DOI: 10.1504/IJBRA.2025.10065722
   
  
• Optimizing Multi-User Massive MIMO Systems through Particle Filter Precoding: A Comprehensive Performance Analysis  
  by Terefe Abebe Beyene, Satyasis Mishra, Tadesse Hailu Ayane, Davinder Rathee, Bijaya Paikray 
  Abstract: Due to globalisation, the demands of network subscribers for different services and the total number of users in the communication industry is growing day to day. To address the demand of these users, the communication industry needed to be upgraded to support the growing demand. This research proposes an optimised particle filter, and nonlinear precoding technique to mitigate this issue. The computational complexity analysis of TH, DPC, and VP and proposed optimised particle filter precoding techniques. Signal-to-noise ratio versus bit error rate, signal-to-noise ratio versus spectral efficiency, the number of transmits antennas versus average sum spectral efficiency, and others were used to analyse the results. The simulation result shows that the proposed optimised particle filter precoding outperforms the existing nonlinear precoding those are, Tomlinson-Harashima (TH), dirty research coding (DPC), and vector perturbation (VP) precoding techniques in average sum spectral efficiency and bit error rate performance on different modulation techniques.
  Keywords: optimised particle filter precoding; linear precoding; nonlinear precoding; spectral efficiency; bit error rate; BER; massive MIMO.
  DOI: 10.1504/IJBRA.2025.10065870
   
  
• Mobile Application based Pulmonary Disease Prediction using Respiratory Sound and Deep Learning  
  by Hiwot Habtamu, Mesfin Abebe, Sudhir Kumar Mohapatra 
  Abstract: Pulmonary diseases are contagious illnesses that disrupt the respiratory system, often affecting the lungs. Diagnosing these diseases can be challenging due to similarities with other lung conditions. While many studies use pulmonary sounds for prediction, this study integrates patient medical history and respiratory sounds to enhance prediction accuracy using deep learning. By combining these data sources, a pulmonary disease prediction model was developed and integrated into a mobile app using TensorFlow Lite, improving accessibility. The model, leveraging Melspectrogram characteristics, achieved 97.0% accuracy, significantly higher than the 73.15% accuracy with only Spec-Augmentation. Evaluation by 10 experts on 30 use cases showed 26 accurate classifications, demonstrating the model's effectiveness and the benefits of using combined data for pulmonary disease prediction. In general, this study demonstrated that building a model using patient symptoms and pulmonary sound and embedding it in a mobile application improves the prediction of pulmonary disease significantly.
  Keywords: Artificial Intelligence; Deep Learning; Respiratory Sound; Pulmonary Diseases; Mobile application.
  DOI: 10.1504/IJBRA.2025.10065933
   
  
• LeNet-Xception: An Advanced Deep Learning Model for Early Covid-19 Detection from CT Scan Images  
  by Noor Fathima K, Renukalatha S 
  Abstract: The COVID-19 pandemic has necessitated the deep learning, a subset of artificial intelligence, has had noteworthy development in the field of COVID-19 identification. Deep learning algorithms can analyse medical images, like CT scan images to aid in the swift and precise diagnosis of COVID-19. Deep learning models, such as LeNet and Xception, have been used in recent studies to diagnose COVID-19 from images of CT with high accuracy. This paper presents a deep learning approach for the detection of COVID-19 using computed tomography (CT) images by proposing a hybrid model, called LeNet-Xception. Various performance metrics, including specificity, sensitivity, and accuracy, were used to estimate the performance of the presented method. LeNet-Xception model attained an accuracy of 95.9%, a sensitivity of 97.5%, and a specificity of 93.8%. According to the results, the suggested technique suggests that can precisely identify cases of COVID-19 by utilising images of CT scans with high accuracy.
  Keywords: Covid-19; CT images; Disease detection; LeNet; Deep learning.
  DOI: 10.1504/IJBRA.2025.10065939
   
  
• Analysis of the Impact of Loss Functions in U-Net Architecture for Segmentation of Right Ventricle  
  by Mahesha Y.  
  Abstract: The present paper sheds light on the effect of loss functions in U-Net architecture for the segmentation of the right ventricle. Five loss functions namely binary cross-entropy, dice, inverse dice, dice combo and combined have been tested using optimisers such as Adam, stochastic gradient descent and root mean square propagation. The accuracy of the U-Net model is measured using the popular dice coefficient metric. The two loss functions dice and dice combo achieved maximum dice coefficients of 0.7825 and 0.7633 with stochastic gradient descent respectively. The result also shows that the loss functions such as dice and dice combo give acceptable dice coefficients with all three chosen optimisers. The loss functions binary cross entropy, combined and inverse dice have achieved moderate dice coefficients value with Adam and root mean square propagation optimisers but have shown very poor performance with stochastic gradient descent optimiser. The dice and dice combo loss functions with stochastic gradient descent optimiser are good candidates for segmentation of the right ventricle in U-Net architecture.
  Keywords: U-Net; binary cross entropy; dice; inverse dice; right ventricle.
  DOI: 10.1504/IJBRA.2025.10065947
   
  
• Long Short-Term Memory based Model Predictive Control of Blood Glucose Level for Type 1 Diabetes Mellitus Treatment  
  by Nitesh Kumar Barnawal, Hoo Sang Ko, Sarah Park, H. Felix Lee, Guim Kwon 
  Abstract: This paper presents a novel method to control blood glucose levels (BGL) based on predictions made by a long short-term memory (LSTM) network. An initial LSTM model was trained with data from rats with type 1 diabetes mellitus (T1DM) using Open Artificial Pancreas System (OpenAPS). Transfer learning was applied to develop an individualised prediction model based on the initial model. The LSTM model predicted BGL with a root mean squared error (RMSE) of 11.8240 mg/dl. The model was integrated into model predictive control (LSTM-MPC), which optimised insulin injection based on BGL predictions. Evaluated against a neural network-based MPC (NN-MPC) and OpenAPS using different diets and rats, LSTM-MPC outperformed both in control performance. This study demonstrated a closed-loop BGL control system tested with in vivo diabetic rats. The prediction model is re-trainable quickly using small datasets obtained from individual rats, which provides a feasible solution for individualised T1DM treatment.
  Keywords: LSTM; blood glucose level prediction; type 1 diabetes mellitus (T1DM); model predictive control; transfer learning; time series forecasting; artificial pancreas system..
  DOI: 10.1504/IJBRA.2025.10065994
   
  
• Fine-Tuning Predictive Models: A Comprehensive Analysis for Accurate Diabetes Risk Stratification  
  by Nuzhat Yatoo, I. Sathik Ali 
  Abstract: Diabetes is a major global health concern since it causes serious complications like kidney disease, heart problems, and eyesight loss. In pursuit of accurate disease diagnosis machine learning (ML) methods have been employed resulting in favourable outcomes. In this study, an innovative diabetes prediction model is introduced that incorporates a comparison of various ML techniques, including logistic regression, K-nearest neighbour, naive Bayes, decision tree, and CatBoost on a diabetes database in order to improve on existing systems for disease prediction, specifically concerned with Diabetes by establishing the best performing model based on performance metrics such as accuracy, recall, precision, F1 score, Mathews correlation coefficient (MCC), Cohen Kappa, index of agreement and area under the curve (AUC). To optimise their results the techniques are subjected to hyperparameter tuning. The metric values thus obtained from the proposed methodology establish CatBoost as the best performing model and, hence, the most viable for diabetes prediction.
  Keywords: diabetes; machine learning; feature selection; SMOTE Tomek; hyper parameter optimisation; prediction.
  DOI: 10.1504/IJBRA.2025.10066073
   
  
• LARSE: Level-Based Associated Residual Network with Squeeze-and-Excitation for Breast Cancer Detection and Classification  
  by A.N.U. Rakhi P. S, Rajesh R. S 
  Abstract: Breast cancer is considered as a serious disease causing a high mortality rate amongst women. In recent years, computer aided diagnosis (CAD) techniques have the radiologists to make proper decisions on mammograms more accurately. The existing CAD method may not contribute significant results for the early identification of breast mass especially at stages 1 and 2. This work introduces a level-based associated residual network with squeeze-and-excitation (LARSE) block for breast cancer classification. Initially, the input image undergoes pre-processing using the contrast limited adaptive histogram equalisation (CLAHE) model. Then, the feature extraction process is done by utilising a dual ResNet-based feature extraction model, LARSE. The LARSE model is used for multilevel breast cancer classification based on BI.RADS categories, tested on CBIS.DDSM and INbreast mammogram datasets. The LARSE model achieved an accuracy of 96.9% (
  Keywords: mammography; CAD; breast cancer; classification; deep learning.
  DOI: 10.1504/IJBRA.2025.10066341
   
  
• Enhancing Data Rate Efficiency in Multi-Cell Massive MIMO Systems Through Pilot Resource Allocation  
  by S. Mishra, Betelhem Gudina, Demissie Jobir, Davinder Rathee, Bijay Paikaray 
  Abstract: Pilot contamination (PC) limits the system's performance and highly reduces the data rates. To overcome this problem and improve the data rate of a multicell massive MIMO system, a modified soft pilot reuse (mSPR) is proposed. For clustered users, mSPR assigns the same pilot set to low PC severity centre zone users, reusing them in a grouped cell. Centre zone users are randomly assigned pilots from the same group, with the maximum user serving as the pilot reference for orthogonal allocation. The simulation result shows the data rate improvement of the multi-cell multi-user massive MIMO system with the proposed mSPR pilot resource allocation. The number of orthogonal pilots is reduced, improving the achievable uplink rate and spectral efficiency over pre-existing schemes. With different numbers of base station antennas and transmission power, the proposed mSPR pilot resource allocation scheme provides average data rates of 3.5 bps/Hz and 2.78 bps/Hz, respectively.
  Keywords: pilot contamination; massive MIMO; modified soft pilot reuse; mSPR; time-division duplex.
  DOI: 10.1504/IJBRA.2025.10066367
   
  
• ResNet-Based Deep Learning Approach for Automated ECG Arrhythmia Recognition System  
  by Soumen Ghosh, Satish Chander 
  Abstract: This study presents approach to signal classification, focusing on electrocardiogram signals using deep learning techniques Leveraging TensorFlow and ResNet50 architecture, research aims to develop robust model for accurate classification ECG signals, crucial various medical diagnostics ,healthcare applications Methodology involves pre-processing ECG dataset, constructing deep neural network model based on ResNet50, and training model on labeled dataset Subsequently, trained model undergoes comprehensive evaluation using performance metrics such as accuracy, confusion matrix analysis, classification report, F1-score, and ROC curve analysis The results demonstrate the effectiveness of proposed approach in accurately classifying ECG signals, showcasing its potential for enhancing medical diagnostics and improving patient care In this study we achieved 92 71% accuracy score by using proposed approach This research contributes to advancing field of signal classification in healthcare, offering a promising methodology for automated analysis and interpretation of ECG signals, ultimately aiding healthcare professionals in timely diagnosis and treatment of cardiovascular conditions.
  Keywords: Deep learning; Signal classification; Electrocardiogram (ECG); ResNet50; TensorFlow; Medical diagnostics; Healthcare; Accuracy; Performance evaluation; Confusion matrix; Classification report.
  DOI: 10.1504/IJBRA.2025.10066430
   
  
• Deep Feature Fusion and Ensemble Learning to Create an Effective CNN Brain Tumour Classification Model  
  by Sathees Kumar  
  Abstract: Early brain tumour exploration can streamline treatment. Some automated diagnosis system aids radiologists in distinguishing between normal and abnormal brain tissues, simplifying clinical and diagnostic processes. However, categorizing MRI images is challenging due to low contrast, noise, tumour shape and localisation dissimilarity, and similarity between ordinary and cancerous regions of interest (ROIs). This study uses a deep convolutional neural network with feature blending and ensemble learning to analyse MRI abnormalities, followed by detection and classification tasks. The ensemble learning method effectively distinguishes between ordinary and cancerous tumour ROIs, yielding reliable results. Feature fusion identifies discriminative features between classes. To address overfitting in smaller data sets, depth-wise separable convolution and spatial drop-out techniques are explored for MRI brain image classification. The proposed approach has been validated on two freely available datasets, Kaggle and BrATS, with the BrATS dataset showing superior outcomes in accuracy, specificity, and sensitivity (0.995, 0.996, 0.996).
  Keywords: Brain tumour; Feature Fusion; (CAD) Computer-Aided Diagnosis; Ensemble Learning; Brain tumour classification; Regions of Interest (ROIs); Psychological Health; Malignant Primary Brain Tumors,.
  DOI: 10.1504/IJBRA.2025.10066432
   
  
• Unmasking Poly Cystic Ovarian Syndrome(PCOS): Harnessing Deep Learning in Ultrasound Imaging Analysis  
  by Nusrath Fathima, Pradeep Kumar 
  Abstract: A large proportion of women globally suffer from PCOS, a hormonal condition that impacts reproductive health and poses major dangers to their metabolic and cardiovascular health. PCOS diagnosis at an early stage is crucial to mitigate these risks and provide timely interventions. The challenge in diagnosing the PCOS is to count the follicles and calculate their volume in the ovaries, which is currently done manually by doctors and radiologists utilising ovary ultrasonography. In this study, a shallow robust deep learning model is proposed with three alternate convolution and max pooling layers followed by flatten, dropout and dense layer that automatically detects PCOS from ultrasound images with low computational complexity. The performance of the proposed model is compared with the Inception V3 and Dense Net 201 deep learning models. The benchmark PCOS dataset from Kaggle was used for the study and dataset was split as 70:30 for training and testing. In conclusion, our study highlights the potential of deep learning in the field of gynaecology and reproductive medicine. It can revolutionise PCOS diagnosis and contribute to better health outcomes for women with PCOS.
  Keywords: PCOS; Deep Learning; CNN ; Ultrasound Images; Medical Imaging; Early Diagnosis; Automated Detection.
  DOI: 10.1504/IJBRA.2025.10066434
   
  
• An Optimised InceptionResNetV2 Model for Breast Cancer Histopathology Image Classification  
  by Keren Evangeline I, Glory Precious J, Anand C. D, Angeline Kirubha S. P. 
  Abstract: Breast cancer usually develops in women due to uncontrolled cell division. The clinical gold standard for diagnosing this disease is breast histopathology. Automating breast cancer detection saves time and aids pathologists. Deep learning is vital here. This study investigates utilising convolutional neural networks and transfer learning to identify breast cancer from histopathological image patches of all magnifications. Thus, an optimised deep learning model for breast cancer image classification was created by adding and modifying InceptionResNetV2 layers. Transfer learning was used to train and fine-tune it. The model was then compared to VGG-16, DenseNet-121, and original InceptionResNetV2 networks. The optimised InceptionResNetV2 model outperforms all other models with images of all magnification factors. For 400X magnification image classification, the optimised InceptionResNetV2 model has the maximum accuracy of 98%. Hence, the model predicts benign and malignant cancer image patches more accurately.
  Keywords: image patches; optimised InceptionResNetV2; deep learning; diagnosis; histopathology; transfer learning; breast cancer.
  DOI: 10.1504/IJBRA.2025.10066542
   
  
• An improved Machine Learning Based System for Depression Detection with RFLR Model  
  by S.Nalini Poornima, S. Geetha 
  Abstract: The world is changing at a dizzying pace due to technological advancements and improved human abilities. Physical and emotional well-being suffer due to the immense strain of keeping up with the fast-paced society around us. Depression is a prevalent mental condition that affects everyone at some time. Globally, millions of individuals suffer from depression, making it one of the most prevalent mental illnesses. Prolonged and excessive fretting about several issues that a healthy person would often dismiss as unimportant characterises depression. Machine learning algorithms are crucial for deciphering healthcare data and revealing hidden information. In the investigated approach, a hybrid model was utilised to combine RF and LR using a Voting Classifier to construct a depression prediction model. After acquiring a suitable dataset from Kaggle, the suggested method for depression prediction moves on to pre-processing, where data is cleaned and scaled to guarantee consistency and quality. The proposed model is trained and tested with the b_depressed.csv dataset obtained from a Kaggle source with 1,767 records. The model demonstrates superior performance in both accuracy and overall effectiveness compared to other models, as indicated by the findings.
  Keywords: Depression Prediction; Classification; Preprocessing; Random Forest (RF); Logistic Regression (LR); Machine Learning Based System; Depression detection; RFLR model.
  DOI: 10.1504/IJBRA.2025.10066894
   
  
• An Efficient MedicalEmergency Prediction ModelFrom Unstructured Medical Transcript Using Dual Attentional Bilstm Model  
  by Amita Mishra, Sunita Soni 
  Abstract: This study tackles the challenge of predicting medical emergencies from unstructured transcripts, enhancing clinical decision-making. Employing a dual attention Bi-LSTM model, it integrates pre-processing, feature extraction, and summarisation steps. A weighted hybrid distance-based graph embedding technique captures relevant features, and a fine-tuned BERT model with dual attention effectively summarises the text. Additionally, an adaptive rule-based Bi-LSTM captures temporal connections and contextual information. The model performs exceptionally well with 96.44% accuracy, 98.18% sensitivity, 95.95% specificity, and a 96.42% F1-score for classification. Summarisation results surpass existing methods with BLEU 0.43, CIDER 0.74, METEOR 0.22, ROUGE 0.53, and SPICE 19.74. This dual attention Bi-LSTM model holds promise for enhancing clinical workflows by extracting critical information from medical transcripts.
  Keywords: deep Bi-LSTM; dual attention Bi-LSTM; BERT; TF-IDF; weighted hybrid distance-based graph embedding.
  DOI: 10.1504/IJBRA.2025.10067237
   
  
• Multiepitope Vaccine Design: a Promising in Silico Strategy to Combat the Threat of Emerging Viruses like Zwiesel bat Banyangvirus  
  by Amzad Hossain, Forsan Amin, Md Rakibul Islam 
  Abstract: The emerging bat virus, Zwiesel bat banyangvirus (ZbbV), was recently discovered. This virus family causes severe clinical symptoms and many deaths. Since this virus has spread to almost all continents, preventive measures are necessary. Current antiviral treatments cannot alleviate banyangvirus's clinical complexities, so this study aims to develop a multiepitope vaccine using ZbbV's nucleocapsid sequence and its conserved regions with other family members. The 329-amino-acid multiepitope vaccine includes B-, TH-, and TC-cell epitopes linked by flexible linkers. The vaccine produces stable mRNA and is non-allergenic, non-toxic, and immunogenic. Its structural strengths include temperature resistance, flexibility, mammalian physiological suitability, and humoral and cellular immune elicitation. The vaccine binds TLR4 and MHC with high binding energy. Molecular dynamic simulation assessed the vaccine's flexibility and stability. An enhanced in silico cloning method in a constitutive expression vector could help purify the vaccine. This research is essential to combating the imminent Zwiesel bat banyangvirus.
  Keywords: Zwiesel bat banyangvirus; multiepitope vaccine; immunoinformatics; vaccine design; emerging virus.
  DOI: 10.1504/IJBRA.2025.10067253
   
  
• A Mobile Healthcare: Empowering Chronic Kidney Disease Management in Eastern Economic Corridor Thailand  
  by Boy Xayavong, Supet Jirakajohnkool, Nattadon Pannucharoenwong, Wachirathorn Janchompu, Damrongrit Niammuad, Kammal Kumar Pawa 
  Abstract: Chronic kidney disease (CKD) is a global health issue with widespread implications for public health. Primarily stemming from non-communicable diseases such as hypertension, heart disease, stroke, and diabetes, Studies have shown that in the eastern economic corridor there is a high risk of chronic non-communicable diseases has seen a substantial increase in mortality. Researchers have developed the Ticare mobile health application (mHealth) for CKD patients. The primary objective of this research is to create a model mHealth app focusing on improving public access to healthcare. Data collection for app development was carried out through surveys, interviews, and insights from healthcare experts. The Ticare app is compatible with all operating systems and offers key functionalities, including a disease news information system, hospital search and navigation, online service reservation, telemedicine, in-app medication ordering, and data visualisation. The satisfaction survey involving 400 participants revealed a high level of user satisfaction.
  Keywords: non-communicable diseases; NCDs; eastern economic corridor; EEC; mHealth; Thailand.
  DOI: 10.1504/IJBRA.2026.10067444
   
  
• Unravelling the Functionally Enriched HUB Gene Targets of Alzheimer's Disease: a Bioinformatics Approach  
  by Polani Ramesh Babu, Kumaravel Appavoo, Subbiah Suresh Kumar, Prabhu Manickam Natarajan 
  Abstract: Alzheimer disease (AD) is a complex neurological disorder for which suitable therapeutic approaches are not yet developed. Identifying a crucial functional protein using gene ontology (GO) tools and protein protein interaction (PPI) analysis is a promising approach in finding a potential protein drug target. We selected 48 crucial genes previously proven as HUB genes associated with AD symptoms and ranked them based on their functional enrichments analysis. We identified 14 HUB genes with high functional enrichments scores in carbohydrate metabolism associated with gluconeogenesis, oxidative phosphorylation, energy metabolisms, a few signaling pathways, MTOR, AMPK, H1F1 pathways in certain infections and cancers. PPI analysis across different GO databases indicated ENO1, ENO2, PFKM, GP1, ALDOC, MDH1 and GAPDH as top ranked targetable HUB genes majorly associated with the abnormalities of neuronal glucose metabolism in the early onset of AD, which could potentially be used as therapeutic targets in disease management of AD.
  Keywords: Alzheimer's; HUB genes; gene ontology; protein protein interactions and protein targets.
  DOI: 10.1504/IJBRA.2025.10067447
   
  
• In Vitro and in Silico Analysis of Marine Actinobacterial Bioactive Compounds: Inhibitors of Cancer  
  by N.S. Swarnakumar, Kalaivanan Uma Nageswari, Elangovan Dilipan 
  Abstract: Marine actinobacteria are important screening targets because of their variety and propensity to create new metabolites and other pharmaceutical molecules. The study used Hep2 cancer cell lines and Vero normal cell lines to test the cytotoxicity of the SNI5 and SNI10 strains of actinobacteria. The ethyl acetate extract of two strains of actinobacteria was tested, and the most effective extract was put through GC-MS profiling of metabolites. The discovered compounds were docked against an O-glycosylated (MUC16) protein molecule, and their protein-ligand interactions were investigated. Extracts of the strain SNI5 showed 100% cytotoxicity only on the Hep2 cells. However, the SNI10 extract showed 100% cytotoxicity both in Hep2 cells and Vero cells at all concentrations (500 to 65.2
  Keywords: Marine actinobacteria; cytotoxicity; GCMS; molecular docking; cancer.
  DOI: 10.1504/IJBRA.2025.10067528
   
  
• HCFPC: A New Hybrid Clustering Framework using Partition-Based Clustering Algorithms to Group Functionally similar Genes from Microarray Gene Expression Data  
  by Shilpi Bose, Chandra Das, Aishwariya Barik, Kuntal Ghosh, Matangini Chattopadhyay, Samiran Chattopadhyay 
  Abstract: Clustering of genes from microarray gene expression data is one of the important analysis tasks as it has a great impact on the functionality prediction of novel genes and also in gene regulatory network formation. Partition-based clustering algorithms are the most popular and most of the time give consistent results but these methods cannot handle noise points. Apart from this, hard-k-means cannot handle overlapping data. Although fuzzy-k-means and possibilistic-k-means can handle overlapping data, these methods have several shortcomings related to their membership function. In this regard, here, a hybrid clustering framework using different partition-based clustering algorithms (hard-k-means, fuzzy-k-means, and possibilistic-k-means) named HCFPC is proposed to find clusters from overlapping objects via eliminating noise from highly noisy datasets. It also can minimise different drawbacks that appear in fuzzy and possibilistic approaches. To validate this method, it is applied to highly noisy gene expression datasets to find clusters from overlapping genes via eliminating noise. Experimental results show the efficiency of the proposed hybrid framework.
  Keywords: clustering; k-means; fuzzy-k-means; possibilistic k-means; microarray technology; gene expression data.
  DOI: 10.1504/IJBRA.2026.10067605
   
  
• Breast Cancer Detection and Classification using Optimisation enabled Deep Learning Model  
  by Dhanya Mathew, Vijula Grace K.S., Mary Synthuja Jain Preetha 
  Abstract: This work aspires to discover a scheme for breast cancer disease detection and presents the integration of deep learning methods to minimise breast cancer risk in women's lives. The initial phase involves pre-processing the images through contrast enhancement and resizing. Subsequently, the pre-processed outputs undergo data augmentation, which has become a significant focus in recent deep learning domain. Subsequently, for segmenting cancer region in the augmented images, the MIScnn scheme was employed. Finally, the breast cancer detection process is performed using the Hybrid LeNet-ZFNet, which is the combination of LeNet and ZFNet scheme, to classify as either normal or malignant. Finally, experimental analysis revealed that proposed approach achieved maximal accuracy of 95.8%, sensitivity of 97.3%, and specificity of 93.3%.
  Keywords: Breast cancer; detection; Deep Learning; classification; MIScnn.
  DOI: 10.1504/IJBRA.2026.10067700
   
  
• Expression of flmA and csgA genes in Uropathogenic Escherichia Coli Treated with Supernatant of Lactobacillus Plantarum plus Gentamicin  
  by Mohamad Kamel Koodi, Asra’a Adnan Abdul-Jalil, Laith Muslih Najeeb 
  Abstract: A total of 422 patients with type 2 diabetes mellitus (T2DM) were analysed via standard microbiological techniques to detect the presence of Escherichia coli. This study aimed to assess UPEC isolates' adhesion to urothelial cells and their ability to produce haemolysin via sheep blood agar. The biofilm formation of E. coli isolates was investigated. The impact of each antibiotic, both individually and in combination, on biofilm gene expression was investigated. A total of 26.19% of the isolates were identified as E. coli. The minimum inhibitory concentration (MIC) for the CFS of Lactobacillus was 12.5 U/ml. The results indicated that all strains of Escherichia coli exhibited a 50% success rate in sticking to epithelial cells, and all isolates demonstrated the ability to form biofilms. In response to the combination of CFS and antibiotics, RT-PCR analysis revealed significant suppression of biofilm-producing genes (fimA and csgA) compared to those in response to monotherapy.
  Keywords: biofilm; E. coli; Lactobacillus plantarum; flmA; csgA; synergistic; gene expression; uropathogenic; supernatant.
  DOI: 10.1504/IJBRA.2026.10067789
   
  
• Antibiotic Genomic Resistance Prediction Using Deep Learning Models  
  by Shifana Rayesha, W. Aisha Banu, Afzalur Rahman 
  Abstract: The classification of antibiotic resistance genomes presents a substantial challenge in the field of computational biomedical data analysis. Several machine-learning techniques have been used to solve this challenge in recent years. However, training data that is not independent and identically distributed makes taught models prone to out-of-distribution generalisation difficulties, making it a major challenge. Antibiotic resistance data generally match observations from similar phylogenetic datasets, making this important. To identify antibiotic resistance in drug discovery, antimicrobial characteristics must be extracted from large datasets. This abstract discusses retrieving antibiotic-resistance genes from the 89,491-feature pseudomonas aeruginosa integrin nucleotide sequence. For data separation and processing, 1D CNNs and ANNs were used to handle this complex text. The model extracts antimicrobial resistance genomes efficiently in convolution neural networks. The 1D convolutional neural network has 98.85% efficiency, while the artificial neural network has 80.46 %. This article uses the convolutional neural network, which extracts antibiotic resistance genomic information well. These findings could help us understand antibiotic resistance and improve medication discovery. These cutting-edge machine-learning approaches provide hope in the fight against emerging microbial dangers in an era of antibiotic resistance.
  Keywords: machine-learning; 1D convolutional neural networks; CNN; artificial neural networks; ANN; phylogenetic; nucleotide.
  DOI: 10.1504/IJBRA.2025.10067903