Bangabandhu Sheikh Mujibur Rahman Science and Technology University

Blood cell classification and counting are vital for the diagnosis of various blood-related diseases, such as anemia, leukemia, lymphoma, and thrombocytopenia. The manual process of blood cell classification and counting is time-consuming, prone to errors, and labor-intensive. Therefore, we have proposed a deep learning (DL)-based automated system for blood cell classification and counting from microscopic blood smear images. We classify a total of nine types of blood cells, including Erythrocyte, Erythroblast, Neutrophil, Basophil, Eosinophil, Lymphocyte, Monocyte, Immature Granulocytes, and Platelet. Several preprocessing steps like image resizing, rescaling, contrast enhancement and augmentation are utilized. To segment the blood cells from the entire microscopic images, we employed the U-Net model. This segmentation technique aids in extracting the region of interest (ROI) by removing complex and noisy background elements. Both pixel-level metrics such as accuracy, precision, and sensitivity, and object-level evaluation metrics like Intersection over Union (IOU) and Dice coefficient are considered to comprehensively evaluate the performance of the U-Net model. The segmentation model achieved impressive performance metrics, including 98.23% accuracy, 98.40% precision, 98.26% sensitivity, 95.97% Intersection over Union (IOU), and 97.92% Dice coefficient. Subsequently, a watershed algorithm is applied to the segmented images to separate overlapped blood cells and extract individual cells. We have proposed a BloodCell-Net approach incorporated with custom light weight convolutional neural network (LWCNN) for classifying individual blood cells into nine types. Comprehensive evaluation of the classifier’s performance is conducted using metrics including accuracy, precision, recall, and F1 score. The classifier achieved an average accuracy of 97.10%, precision of 97.19%, recall of 97.01%, and F1 score of 97.10%. A 5-fold cross-validation technique is applied to split the data, which not only aids in reducing overfitting but also helps in generalizing the model.

Microarray gene expression data are often classified by cell line or tumor type to assist in the diagnosis and prediction of human cancer. While microarray analysis demonstrates potential, choosing the most suitable machine learning approach is crucial for accurate cancer diagnosis and prediction. In this beginner’s guide, we outline how to select an optimal machine learning model for cancer phenotype prediction by comparing various existing methods. This study used three well-known machine learning methods: linear discriminant analysis, support vector machines, and random forest. To assess prediction performance, several performance metrics were considered, including model prediction accuracy (AC), the area under the curve (AUC), F-measure, the receiver operating characteristic (ROC) curve, and the precision-recall curve (PRC). Microarray gene expression data from two cancer types, leukemia and colon cancer were analysed. A cross-validation process with 100 resampling iterations was implemented to compute average performance measures (APM), ensuring the reliability of the results. Findings emphasize the significance of selecting the right machine learning model for accurate predictions of new samples. The methods employed provided satisfactory results, validated by various APM for both leukemia and colon cancer datasets. Notably, the random forest classifier exhibited the best performance in cancer prediction. International Journal of Statistical Sciences, Vol. 25(2), November, 2025, pp 47-57

Gigantol is a biphenyl natural compound obtained from various medicinal orchids. This study sought to evaluate gigantol’s anticancer effects, the mechanisms underlying this effect, and its pharmacokinetic characteristics. Gigantol exhibits potential anticancer properties in preclinical pharmacological test systems that are mediated via a range of molecular mechanisms, including the induction of cytotoxicity, Apoptosis signalling pathway activation, Wnt/β-catenin signalling pathway inhibition, HSP90/Akt/CDK1 signalling pathway suppression, PI3K/AKT/mTOR signalling pathway modulation, interfering with JAK/STAT signalling pathway, MAPK signalling pathway inhibition, and MYC signalling pathway modulation. Further, the research emphasised the pharmacokinetic profile of gigantol, which is rapidly absorbed upon oral administration and primarily excreted from the body via metabolic processes. Thus, gigantol shows promising potential as a chemotherapeutic agent. However, in-depth clinical investigations are needed to determine its safety, efficacy, and therapeutic potential in cancer.

ABSTRACT Microbial inoculants that integrate plant growth promotion, pathogen suppression, and pollutant degradation are central to sustainable agriculture. Here, we report the genomic and phenotypic characterization of Lysinibacillus capsici BCSIR-Raj-01, an endophyte isolated from citrus leaves in Bangladesh. The 4.62 Mb genome encodes 4,699 coding sequences and displays an open pangenome with loci linked to stress adaptation, xenobiotic degradation, and secondary metabolite biosynthesis. Genome mining revealed complete pathways for indole-3-acetic acid (IAA) production, siderophore biosynthesis, and phosphate metabolism, alongside gene clusters for bacteriocins, lantibiotics, terpenes, and polyketides. Nitrogen cycling genes indicated diazotrophic growth potential and a truncated denitrification pathway. Phenotypic assays validated these predictions: L. capsici BCSIR-Raj-01 solubilized phosphate, produced IAA and siderophores, grew on hydrocarbons, and exhibited diazotrophic activity. Notably, the strain showed potent larvicidal activity against Culex and Anopheles larvae, suggesting ecological roles beyond plant association. Together, these results establish L. capsici BCSIR-Raj-01 as a multifunctional bacterium with potential applications in crop improvement, bioremediation, and biological vector control. IMPORTANCE The challenges of modern agriculture and environmental sustainability demand microbial solutions that deliver multiple functions. Lysinibacillus species are increasingly recognized for such versatility, but their genomic basis remains poorly resolved. Our study provides the first genome-enabled framework for understanding multifunctionality in L. capsici . By combining whole-genome sequencing with phenotypic validation, we show that BCSIR-Raj-01 harbors biosynthetic clusters, stress adaptation loci, and xenobiotic degradation genes while also expressing experimentally confirmed traits, including phosphate solubilization, IAA and siderophore production, nitrogen fixation, hydrocarbon utilization, and larvicidal activity. This integration of plant growth promotion, biocontrol potential, and pollutant degradation positions L. capsici BCSIR-Raj-01 as a next-generation bioinoculant. Beyond broadening the functional scope of Lysinibacillus , our findings highlight how genome-guided approaches can identify microbes with cross-sectoral applications in agriculture, bioremediation, and vector management.

Climate change and global warming overshadow concerns about the sustainability of human existence and progress. As a result, there have been numerous worldwide efforts to reduce the negative effects of climate change and promote a sustainable future. Companies use corporate carbon accounting (CCA) as a key strategy to adapt to climate impacts. Therefore, this study aims to evaluate the current state of CCA practices in Bangladesh. The authors gathered secondary sources of information for the study. The researchers analysed the annual reports of 154 companies across 11 industries listed on the Dhaka Stock Exchange (DSE). The authors developed the 'Corporate Carbon Accounting Practice Index (CCAP)' based on existing literature and international standards. This study used content and thematic analysis, following specific criteria. The results indicated that a mandatory legal framework influences CCA practices. However, most companies have not yet implemented CCA procedures, with only a few meeting certain standards. The main reason for Bangladesh's current CCA situation is the absence of a strong legislative framework. This study provides specific guidance for policymakers interested in CCA, helping them identify the best actions to mitigate climate change risks and support sustainable development, including establishing and enforcing a robust mandatory legal framework.

Discovery of the genus Tarenna campaniflora and its traditional uses was obtained by living in forest or semi-forest areas and by close observations of the indigenous populations. This study evaluated the phytochemical constituents, antioxidant potential, antimicrobial efficacy, and thrombolytic activity of ethanolic leaf extract of Tarenna campaniflora. Qualitative phytochemical screening revealed the presence of various bioactive metabolites. DPPH assay demonstrated notable antioxidant activity (IC50 = 0.877 µg/mL), comparable to ascorbic acid (IC50 = 0.839 µg/mL). Disc diffusion assay showed antimicrobial activity, with the highest inhibition against Escherichia coli (18 mm). The extract also exhibited clot lysis (22.42%), compared to streptokinase (31.17%). The findings support the therapeutic potential of T. campaniflora, warranting further pharmacological investigations.

Mental health issues, including depression, anxiety, and insomnia, are increasingly prevalent among university students and graduates, especially those involved in academic research. The impact of research-related characteristics on mental health remains underexplored. We examined this relationship using machine learning alongside traditional statistical analyses and GIS mapping. Data from 508 university students and graduates were collected, and encompassed socio-demographics, academic information, research related information, and mental health outcomes. Statistical analyses were performed using SPSS, while spatial analysis was conducted using QGIS and machine learning models were developed with Python with Google Colab. High prevalence rates of depression (39.8%), anxiety (29.3%), and insomnia (12.2%) emerged. Feature selection highlighted research experience (excluding thesis), research courses during the bachelor's program, and interest in a research-related career as significant predictors of mental health outcomes. CatBoost modeling performed best in accuracy and precision of risk prediction of mental health conditions. Support Vector Machine model performed well in predicting depression, while Random Forest showed consistent low log loss, indicating better calibration across mental health issues. GIS mapping revealed no significant regional heterogeneity in mental health outcomes. Research-related factors, such as research experience and academic pressures, significantly impact the mental health of university students and graduates. Machine learning models may enable institutions to more effectively identify at-risk students and provide personalized support to foster a supportive research environment, ultimately improving both mental health outcomes and academic success.