Improving Accuracy and Efficiency of Gene Expression Data Analysis Using an Ensemble of GSOM and Hierarchical Clustering
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Clustering of gene expression data is essential for identifying co-expressed genes, functional modules, and disease biomarkers. Traditional clustering methods, such as hierarchical clustering and k-means, often struggle with the high dimensionality and noise in gene expression datasets, limiting their biological interpretability. This paper presents GSOM+HClust, a framework that integrates the Growing Self-Organizing Map (GSOM) with hierarchical clustering (HClust) to improve the accuracy and efficiency of gene expression data analysis. The framework is evaluated on the Alzheimer’s disease gene expression dataset (GSE5281) and compared against raw hierarchical clustering, PCA+HClust, and SOM+HClust. Results show that GSOM+HClust achieves superior clustering quality across multiple indices (CCC = 0.962, Silhouette = 0.935, DBI = 0.145) while reducing runtime from several hours to under 20 minutes. Furthermore, all 14 known Alzheimer’s biomarkers were clustered into coherent groups, providing strong biological validation. Findings of this work establish GSOM+HClust as a scalable, interpretable framework for biomarker discovery and disease stratification, providing a practical tool for large-scale gene expression analysis.
Fields of Study & Subjects
Gene expression and cancer classification
Bioinformatics and Genomic Networks
Advanced Clustering Algorithms Research
Publication details
- Journal / source
- IEEE International Conference on Industrial and Information Systems (ICIIS)
- Pages
- 460–465
- Publication date
- 16 January 2026
- Language
- EN