Selective Omics Fusion with Slime Mould Algorithm-Based Feature Selection for Subtype Classification in Breast Cancer
Procedia Computer Science · 2026 · 283 · 5443-5453
DOI: 10.1016/j.procs.2026.06.594 ↗Kurz gefasst
Fachartikel in der Zeitschrift Procedia Computer Science, 2026. Verfasst von Ayushi Singh und 2 weiteren Autor:innen. Die Arbeit nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Breast cancer is a complex illness with several molecular subtypes, each needing a unique therapeutic strategy. The accurate classification of these subgroups is crucial for better diagnosis and personalized treatment. This paper proposes a deep learning approach that employs multi-omics data: mRNA expression, DNA methylation, and copy number alteration (CNA) to classify breast cancer into five major subtypes: Basal, Her2, LumA, LumB, and Normal. To reduce data dimensionality and improve learning, we used the Slime Mould Algorithm (SMA) for feature selection, choosing biologically important features from each omics layer. A multibranch neural network is used to create a selective omics fusion model, in which each type of omics is processed independently before being fused. Selective masking is employed to omit less valuable omics data for specific subtypes, including Her2, Normal, and LumB. The model is trained and tested using the TCGA BRCA dataset, and it achieved a final test accuracy of 98.98%, with F1-scores of 1.00 for Basal, LumA, and LumB, indicating excellent subtype categorization. SMA is also tested against other metaheuristic feature selection methods such as PSO, DE, and GWO, and it is found to be effective at selecting interpretable and discriminative features. This study demonstrates that combining SMA with selective fusion increases both performance and biological interpretability, resulting in a reliable method for multi-omics-based breast cancer subtype categorization.
Zitieren
APA
Singh, A., Sikarwar, G., & Vijay, R. K. (2026). Selective Omics Fusion with Slime Mould Algorithm-Based Feature Selection for Subtype Classification in Breast Cancer. Procedia Computer Science, 283, 5443-5453. https://doi.org/10.1016/j.procs.2026.06.594
Vancouver
Singh A, Sikarwar G, Vijay RK. Selective Omics Fusion with Slime Mould Algorithm-Based Feature Selection for Subtype Classification in Breast Cancer. Procedia Computer Science. 2026;283:5443-5453. doi:10.1016/j.procs.2026.06.594
BibTeX
@article{Singh2026,
author = {Singh, Ayushi and Sikarwar, Gargi and Vijay, Rahul Kumar},
title = {Selective Omics Fusion with Slime Mould Algorithm-Based Feature Selection for Subtype Classification in Breast Cancer},
journal = {Procedia Computer Science},
year = {2026},
volume = {283},
pages = {5443--5453},
doi = {10.1016/j.procs.2026.06.594}
}
EUVIMED