Related Experiment Video
Updated: Feb 28, 2026
08:14
MicroRNA Based Liquid Biopsy: The Experience of the Plasma miRNA Signature Classifier MSC for Lung Cancer Screening
Published on: October 26, 2017
16.3K
miRNA-Based Breast Cancer Subtyping Using AHALA Multi-Stage Classification Approach.
Mohammed Qaraad1, Eric P Rahrmann1, David Guinovart1
1The Hormel Institute, University of Minnesota, 801 16th Ave NE, Austin, MN 55912, USA.
Cancers
|February 27, 2026
Summary
A new algorithm, Adaptive Hill Climbing Artificial Lemming Algorithm (AHALA), accurately subtypes breast cancer using microRNA (miRNA) expression. This precision medicine approach identifies key biomarkers for improved diagnosis and treatment.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Breast cancer is molecularly heterogeneous, necessitating precise subtyping for effective precision medicine.
- MicroRNA (miRNA) expression profiles offer potential for breast cancer subtyping but face challenges in feature selection and algorithm optimization.
- Accurate subtyping is crucial for tailoring treatments to individual breast cancer molecular profiles.
Purpose of the Study:
- To develop and validate a novel optimization framework for accurate miRNA-based breast cancer subtyping.
- To enhance the potential of miRNAs as diagnostic biomarkers through advanced feature selection and machine learning.
- To improve the precision of breast cancer classification for personalized treatment strategies.
Main Methods:
- Proposed the Adaptive Hill Climbing Artificial Lemming Algorithm (AHALA), a hybrid optimization framework.
- Applied low-variance filtering and differential gene expression analysis for dimensionality reduction.
- Utilized AHALA to optimize deep neural network hyperparameters for multi-class breast cancer subtype classification using miRNA data.
- Validated the method on TCGA breast cancer miRNA expression data and benchmarked against other optimization algorithms.
Main Results:
- AHALA achieved high classification performance: 95.74% accuracy, 95.98% precision, 95.74% recall, 95.74% F1 score, and 0.9682 AUC.
- The algorithm demonstrated superior convergence and significance compared to existing optimization methods.
- Identified specific miRNAs (e.g., hsa-miR-190b, hsa-miR-429) associated with distinct breast cancer subtypes.
Conclusions:
- The AHALA framework provides an efficient and potent method for miRNA-based breast cancer subtyping.
- The algorithm's integration of global exploration and local search enhances classification performance and stability.
- AHALA effectively identifies biologically significant biomarkers, marking it as a promising tool for breast cancer diagnostics.
