Related Experiment Video
Updated: May 12, 2026
08:40
Methyl-binding DNA capture Sequencing for Patient Tissues
Published on: October 31, 2016
8.6K
EpiBrCan-Lite: A lightweight deep learning model for breast cancer subtype classification using epigenomic data.
Punam Bedi1, Surbhi Rani1, Bhavna Gupta2
1Department of Computer Science, University of Delhi, Delhi, India.
Computer Methods and Programs in Biomedicine
|December 12, 2024
Summary
A new lightweight model, EpiBrCan-Lite, accurately classifies breast cancer subtypes using DNA methylation data. This approach significantly reduces trainable weight parameters while maintaining high performance, addressing limitations of existing methods.
Area of Science:
- Computational Biology and Bioinformatics
- Machine Learning in Oncology
- Genomic Data Analysis
Background:
- Accurate breast cancer subtype classification is crucial for patient prognosis and survival rates.
- Existing Machine Learning and Deep Learning models often suffer from high trainable weight parameters, low performance, and class imbalance issues.
- DNA methylation data offers a promising avenue for subtype classification but requires efficient analytical models.
Purpose of the Study:
- To develop a lightweight model for breast cancer subtype classification using DNA methylation data.
- To address the shortcomings of existing models, specifically large trainable weight parameters and class imbalance.
- To improve the efficiency and deployability of breast cancer classification models on resource-constrained devices.
Main Methods:
- Proposed EpiBrCan-Lite, a novel lightweight model comprising Data Encoding, TransGRU, and Classification blocks.
- The TransGRU block modifies the traditional Transformer Encoder by replacing the MLP module with a GRU module to reduce trainable weight parameters and capture long-range dependencies.
- Utilized Synthetic Minority Oversampling Technique (SMOTE) to mitigate class imbalance in the TCGA breast cancer dataset.
Main Results:
- EpiBrCan-Lite achieved high performance metrics: 95.85% accuracy, 95.96% recall, 95.85% precision, and 95.90% F1-score.
- The model demonstrated significantly reduced trainable weight parameters, using only 1/1500 compared to state-of-the-art models.
- Low False Positive Rate (FPR) of 1.03% and False Negative Rate (FNR) of 4.12% were recorded.
Conclusions:
- The EpiBrCan-Lite model provides efficient and accurate classification of breast cancer subtypes.
- Its lightweight architecture makes it suitable for deployment on devices with limited computational power.
- This study offers a viable solution for improving breast cancer prognosis through advanced, resource-efficient machine learning.
Keywords:
Breast cancer diseaseDNA methylation dataEpigenomicGated recurrent unitSMOTETransformer encoderMore Related Videos
Related Concept Videos
01:22Classification of Epithelial Tissues: Overview
Epithelial tissues are classified according to the shape of the cells and the number of cell layers formed. Cell shapes can be squamous (flattened and thin), cuboidal (square-like, as wide as it is tall), or columnar (rectangular, taller than it is wide). Additionally, the nucleus shape helps identify the type of epithelial cells. Squamous cells have flattened disc-shaped nuclei, cuboidal cells have spherical nuclei, and columnar cells have elongated nuclei.
Based on the number of cell layers,...
Based on the number of cell layers,...
01:29Classification of Epithelial Tissues: Stratified Epithelium
Stratified epithelium consists of several stacked layers of cells. They provide the durability to withstand constant physical and chemical attacks. Stratified epithelium is named after the shape of the most apical layer of cells. Stratified squamous epithelium is the most common type found in the human body. In this tissue, the apical cells are squamous, whereas the basal layer contains either columnar or cuboidal cells. The basal cells divide to form new daughter cells, which gradually become...