RNA-Seq analysis for breast cancer detection: a study on paired tissue samples using hybrid optimization and deep...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Breast cancer poses a significant global health challenge, with high mortality rates.
• Early detection is hindered by the complexity and high dimensionality of gene expression data.
• Accurate classification of breast cancer is crucial for effective treatment and improved patient outcomes.

Purpose of the Study:

• To develop an advanced deep learning model for accurate breast cancer detection using RNA-Seq gene expression data.
• To address the challenges of high dimensionality and complexity inherent in gene expression datasets.
• To enhance feature selection and classification accuracy through a novel hybrid optimization approach.

Main Methods:

• A hybrid gene selection method combining Harris Hawk Optimization (HHO) and Whale Optimization (WO) with deep learning was developed.
• The model's performance was evaluated against conventional optimization algorithms: Genetic Algorithm (GA), Artificial Bee Colony (ABC), Cuckoo Search (CS), and Particle Swarm Optimization (PSO).
• RNA-Seq data from 66 paired normal and cancerous breast tissue samples were utilized.

Main Results:

• The proposed hybrid HHO-WO deep learning model achieved a mean classification accuracy of 99.0%.
• The model consistently outperformed GA, ABC, CS, and PSO methods in breast cancer detection.
• The dataset included samples from 55 female breast cancer patients across different stages and age-matched healthy controls.

Conclusions:

• The hybrid gene selection approach using HHO and WO with deep learning is a highly accurate tool for breast cancer detection.
• This advanced method shows significant promise for improving early detection rates.
• The findings could pave the way for personalized treatment strategies and better patient outcomes.