Cancer gene search with data-mining and genetic algorithms

Area of Science:

• Oncology
• Bioinformatics
• Computational Biology

Background:

• Cancer is a leading cause of death, emphasizing the need for early and accurate detection.
• Gene expression data analysis is crucial for cancer identification, classification, treatment selection, and drug development.

Purpose of the Study:

• To propose an integrated gene-search algorithm for analyzing gene expression data to improve cancer detection and classification.
• To identify significant genes associated with cancer types and enhance prediction accuracy.

Main Methods:

• Utilized gene expression datasets for ovarian, prostate, and lung cancer.
• Developed an integrated algorithm combining a genetic algorithm and correlation-based heuristics for data preprocessing and mining.
• Employed decision tree and support vector machine algorithms for predictions, with bagging and stacking for accuracy enhancement.

Main Results:

• The proposed algorithm achieved high classification accuracy in identifying significant genes.
• Enhanced classification accuracy was observed after applying bagging and stacking techniques.
• Results demonstrated the algorithm's effectiveness in cancer detection and classification.

Conclusions:

• The integrated gene-search algorithm provides a robust method for cancer detection and classification using gene expression data.
• Mapping genotype to phenotype parameters can simplify and reduce the cost of cancer diagnostics.
• This approach holds potential for advancing personalized cancer treatment and drug discovery.