A Boruta-SMOTE Integrated Approach for Rapid Donkey Breed Classification Using SNP Data: Addressing...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Genomic data classification faces challenges like the curse of dimensionality and class imbalance.
• Traditional methods struggle with high-dimensional single nucleotide polymorphism (SNP) data.
• Class imbalance hinders accurate analysis in genetic studies.

Purpose of the Study:

• To develop an effective genomic data classification method addressing dimensionality and imbalance.
• To integrate Boruta for feature optimization and SMOTE for class balancing.
• To improve classification performance for genomic datasets, particularly in animal genetics.

Main Methods:

• Feature selection using the Boruta algorithm to identify key genetic markers.
• Application of Synthetic Minority Over-sampling Technique (SMOTE) to balance class distributions.
• Comparative analysis against traditional classifiers like Random Forest, KNN, XGBoost, and CNN.

Main Results:

• The Boruta-SMOTE integrated method significantly reduced data complexity.
• SMOTE effectively balanced minority class distributions, mitigating imbalance issues.
• The proposed method demonstrated superior performance across accuracy, precision, recall, and F1-score compared to traditional methods.

Conclusions:

• The Boruta-SMOTE approach offers a robust solution for high-dimensional, imbalanced genomic data.
• This method enhances the classification of genetic resources, aiding breed improvement and commercial applications.
• Provides a valuable tool for advancing genomic data analysis in various biological fields.