ILRC: a hybrid biomarker discovery algorithm based on improved L1 regularization and clustering in microarray data.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-dimensional gene chip data presents challenges for disease diagnosis and drug development.
• The curse of dimensionality hinders the identification of significant genes or proteins.
• Machine learning is crucial for feature selection and accurate classification models.

Purpose of the Study:

• To propose an effective feature selection algorithm for high-dimensional gene expression data.
• To address the challenge of the curse of dimensionality in biological data analysis.
• To develop a method for identifying potential biomarkers for disease diagnosis and drug development.

Main Methods:

• A novel feature selection algorithm, ILRC (improved L1 regularization with clustering), was developed.
• The method involves clustering features and removing redundant ones within sub-clusters.
• Remaining features are iteratively evaluated using ILR, with results determined by cumulative weight reordering.

Main Results:

• The ILRC method demonstrated superior performance compared to existing advanced hybrid and traditional feature selection techniques on public microarray datasets.
• Selected biomarkers using the ILRC method showed agreement with clinical data from a cooperative hospital for cleft lip and palate.
• The algorithm achieved high stability and classification accuracy.

Conclusions:

• The proposed ILRC algorithm effectively removes redundant features from high-dimensional data.
• The method offers high stability and classification accuracy, enabling the selection of potential biomarkers.
• This approach holds promise for advancing disease diagnosis and drug development through precise gene identification.