Subspace Weighting Co-Clustering of Gene Expression Data

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Microarray technology generates extensive gene expression data.
• Clustering algorithms are vital for exploring this data.
• Co-clustering is beneficial as genes may correlate with sample subsets.

Purpose of the Study:

• To propose a novel co-clustering algorithm, Subspace Weighting Co-Clustering (SWCC), for high-dimensional gene expression data.
• To introduce a gene subspace weight matrix to identify gene contributions to sample clusters.
• To develop an iterative algorithm for solving the co-clustering objective function.

Main Methods:

• Developed the Subspace Weighting Co-Clustering (SWCC) algorithm.
• Introduced a gene subspace weight matrix into the co-clustering objective function.
• Designed an iterative algorithm to compute the subspace weight matrix during co-clustering.

Main Results:

• SWCC demonstrated encouraging performance compared to six state-of-the-art clustering algorithms on ten gene expression datasets.
• The algorithm was successfully applied to gene clustering and selection.
• Selected genes using SWCC improved the classification performance of Random Forests.

Conclusions:

• SWCC is an effective novel algorithm for co-clustering high-dimensional gene expression data.
• The gene subspace weight matrix is crucial for identifying gene contributions.
• SWCC facilitates improved gene selection for enhanced biological data analysis and machine learning applications.