Row and Column Structure-Based Biclustering for Gene Expression Data.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-throughput gene analysis technologies necessitate advanced biclustering methods.
• Existing biclustering algorithms often suffer from high time and space complexity.
• Checkerboard patterns in microarray data require efficient detection methods.

Purpose of the Study:

• To propose a novel biclustering method, Row and Column Structure-based Biclustering (RCSBC), with low time and space complexity.
• To effectively identify checkerboard patterns in gene expression data.
• To address the limitations of existing biclustering techniques for large-scale datasets.

Main Methods:

• Describing bicluster structure using row and column relationships.
• Selecting representative rows and columns via two specific algorithms.
• Performing biclustering on the space defined by representative rows and columns.

Main Results:

• RCSBC demonstrates superior performance compared to state-of-the-art algorithms in terms of clustering accuracy.
• The proposed method significantly reduces time and space complexity.
• Effectiveness validated on both synthetic and real-world gene expression datasets.

Conclusions:

• RCSBC offers a novel approach by exploiting gene expression matrix row/column structure for bicluster identification.
• The method enables efficient biclustering of large-scale gene expression data, even without loading the entire dataset into memory.
• RCSBC provides a valuable tool for genomic data analysis, improving upon existing methods.