RecBic

RecBic: Algorithm for Trend-Preserving Bicluster Identification in Gene Expression Data

RecBic identifies trend-preserving biclusters in gene expression matrices, with emphasis on narrow biclusters where the number of genes exceeds the number of conditions or samples.

Key Features:

• Trend-Preserving Bicluster Identification: Detects biclusters that preserve expression trends across subsets of genes and conditions, particularly narrow biclusters.
• High Accuracy and Efficiency: Accurately identifies implanted biclusters in simulated datasets and distinguishes bicluster elements from background matrix noise with similar distributions.
• Robustness to Noise and Overlaps: Maintains performance in the presence of noise and overlapping biclusters.
• Functional Gene Identification: Identifies functionally related genes in real gene expression datasets.

Scientific Applications:

• Gene Expression Analysis: Detects functional gene modules in genomics, disease research, functional genomics, and systems biology studies.

Methodology:

RecBic operates on a gene expression matrix, initializes bicluster detection with a column seed, and expands the seed into a full-sized bicluster through iterative real-number comparisons to ensure precise and efficient cluster growth.