GENECLUST

GENECLUST implements the gene shaving methodology to identify coherent gene clusters from gene expression microarray data for discovery of co-regulated gene sets and discrimination of tissue and disease subtypes.

Key Features:

• Gene Shaving Methodology: Implements gene shaving to identify co-regulated gene families and detect subtle expression patterns across tissue samples.
• Speed and Flexibility: Supports rapid processing, variable cluster sizes, and both partial and full supervision during clustering.
• High Coherence in Gene Expression Profiles: Produces clusters with high internal coherence that exhibit significant variation across tissue samples.
• Versatility Across Data Sets: Applied to Affymetrix oligonucleotide arrays from colon (normal and tumor) and leukemia tissues, distinguishing AML versus ALL and identifying ALL-Tcell specific genes.

Scientific Applications:

• Identification of Co-regulated Gene Families: Clusters genes into co-regulated families such as ribosomal proteins and smooth muscle genes in colon datasets.
• Tissue Type Differentiation: Separates tissue types (e.g., tumor vs. normal) by detecting distributed gene expression patterns.
• Leukemia Subtype Classification: Categorizes leukemia samples according to external classifications, aiding subtype discrimination such as AML versus ALL and identification of ALL-Tcell subgroup.

Methodology:

Initial gene filtering, gene shaving-based cluster formation, and subsequent tissue clustering based on those gene clusters, with ranking and classification of gene clusters.