Use of keyword hierarchies to interpret gene expression patterns

Area of Science:

• Bioinformatics
• Computational Biology
• Gene Expression Analysis

Background:

• High-density microarray technology enables simultaneous monitoring of thousands of genes.
• Interpreting large-scale gene expression data presents a significant challenge.
• Existing statistical methods identify gene clusters but lack biological context.

Purpose of the Study:

• To develop a data mining method for interpreting gene expression clustering results.
• To leverage published literature for understanding biological similarities among genes.
• To enhance the biological relevance of microarray data analysis.

Main Methods:

• Utilizes indexing terms (keywords) from published literature linked to specific genes.
• Employs a data mining approach to assess conceptual similarity.
• Leverages the hierarchical structure of Medical Subject Headings (MeSH) from MEDLINE.
• Incorporates enzyme registry numbers for enhanced data linkage.

Main Results:

• Presents a novel view of conceptual gene similarity within clusters.
• Facilitates the interpretation of gene expression patterns by connecting them to existing biological knowledge.
• Demonstrates a method to bridge statistical clustering with biological meaning.

Conclusions:

• The described data mining method aids in the biological interpretation of gene expression data.
• Linking gene expression clusters to literature-based keywords provides valuable biological insights.
• This approach enhances the utility of microarray data analysis by integrating textual information.