Detecting pathways transcriptionally correlated with clinical parameters

Area of Science:

• Computational biology
• Bioinformatics
• Genomics

Background:

• Microarray data analysis requires advanced computational methods.
• Protein interaction networks offer insights into gene function and biological pathways.
• Identifying functional modules in gene expression data is crucial for understanding disease.

Purpose of the Study:

• To develop a novel computational methodology for extracting connected network modules from microarray data.
• To correlate these modules with specific clinical parameters (numerical and logical).
• To apply the method to breast cancer data for identifying biologically relevant modules.

Main Methods:

• Utilized human protein interaction networks.
• Developed a method to extract connected network modules with coherent gene expression patterns.
• Applied the methodology to a large breast cancer dataset, correlating modules with nine clinical parameters.

Main Results:

• Identified biologically relevant modules associated with patient age, tumor size, and metastasis-free survival in breast cancer.
• The method successfully detected disease-relevant pathways missed by other approaches.
• Results support existing hypotheses and suggest novel molecular pathways in breast tumor diversity.

Conclusions:

• The novel methodology effectively integrates protein interaction networks and gene expression data for clinical parameter correlation.
• This approach enhances the discovery of disease-relevant biological pathways, particularly in complex diseases like breast cancer.
• The findings contribute to a deeper understanding of breast tumor heterogeneity and molecular mechanisms.