Relative expression analysis for identifying perturbed pathways.

Area of Science:

• Computational biology
• Genomics
• Personalized medicine

Background:

• Accurate cancer classification is crucial for effective treatment, especially for histopathologically similar tumors like GIST and LMS.
• Existing methods may struggle to differentiate between cancers with overlapping features.

Purpose of the Study:

• To introduce and validate Differential Rank Conservation (DIRAC) as a novel computational approach for molecular classification.
• To identify disease-perturbed pathways and gene expression patterns for personalized medicine.

Main Methods:

• Utilized global datasets to identify stable and predictive patterns of gene and protein relative expression reversals.
• Applied DIRAC to analyze gene ordering within pathways based on relative expression ranks.
• Examined gene expression in GIST and LMS tumor profiles to identify conserved and variable pathway rankings.

Main Results:

• DIRAC quantifies differences in pathway rankings within and between phenotypes.
• Identified tightly regulated pathways with conserved gene ordering in GIST and LMS.
• Discovered variably expressed pathways that serve as molecular classification signatures, accurately classifying microarray samples.

Conclusions:

• DIRAC is a powerful tool for molecular classification and identifying disease-specific pathways.
• This approach enhances personalized medicine by enabling accurate differentiation of similar cancers.
• The study validates DIRAC's effectiveness through high accuracy in classifying GIST and LMS samples.