Statistical assessment of functional categories of genes deregulated in pathological conditions by using microarray...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Identifying deregulated cellular processes from gene expression profiles is a key challenge in biomedical research.
• Current methods compare gene sets to differentially expressed gene lists using univariate statistics.

Purpose of the Study:

• To propose a statistically sound method for evaluating the relevance and significance of predefined gene lists from raw microarray data.
• To assess if a functional gene category can predict disease occurrence based on its members' expression levels.

Main Methods:

• Utilizing prediction accuracy as a measure of gene list relevance.
• Employing a multiple random validation strategy for accuracy estimation.
• Assessing statistical significance using two independent permutation tests against null hypotheses.

Main Results:

• The proposed method effectively measures the relevance and statistical significance of gene lists.
• Demonstrated utility in analyzing Gene Ontology biological process terms in colon and prostate cancer using multiple microarray datasets.
• Outperformed current approaches in identifying relevant biological processes.

Conclusions:

• The developed methodology provides a robust framework for analyzing gene expression data to uncover disease-associated cellular processes.
• This approach enhances the understanding of pathology by leveraging functional gene categories and prediction accuracy.