Bayesian ranking and selection methods using hierarchical mixture models in microarray studies

Area of Science:

• Genomics
• Bioinformatics
• Statistical Genetics

Background:

• Microarray studies aim to identify differentially expressed genes for further investigation.
• Gene prioritization is crucial due to limited resources in the screening stage.
• Conventional methods may lack efficiency in ranking genes based on differential expression.

Purpose of the Study:

• To develop and evaluate three empirical Bayes methods for gene ranking in microarray studies.
• To improve the accuracy of identifying differentially expressed genes.
• To provide a robust statistical framework for gene prioritization.

Main Methods:

• Development of three empirical Bayes methods using hierarchical mixture models.
• Methods based on minimizing mean squared errors (parameters and ranks) and maximizing sensitivity.
• Incorporation of mixture structures for differential and nondifferential gene components.

Main Results:

• Empirical Bayes methods demonstrated improved accuracy in gene ranking compared to conventional approaches.
• Simulation studies validated the performance of the proposed methods.
• Application to breast cancer data showcased practical utility.

Conclusions:

• The proposed empirical Bayes methods offer enhanced accuracy for gene ranking in differential expression analysis.
• These methods facilitate more effective identification of candidate genes in genomic studies.
• The approach is applicable to clinical studies, including cancer research.