Bayesian Fourier clustering of gene expression data

ABSTRACT

Clustering gene expression data are an important step in providing information to biologists. A Bayesian clustering procedure using Fourier series with a Dirichlet process prior for clusters was developed. As an optimal computational tool for this Bayesian approach, Gibbs sampling of a normal mixture with a Dirichlet process was implemented to calculate the posterior probabilities when the number of clusters was unknown. Monte Carlo study results showed that the model was useful for suitable clustering. The proposed method was applied to the budding yeast Saccaromyces cerevisiae and provided biologically interpretable results.

KEYWORDS:

• Dirichlet process
• Fourier coefficient
• Model-based cluster
• Normal mixture
• Yeast cell cycle

MATHEMATICS SUBJECT CLASSIFICATION:

• Primary 62H30
• Secondary 62G08

Acknowledgments

The authors would like to thank the referees and the associate editor for providing informative feedback which have led to an improved version of a previous draft. The work of the first author was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology grant 2014-044670, and the work of the corresponding author was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science, ICT & Future Planning 2012R1A1A1008444.