Joint identification of differentially expressed gene and phenotype associated genes

Abstract

The emergence of a wide variety of new techniques has led to the production of diverse types of biological data. Among them microarray technology has brought innovative changes in biological field and is still most commonly used in various research fields. For the last decade, many analytical methods and tools have been developed. In general, the detection of differentially expressed genes (DEGs) among different treatment groups is often a primary purpose of microarray data analysis. In addition, the association studies investigating the relationship between genes and the phenotype of interest such as survival time became also popular in microarray data analysis. Such association analysis provides the list of phenotype associated genes (PAGs). In this study, I consider a joint identification of DEGs and PAGs in microarray data analyses. The first approach is a naïve approach which detects DEGs and PAGs separately, and then identifies the intersection genes of PAGs and DEGs. The second approach is a hierarchical approach which detects DEGs first and then chooses PAGs among DEGs, or visa versa. I propose a new model-based approach for a joint identification of DEGs and PAGs simultaneously. Through a real microarray data analysis, I show that our model-based approach provides a more powerful result than the naïve approach and the hierarchical approach.