Computational analysis of circadian-rhythm dysregulated genes exhibiting undifferential expression in Oryza sativa

Abstract

ABSTRACT Differential expression(DE) analysis identifies genes with differential expressions in a phenotype, but it cannot detect coordinated regulations of a gene from its transcription factors(TFs). This study selected rice diurnal rhythmic samples to train a transcriptional regulatory relationship model, which was used to predict the expression levels of target genes based on those of TFs in the other samples. The difference between the predicted and real expression levels of a target gene in a sample was defined as this gene’s model-based quantitative transcription regulation alternation (mqTrans) value. This study paid special attention to those genes with significant changes in their mqTrans values but without differential expression patterns, which were called dark biomarkers of rice circadian rhythm. The investigation revealed three dark biomarkers that were otherwise ignored by the conventional biomarker studies due to the undifferential expressions. One detected dark biomarker Os11g0135400 was crucial for rice circadian changes, but they were not on the list of published genes that were differentially expressed in circadian rhythm changes. This study captured important circadian rhythmic biomarkers ignored by traditional biological and computational methods. The proposed mqTrans analysis and the detected dark biomarkers contributed complementary information to the understanding of the rice circadian rhythmic mechanism.