Abstract
We conducted genome-wide identification of R-loops followed by integrative analyses of R-loops with relation to gene expression and epigenetic signatures in the rice genome. We found that the correlation between gene expression levels and profiled R-loop peak levels was dependent on the positions of R-loops within gene structures (hereafter named "genic position"). Both antisense only (ASO)-R-loops and sense/antisense (S/AS)-R-loops sharply peaked around transcription start sites (TSSs), and these peak levels corresponded positively with transcript levels of overlapping genes. In contrast, sense only (SO)-R-loops were generally spread over the coding regions, and their peak levels corresponded inversely to transcript levels of overlapping genes. In addition, integrative analyses of R-loop data with existing RNA-seq, chromatin immunoprecipitation sequencing (ChIP-seq), DNase I hypersensitive sites sequencing (DNase-seq), and whole-genome bisulfite sequencing (WGBS or BS-seq) data revealed interrelationships and intricate connections among R-loops, gene expression, and epigenetic signatures. Experimental validation provided evidence that the demethylation of both DNA and histone marks can influence R-loop peak levels on a genome-wide scale. This is the first study in plants that reveals novel functional aspects of R-loops, their interrelations with epigenetic methylation, and roles in transcriptional regulation.