*4.4. RNA-Seq Data Processing and Gene Expression Calculation*

Raw reads were first filtered by Fastx program to remove disqualified, short or ribosome RNA reads, resulting in clean reads. Clean reads % was calculated as (clean reads/raw reads) %; mRNA % was as ((clean reads-rRNA reads)/clean reads) %; Subsequent genome mapping was by spliced mapping algorithm in Tophat program [45] to a reference genome, *Oryza sativa japonica* NPB version 7.0 from ftp://ftp.plantbiology.msu.edu, and generated BAM files.

To estimate the abundance of gene expression, reads number of uniquely mapped genes were normalized to Fragments Per Kilobase of exon model per Million mapped reads (FPKM) by Cufflinks program [46]. FPKM was calculated as (total exon fragments in reads)/((mapped reads in millions) × exon length in Kilo base pair)), where total exon fragments was the number of reads mapped to exons, mapped reads was the number of reads mapped to the reference genome, and exon length was total base pair number of exons in Kilo base pair.

To generate a list of differentially expressed genes (DEG) between treatments, FPKM was used to calculate the fold change and false discovery rate (FDR, an adjusted *p*-value, i.e., *q*-value) by Cuffdiff program [18]. The threshold of DEG was set as *q*-value≤ 0.05 and fold change ≥ 2.
