*2.2. RNA Extraction and Sequencing*

Total RNA was extracted using the Qiagen RNeasy Plant Mini Kit (Qiagen, Limburg, The Netherlands). RNA concentration was quantified using Nanodrop spectrophotometer (ND-1000; Nanodrop Technologies, Wilmington, DE, USA), while purity and integrity were established using an Agilent 2100 BioAnalyzer RNA 6000 Kit (Agilent Technologies, Santa Clara, CA, USA), with a RIN value of 8 used as quality threshold. Illumina library preparation and sequencing were completed following the standard protocols of Macrogen Inc. (Seoul, South Korea). Using Illumina HiSeq 2000 and HiSeq 4000 (Illumina, Inc., San Diego, CA, USA) platforms for the whole panicle and flag-leaf tissues, respectively, 101 bp aired-end sequencing was done. A quality check of raw RNA-Seq reads was performed using FastQC software (version 0.11.5) [51]. The sequencing data have been deposited in NCBI's Gene Expression Omnibus (GEO) database under the accession number GSE145870.

### *2.3. Transcriptome Assembly and Expression Level Quantification*

Raw fastq reads were filtered using Trimmomatic software, version 0.36 [52] using default settings. An indexed transcriptome fasta file was built from the rice japonica genome (IRGSP1.0) and annotation gff3 file from Rice Genome Annotation release 7 (RGAP 7, http://rice.plantbiology.msu.edu/), using the "gffread" function of Cufflinks (version 2.2.1) [53], and Salmon (version 0.7.2) [54] "index" function. Salmon then quantified transcript abundance in quasi-mapping mode directly using the indexed transcriptome and the trimmed high-quality paired-end reads with parameters "-l A, -seqBias and -gcBias" to allow automatic inference of library type, learn and to correct for sequence-specific and fragment-level GC content biases, respectively. Gene expression levels were normalized using the transcripts per million mapped reads (TPM) method, exported as estimated transcript abundance, and aggregated to gene-level expression using the Bioconductor package tximport (version 1.2.0) [55] complemented with the reader package (version 1.1.1), within the R environment (version 3.3.3). Principal component analysis (PCA) was performed in R using ggplot2 (version 2.2.1) and gplots package (version 3.0.1) to determine relationships between samples.
