RNA sequencing and analyses
The RNAprep Pure Plant Plus Kit was used to extract total RNA from the roots, stems, and leaves of rosemary samples according to the manufacturer’s instructions. The RNA-seq library was constructed on the Illumina HiSeq X Ten platform and sequenced to 150-bp paired reads, and then de novo assembled. The quality control of the assembled sequence included low-quality filtering, removal of N-containing bases, and removal of 3’ and 5’ end sequencing adapters. Raw data were filtered using SOAPnuke software (www.bgitechsolutions.com) with the following filtering parameters: -n 0.01 -l 20 -q 0.4 -A0.25 –cutAdaptor -Q 2 -G –polyX50 –minLen 150. After the raw data had been filtered, FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to monitor data quality. Three sets of biological replicates were made for each of the three parts of the roots, stems and leaves, and 9 sample RNA sequences were obtained for subsequent transcriptome analyses.
The ‘WGCNA’ package of R (v4.0) was used to separate all genes into modules of color blocks, calculate the PCCs between gene expression and the content of differential metabolites, and construct a weighted gene co-expression network based on PCCs. In addition, genes were regarded as highly corelated to candidate metabolites if PCCs were > 0.6 and p-values < 0.05.