Ancient WGD event prediction
To identify the ancient replication events of S. rosmarinus , WGD analysis was performed with the predicted proteins file from its genome. We used wgd (v1.0.1) (https://github.com/arzwa/wgd) (Zwaenepoel & Van de Peer, 2019) and MCScanX (Wang et al., 2012) to detect genome-wide replication events in the species by calculating the synonymous substitution rate (Ks ). On the basis of the most representative transcripts identified using the script of CAFÉ 5 (‘cafetutorial_longest_iso.py’) (https://github.com/hahnlab/CAFE) (Mendes et al., 2020), sequence alignments and analysis of all gene clustering were performed by wgd. We then calculated the Ks distribution and plotted this using R (v4.0.1) (https://www.r-project.org/). And we calculated and compared the Ks distribution on the internal collinearity of gene pairs of S. rosmarinus and the Ks distribution based on the orthologous gene pairs between S. rosmarinus and S. miltiorrhiza andS. splendens . According to the normal distribution peaks in the distribution, putative whole-genome duplication events could be identified within species.