Mtgenome annotation and characteristics analysis
The rough annotation of protein-coding genes (PCGs), transfer RNA genes (tRNAs), ribosomal RNA genes (rRNAs), and CR was initially identified using MITOS (http://mitos.bioinf.uni-leipzig.de/index.py) (Berntet al. , 2013), and then determined in comparison of published homologous mtgenome sequences in phylogeny-close species using MEGAX (Kumar et al. , 2018). The tRNAs secondary structures were predicted using tRNAscan-SE Search Server v. 1.21 (http://lowelab.ucsc.edu/tRNAscan-SE/) (Lowe & Eddy, 1997). The annotation of the mtgenomes was corrected manually using the Geneious v. 4.8.5 (Kearse et al. , 2012), and final mtgenomes were submitted to the GenBank database. The secondary structures of the tRNAs were visualized and manually edited using VARNA (http://varna.lri.fr) (Dartyet al. , 2009). The mtgenomes were visualized using the Chloroplot online server with default parameters (Zheng et al. , 2020). Base composition and relative synonymous codon usage (RSCU) of 90 species of mtgenome were computed with PhyloSuite desktop platform (Zhang et al. , 2020). AT-skew [(A - T) / (A + T)] and GC-skew [(G - C) / (G + C)] were estimated to investigate nucleotide composition bias (Perna & Kocher, 1995), and three-dimensional scatterplots of AT-Skew, GC-Skew and AT% were drawn using Origin Pro v. 9.0 (Mikrajuddin and Khairurrijal, 2009). Selection pressure of the 13 PCGs was analyzed by calculating Ka (non-synonymous mutation rates) and Ks (synonymous mutation rates) values with DnaSP v. 5.0 (Librado & Rozas, 2009), and visualized using RStudio. Sequences saturation was assessed in DAMBE v. 5.0 (Xia, 2013).