*2.6. Phylogenetic Analyses*

In order to accurately infer the phylogenetic relationships of *P. vulgaris* within the Fabaceae family, maximum likelihood (ML) analysis was performed based on the conserved mitochondrial PCGs (amino acid sequences) of 23 higher plants. The NCBI accession numbers and abbreviations of all these observed mitogenomes are listed in Table S2. Apart from the 19 representative Fabaceae species, taxon sampling also included two species of Solanales (*C. annuum* and *N. tabacum*) and two species of Malpighiales (*P. tremula* and *S. suchowensis*) as outgroups. The single-copy orthologous PCGs common among the 23 analysed species were selected with local Perl scripts. All conserved mitochondrial PCGs were extracted from each mitogenome. The conserved gene sequences from the mitogenome were concatenated into a single dataset and aligned using Muscle software with default settings [50]. Poorly-aligned sequences were deleted or manually adjusted for each alignment. Prior to constructing the phylogenetic tree, we applied MEGA X to determine the most appropriate amino acid substitution model [41]. Based on the model selection results, the ML tree based on a JTT + F model with a gamma distribution was constructed using MEGA X. The bootstrap index value (%) in which the associated taxa clustered together was shown next to the branches and was calculated from 1000 replications.
