**3. Materials and Methods**

#### *3.1. Species and Database*

In total, 128 *Bacillus* species genomes available for public use at KEGG (https://www.genome. jp/kegg-bin/show\_organism?category=Bacillus) were used in this study (Table S4). *Bacillus* species used in this study, along with their names, species codes, and individual genome database links, were presented in Table S4.

#### *3.2. Genome Data Mining and Annotation of P450s*

P450 mining in *Bacillus* species was carried out following the methods described elsewhere [25]. Briefly, the whole proteome of *Bacillus* species was downloaded from the databases listed in Table S4, and subjected to the NCBI Batch Web CD-Search Tool (http://www.ncbi.nlm.nih.gov/Structure/ bwrpsb/bwrpsb.cgi). Proteins that belong to a P450 superfamily were selected and based on the International P450 Nomenclature Committee rule; proteins with >40% identity and >55% identity were grouped under the same family and subfamily, respectively [27–29]. Proteins with less than 40% identity were assigned to a new P450 family.

### *3.3. Phylogenetic Analysis of P450s*

The phylogenetic tree of *Bacillus* species P450s was built as described elsewhere [25], with slight modifications. Briefly, the *Bacillus* P450s protein sequences along with the outgroup *M. tuberculosis* CYP51B1 (Rv0764c) protein were aligned by MAFFT v6.864 [49], embedded on the Trex web server [50]. Then, the alignments were automatically subjected to tree inferring and optimization by the Trex web server. Finally, the best-inferred trees were visualized, colored, and generated by iTOL (http://itol.embl.de/about.cgi) [51].
