2.2. Identification of the GH3 Genes

Potential members of the GH3 gene family were identified based on the Hidden Markov model (HMM) and BLAST homology searches [27,28]. Protein and nucleotide sequences of wild rice (Oryza rufipogon, version: OR_W1943.39; Oryza nivara, version: v1.0; Oryza punctata, version: v1.2; Oryza glumaepatula, version: v1.5) and cultivated rice (O. sativa ssp. indica, version: ASM465v1, version: R498) were downloaded from EnsemblPlants (http://plants.ensembl.org/index.html) and MBKBASE (http://www.mbkbase.org/R498/). GH3 protein sequences of Arabidopsis and rice (O. sativa ssp. japonica) were downloaded from the TAIR Database (https://www.arabidopsis.org/) and RiceData (http://www.ricedata.cn/gene/) as query sequences [11,12,29]. These query sequences were used to search for GH3 protein sequences in five Oryza species’ protein databases using local ncbi-blast-2.7.1+ (ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST) in the Blastp method with a cut-off E-value of e⁻⁵. Then, PF03321 was downloaded from Pfam (http://pfam.xfam.org/) and PF03321 was used to query the Oryza species’ proteins database using HMMER 3.0 software (http://hmmer.org/) [27]. Since the GH3 domain is longer than 400, protein sequences with a length of less than 400 were deleted in this study. Finally, the GH3 domains of all the nonredundant protein sequences were verified by SMART (http://smart.embl-heidelberg.de/) and Pfam (http://pfam.xfam.org/search/sequence) [8,27].