2.3　Sequence Assembly and Annotation

Approximately 4，990Mb of high-quality clean reads were produced by the Illumina platform after filtering out low-quality and adapter-contaminated reads.The sequencing reads of the mt genome were extracted from the raw data by mapping to the mt genomes of other Lepidoptera using bwa(Li and Durbin，2009).Approximately 15.26Mb of reads were extracted and then assembled using SOAPdenovo v2.04(Li et al.，2010).Using the PCR product sequencing data，this mt genome was further assembled as a linear contig using SeqMan Pro software(DNASTAR Inc.，Madison，WI，USA).Because insect mt genomes are circular，we filled in the gap by using PCR amplification with primers designed at the 3＇and 5＇termini of this linear contig.Finally，we adjusted the linear representation of the circular DNA by circular permutation so that the sequence started with trnI(GAT)，which is always reported as the beginning of ancestral lepidopteran mt genomes.(https://www.daowen.com)

The mt genome sequences were annotated using MITOS WebServer(http：//mitos.bioinf.uni-leipzig.de)(Bernt et al.，2013).The nucleotide sequences of proteincoding genes(PCGs)were translated to protein sequences using the invertebrate mitochondrial code(transl_table=5).The MITOS predictions were manually curated using other published lepidopteran mt genomes as references.The gene beginnings and ends were modified，if necessary，to be consistent with other species.The rrn L gene was annotated to extend to the boundaries of the flanking trnL and trnV.The 5'end of rrnS was determined via comparison with orthologous genes of other Lepidoptera，while the 3'end was annotated to be adjacent to the start of trnV.The secondary structures of tRNA genes were also predicted using MITOS(Bernt et al.，2013).For the base composition of the nucleotide sequences，the composition skewness was calculated by Lobry as follows：AT skew=［A—T］/［A+T］，GCskew=［G—C］/［G+C］(Lobry，1996).