2.3　Assembly of the mitochondrial genome

Clean data were obtained by filtering out the sequencing adapters and low-quality sequences(parameters：minimum sub-read length=500bp；minimum polymerase read quality=0.80).The mt sequences were extracted from the filtered reads containing both nuclear and mt genomes，using BLASR which matches each read against 201 published fungal mitochondrial genomes(Chaisson，2012).About 2815 subreads in 25.4Mb mt sequencing data were obtained，with an average read length of 8993bp and a longest read length of 39358bp，reaching an average depth of 167 X.The mt genome was assembled through Hierarchical Genome Assembly Process(HGAP)workflow，including preassembly，error correction，Celera assembly and polishing with Quiver(Chin et al.，2013).Long reads were selected to be“seed”reads，which the other subreads were blast against to improve accuracy.The corrected reads were retained and fully assembled by using overlaplayout-consensus(OLC)algorithm in the Celera Assembler program(Myers et al.，2000)，and then further refined with Quiver(Chin et al.，2013).(https://www.daowen.com)