Mitogenome Reconstruction from Simulated Data

Paired-end illumina data sets (read length 150 bp, insert size 300 ± 50 bp, error rate 0) were simulated for five salmonid mitochondrial genomes, i.e. the four genomes used in case study 2 and the mitochondrial genome of Salmo trutta retrieved from GenBank (NC_010007), to a coverage of ∼50× (i.e. 6000 reads per reference) using wgsim from SAMTOOLS (36 (link)). A single pooled data set containing all generated reads was created from the simulated data. Five individual MITObim instances were subsequently run (in proofreading mode), each using a ∼1200 bp COI barcode as initial seed for the reconstruction of individual mitochondrial genomes from the pooled data set. Seeds were obtained from the mitochondrial genomes of the targeted species, respectively, as the proofreading procedure requires 100% sequence identity for successful incorporation of reads into the growing readpool (see Supplementary Methods).