*4.4. Assembly and Sca*ff*olding of the Illumina MiSeq Reads from DNA Sequencing*

Reads were trimmed with Trimmomatic version 0.36 [84] and assembled using the CLC Genomics Workbench (CLC-GWB) Version 10.1.1 (CLC-bio, a Qiagen company; Aarhus, Denmark), (length fraction = 0.9, similarity fraction = 0.95, map reads back to contigs, word size = 45). Two out of 279,689 contigs were identified as mitochondrial sequences based on high mapping coverage and comparison against the NCBI nucleotide collection database. Blastn was used to join the mitochondrial contigs by finding similarities between the sequence endings in order to create the complete DNA sequence of the mitochondrial genome. To verify the assembled sequence, we used the tool "ROUSFinde1\_1", described by Wynn and Christensen [19], with default parameters to identify repeat structures in the mitochondrial sequence of *F. sylvatica* (MT446430). Accordingly, we identified repeat structures in *Quercus variabilis* (MN199236) and *Betula pendula* (LT855379.1) for the comparison of repeats between these three mitochondrial genomes (Table S1).
