*4.7. Annotation of the DNA Sequence of the Mitochondrial Genome of F. sylvatica*

Structural and functional annotation was performed using the GeSeq server [86] with default settings and NC\_028096 (mtDNA sequence of *Populus tremula*) set as reference. Using the Sequin tool v13.05 [87], these draft annotations were corrected where necessary, guided by alignments to other well-characterized eudicot mtDNA sequences including those of *Arabidopsis thaliana* (NC\_037304.1) [88], *Cucurbita pepo* (NC\_014050.1) [65], and *Quercus variabilis*(MN199236) [9]. Additional ORFs including the ORF of the gene *rpl10* were identified by ORF prediction in transcript contigs assembled from RNA-Seq data of two individuals (see Section 4.6) using the "Find open reading frames"—tool of CLC Genomics Workbench version 20.0.3 with the following parameters: Start codons = AUG; Stop codon included in annotation = Yes; Annotate sequences = Yes; Both strands = Yes; Open-ended sequence = No; Genetic code = 1 Standard; Minimum length (codons) = 100. In total, 90 ORFs with a minimal length of 300 bp (including start and stop codons) were identified and gave rise to 24 additional annotations (ORF 1–23 and *rpl10*), which were manually added to the Seqin file (annotated as "hypothetical proteins"). Nested ORFs were not considered. If BlastP analysis of the amino acid sequence of a related ORF versus NCBI non-redundant proteins provided hits with at least 80% query coverage and 90% identity to any known protein sequence; then, this information was included into the protein description (as a comment) in the Seqin file.

To compare and visualize the gene order between different mitochondrial genomes, the web version of the software tool geneCo (gene Comparison) [89,90] was applied (using the "map comparison" option with default parameters, but displaying names of all genes; considering only potential protein-coding genes and excluding trans-spliced genes).
