Assembly and Comparative Analysis of the Mitochondrial Genomes of the Genus Castanea

Plant mitochondrial genomes (mtDNA) are dynamic and rapidly evolving, varying greatly in size, gene content, and structure even among closely related species. This has caused the sequencing and assembly of mtDNA reference sequences to lag significantly behind the total number of reference chloroplast genomes. Recent studies have noted that read-depth of contigs in draft de novo assemblies might assist in partitioning assembled sequences to their respective organellar compartments. We developed an approach combining sequence homology and organellar stoichiometry to identify subsets of contigs from a draft assembly of the Chinese Chestnut (castanea mollissima) belonging to each of three organellar compartments (nucleus, mitochondria, and chloroplast.) Short-read and large-insert sequencing libraries were used to extend and join these contigs into a single circular contig approximately 388kb in size, representing a draft complete mtDNA for C. millissima. Resequence data from seven other species within the genus (~5-7x coverage) was used to generate consensus references for the mtDNA. Full length alignments were used to reconstruct the mitochondrial history of the genus. The mtDNA of the next nearest species (C. henryii) was reassembled and aligned to the reference mtDNA, the results of which suggest the presence of a large number of rearrangements between species. This methodology will be useful for quickly generating new reference mtDNAs for species with genomes at the draft level of assembly, as well as for individuals and species resequenced at low coverage with taxonomically near mtDNA references.