Discussion
This study evaluates simple and cost-effective methods for the discovery of fish mitogenomes for the purpose of building up fish mitochondrial databases. We combined long fragment sequencing and de novo assembly to generate unbiased consensus sequences that do not miss gene duplications, order variations, homopolymeric or tandem repeat regions. We used two mitochondrial DNA enrichment approaches in combination with two different Oxford Nanopore sequencing kits: the mitoenrichment enhances the amount of mtDNA at the extraction and post-extraction steps and can be applied to all fish while the targeted mitosequencing favors the sequencing of mtDNA from previously dephosphorylated genomic DNA with the aid of a CRISPR Cas9 enzyme in combination with five guide RNA sequences that direct the cuts and can be applied to a majority of fish species. Indeed, we obtained novel complete mitogenomes, some following the typical vertebrate mitogenome, some with extremely repetitive segments in the control region and others with gene rearrangements and indels. We generated a new reference mitogenome for chinook (Oncorhynchus tshawytscha ) and complete mitogenome of another two key species of economical and conservation importance, the Pacific hake (Merluccius productus ) and eulachon (Thaleichthys pacificus ). We see the methods tested here as an immediate, speedy, practical sequencing approach that provides certainty and quality to the mitogenome contigs.
Myctophids (lanterfishes and blackchins) are a common finding in our laboratory metabarcoding studies from oceanic samples, being the most abundant family of fishes in the oceans in terms of biomass. Adding resolution further than Myctophidae ASV beyond the phylogenetic effort by Poulsen et al. (2013) is important to resolve species within eDNA metabarcoding work. Myctophiformes is a very diverse family with 250 species described with very divergent larval forms that is far from complete in the molecular databases (Poulsen et al., 2013, also see Fig S3.2). Moreover, very few specimens remain identifiable in survey net hauls because their scales and photophores pattern are often damaged. Getting more voucher specimens sequenced is thus important and de novo assembly of this group of fishes is key because the gene rearrangement made mapping to a close relative mitogenome insufficient, resulting in shorter genomes with missing genes (data not shown). The complete mitogenomes of the blue lanternfish (Tarletonbeania crenularis ), Northern lampfish (Stenobrachius leucopsarus ), and California headlighfish (Diaphus theta ) were generated for the first time in this study. Prior myctophid mitogenomes (Poulsen et al., 2013) were almost complete but were missing the last ca. 2000 bp from the tRNA-Thr, t-RNA-Pro and the control region, due to the difficulty of sequencing such highly polymorphic and repetitive region (e.g. 44 G and >30 repetitions of 18bp on S. leucopsarus ).
Figure 4. Sequential order of the genes found in the six species studied. The tRNA are noted with their single-letter aminoacid code and there are either they are placed on above or below their line to denote in which strand they are transcribed.