Five opah

Despite its commercial value and common occurrence in pelagic fisheries, surprisingly little is known about the basic biology and ecology of the opah (Lampris guttatus) also known as Moonfish. Opah is becoming increasingly popular in seafood markets, a trend that started in the late 1980s when it was served as sushi or sashimi. I find this particularly interesting given that about 35% of an opah's weight is consumable meat, while the remaining 65% are bone and skin. 

A group of researchers from the US had a closer look into the diversity of the opah:

In the course of collecting additional life history data on this species, observations were made by the authors and other researchers that suggested the presence of two distinct morphotypes in the North Pacific. We subsequently initiated a study to investigate these differences and began collecting tissue samples for genetic analysis and compiling standard morphometric data. 

The results are quite interesting:

Sequencing of the mitochondrial cytochrome c oxidase I gene (655bp) for these morphotypes and other specimens collected worldwide (n=480) produced five strongly diverged and well-supported clades. Additional sequence data from the cytochrome b gene (1141bp) as well as the nuclear recombination activating gene 1 (1323bp) corroborated these results, suggesting these five clades likely represent separate species. Our conclusion that opah is a complex of five separate species has implications for management and indicates a need to gather additional data on these poorly understood fishes.

Fig 1: Subset of a lampriform NJ tree of available
COI sequences about 4 years ago.

I recalled that a few years ago I helped to generate a tree for a monograph about a relative of the opah, the king of herrings (Regalecus glesne) and in order to provide some more context I had included some publicly available sequences for both species of the genus Lampris. With the limited dataset at that time I already found two very distinct clades within Lampris guttatus (see figure 1) but we didn't bother to pursue this further. Reading the new publication I immediately became curious how old and new data would fit together.  

Fig 2: NJ tree of Lampris guttatus COI sequences, labels show BIN# and
Lineage assignment according to Hyde et al.

The authors provided GenBank accessions to each of the haplotypes they found for each marker. I fetched the COI sequences from the database and combined them with what I was able to find on BOLD (and GenBank) for the same species. A bit of editing and I had an alignment for a quick NJ tree (see figure 2). I wasn't surprised to retrieve all five lineages that are presented in the paper. What I found surprising was the fact that despite a much smaller sample size (<20) four of those were already present on BOLD and had BINs assigned to them. Most of this cryptic diversity was already hiding in the database as BOLD does BIN assignments automatically on a regular basis. Once more BINs seem to be a good proxy for species. I am not saying this will always be the case as I also know examples where BIN and species are not congruent but there is a good chance that there are many more of such hidden stories just waiting to be discovered by digging deeper into the barcode treasure chest BOLD. Some 3 Million DNA Barcodes certainly justify some 'treasure hunts'.