Results and Discussion
After BLASTN annotation analyses were conducted, K2P distances were compared at different taxonomic levels, revealing distinct features in the sequences both within and between species. With respect to the COI sequences of the 156 species represented, the interspecific K2P distances for the COI sequences from the 68 fish species, the 59 shellfish species, and the 29 insect species ranged from 0% to 45.25% (fish, 0% to 40.99%; insects, 0% to 10.34%; shellfish, 0% to 45.25%) (Fig. 2A), whereas the intraspecific K2P distances with ≥3 sequences ranged from 0% to 0.985% (fish, 0% to 0.985%; insects, 0.005% to 0.635%; shellfish, 0% to 0.817%) (Fig. 2B). The average interspecific distances and average intraspecific distances were, respectively, 3.58 and 0.14 in shellfish, 3.22 and 0.41 in fish, and 2.06 and 0.25 in insects (Table 1). In shellfish, the greatest interspecific K2P differences were 25.57-fold higher than the intraspecific values. The overall base composition in each species of fish, insect, and shellfish was as follows: T (thymine) ranged from 27.4% to 33.7% (highly abundant); G (guanine) ranged from 16.8% to 21.5% (not highly abundant) (Table 1). These findings for fish were consistent with previous studies showing that T occurred more frequently and G occurred less frequently than A (adenine) and C (cytosine) [8].
In our polytypic species analysis with more than 3 individuals in each species, the average intraspecific difference was approximately 0.5%, and the maximum intraspecific divergence was only 1.86% (Table 2). The highest overall GC% content was found in the 18 species of fish. Lower values were found in the 2 species of insects and in the 6 species of shellfish (Table 2). The fish Chelidonichthys spinosus had a high GC% content of 50.9%. The mean GC% content of the 18 barcoded fish species was higher than that of the 6 shellfish species (46.9 ± 2.2% vs. 38.0 ± 4.9%) (see also Table 2). Sixteen of the 21 species with GC% content ≥45% were fish, whereas only 1 shellfish species exhibited GC% content ≥45%. The GC% content can be used in a new approach to evaluate animal evolutionary relationships, although the relationship between GC% content and the evolutionary branching date is not very accurate [14]. Moreover, the average divergence of congeneric species pairs was greater than that found for intraspecific differences, but 10 species in 5 genera had interspecific distances below 0.1% (Table 3). These species included Hexagrammos agrammus/H. otakii, Ampedus humeralis/A. subcostatus, Anomala luculenta/A. mongolica, Chlorostoma argyrostoma turbinatum/C. turbinate, and Omphalius rusticus rusticus/O. pfeifferi carpenteri. In addition, the NJ tree exhibited shallow interspecific divergence except at the first deep divergence (Fig. 3). In fish, several clades had a high level of bootstrap support (≥97%) (Fig. 3A). These clades included Thrysa chefuensis and T. adelae, Hexagrammos otakii and H. agrammus. In insects, the clades that had a high level of bootstrap support (≥95%) included Fusinus forceps, F. longicaudus, Mytilus galloprovincialis, and M. edulis. In shellfish, 2 clades separated out with a high level of bootstrap support (≥99%) (Fig. 3B). These clades included Anomala mongolica and A. luculenta, Ampedus humeralis and A. subcostatus (Fig. 3C).
In conclusion, we obtained DNA barcodes using COI sequences from fish, insects, and shellfish. The aims of this research were species identification and contribution to biodiversity research. At the species level, the rate of correct identifications might be low in a diversified environment. However, DNA barcoded sequences can be used for the interpretation of species-level identification and community-level patterns in fish, insects, and shellfish.