Methods
The first community-level barcoding studies were conducted in the most diverse terrestrial and marine ecosystems in an inland and coastal area of South Korea (include reference). We collected samples to obtain an overview of the variation patterns for 529 COI sequences among 68 fish species, 29 insect species, and 59 shellfish species. Multiple specimens were collected for most of the species. Fish and shellfish were collected from Yeosu in Jeollanam-do; shellfish were collected from Taean; and insects were collected from Chungcheongnam-do, Gangwon-do, Gyeongsangbuk-do, and Jeollabuk-do in South Korea. Samples were collected using different, technically appropriate methods (Fig. 1, Supplementary Table 1) [9]. If possible, the samples were obtained from widely distributed places in South Korea.
Genomic DNA was isolated from samples using the Qiagen DNeasy 96 blood and tissue kit (Qiagen, Valencia, CA, USA) according to the instructions. DNA fragments of target genes were amplified by polymerase chain reaction (PCR) with primers for the COI gene (primer sequences: LCO1490 GGTCAACAAATCATAAAGATATTGG and HCO 2198 TAAACTTCAGGGTGACCAAAAAATCA) [10]. PCR amplification was performed using Top-Taq PreMix (2×; CoreBio, Seoul, Korea) under the following conditions: denaturation (1 min at 94℃), annealing at 51℃ for amplification of the COI gene, and extension (2 min at 72℃). PCR products were purified with the Core-One PCR purification kit (CoreBio), and TA cloning was performed using the pGEM-T Easy Vector system (Promega, Madison, WI, USA) by Macrogen Inc. The clones for each marker were sequenced with forward (SP6) and reverse (T7) primers using an ABI 3730XL sequencer (Applied Biosystems, Foster City, CA, USA). The sequences reported in this paper have been deposited in GenBank under accession numbers HM180413-HM180941.
To obtain the species information for each operational taxonomic unit (OTU) in a phylogenetic tree, a BLAST search was performed using the BLASTN program from NCBI [11]. A cutoff value for the BLAST result was established as follows: query coverage > 90% and identity > 75% for COI. The levels of sequence divergence within and between the selected species were investigated using the pairwise Kimura 2 parameter (K2P) distance model [12]. The neighbor-joining tree, with gap positions ignored on a pairwise basis, was constructed using the neighbor-joining (NJ) method with K2P distances in MEGA4 [13]. These distances were hierarchically arranged in accordance with intraspecific and interspecific species differences within each genus. When the sequence dataset consisted of only 2 genera from the same family, an intergeneric comparison within the family was not performed.