Re-evaluating the genetic variation of the COI gene of Insecta: Implications for DNA barcoding, metabarcoding and species delimitation studies

To analyze the genetic variation of the cytochrome c oxidase subunit I (COI) gene of Insecta, the COI data of Insecta was downloaded from GenBank and the intraspecific Kimura-2 -parameter (K2P) distance of 40,782 species was calculated (each species with three or more sequences). Our result indicated that the maximum intraspecific genetic distance of 8,928 (21.89%) species was over 3%. Using a threshold of 3% in the clustering analysis, 7,123 (17.47%) species can be divided into two or more clusters. We also analyzed 3,189 genera with over three species (25,283 species) and found that the optimal thresholds for these genera ranged from 0.1%-15.7% (average value: 0.03531, median value: 0.02900). In clustering analysis, if the threshold values were set to 0.01, 0.02, 0.022, and 0.03, the numbers of clusters were 39,860, 31,024, 29,954, and 26,527, respectively. In metabarcoding studies, a threshold of 0.03 was recommended to estimate the species diversity of insects in a certain environment. However, using the empirical thresholds mentioned above for operational taxonomic unit (OTU) picking, the average match ratios of the 3,189 genera were 0.5137, 0.6338, 0.6440, and 0.6587. By contrast, if the possible thresholds from the distance matrix, the minimum interspecific genetic distance of congeneric species, and the optimal thresholds were used in clustering analysis, the average match ratios of them were 0.6626 0.7530, and 0.7549. Herein, we recommended the utilizations of the minimum interspecific genetic distance (when it was greater than or equal to 2%) and the optimal thresholds for OTU picking in DNA barcoding and species delimitation studies based on the COI gene of insects.