Preliminary construction and comparative analysis of environmental DNA metabarcoding reference database of freshwater fishes in Hainan Island

In order to determine the optimal reference database and target genes for environmental DNA study of freshwater fishes in Hainan Island, we compared the species coverage, annotation accuracy and threshold values of interspecific difference of COI, 12S and 16S between the self-built database and the public database. The results show that: 1) Seventy-two fish species were collected, among which 16 (COI), 20 (12S) and 22 (16S) species' reference sequences were provided for the first time. 2) Only 68.06% (COI), 66.67% (12S) and 69.44% (16S) of the fish had high similarity sequence in the public database. 3) The annotation accuracy based on the self-built database was significantly higher than that on the public database (COI: 100% vs 69.64%; 12S: 96.15% vs 67.30%; 16S: 96% vs 70%). 4) COI gene was the best target gene for identifying freshwater fishes in Hainan Island, followed by 16S gene. 5) The threshold values of interspecific difference based on K2P genetic distance were 0.006 9 (COI), 0.005 6 (12S) and 0.007 5 (16S), respectively, and the accuracy rates were 94.96% (COI), 89.05% (12S) and 92.70% (16S), respectively. This study reveals that the sequence annotation accuracy of the self-built database is significantly higher than that of the public database, and it is suggested that COI and 16S should be used as the environmental DNA metabarcoding genes of freshwater fishes in Hainan Island.