Canadian Forest Service Publications
Automated high throughput animal CO1 metabarcode classification. 2018. Porter, T.M.; Hajibabaei, M. Scientific Reports 8: 4226.
Year: 2018
Issued by: Great Lakes Forestry Centre
Catalog ID: 39094
Language: English
Availability: PDF (download)
Available from the Journal's Web site. †
DOI: 10.1038/s41598-018-22505-4
† This site may require a fee
Plain Language Summary
Until now, there has been difficulty assigning names to animal barcode sequences isolated directly from eDNA in a rapid, high-throughput manner, providing a measure of confidence for each assignment. To address this gap, we have compiled nearly 1 million marker gene DNA barcode sequences appropriate for classifying chordates, arthropods, and flag members of other major eukaryote groups. We show that the RDP naive Bayesian classifier can assign the same number of queries 19 times faster than the popular BLAST top hit method and reduce the false positive rate by two-thirds. As reference databases become more representative of current species diversity, confidence in taxonomic assignments should continue to improve. We recommend that investigators can improve the performance of species-level assignments immediately by supplementing existing reference databases with full-length DNA barcode sequences from representatives of local fauna.
