Molecular Ecology Notes (2007) doi: 10.1111/j.1471-8286.2007.01748.x © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd Blackwell Publishing Ltd BARCODING Universal primer cocktails for fish DNA barcoding NATALIA V. IVANOVA,* TYLER S. ZEMLAK, ROBERT H. HANNER and PAUL D. N. HEBERT Canadian Centre for DNA Barcoding, Biodiversity Institute of Ontario, University of Guelph, Guelph, Ontario, Canada N1G 2W1 Abstract Reliable recovery of the 5′ region of the cytochrome c oxidase 1 (COI) gene is critical for the ongoing effort to gather DNA barcodes for all fish species. In this study, we develop and test primer cocktails with a view towards increasing the efficiency of barcode recovery. Specifically, we evaluate the success of polymerase chain reaction amplification and the quality of resultant sequences using three primer cocktails on DNA extracts from repre- sentatives of 94 fish families. Our results show that M13-tailed primer cocktails are more effective than conventional degenerate primers, allowing barcode work on taxonomically diverse samples to be carried out in a high-throughput fashion. Keywords: COI, degenerate primers, M13-tailed primers, species identification Received 11 January 2007; revision accepted 7 February 2007 Introduction Fishes comprise nearly half of all vertebrate species; the group includes approximately 15 700 marine and 13 700 freshwater species (FishBase: www.fishbase.org). The Fish Barcode of Life Initiative (FISH-BOL; www.fishbol.org) is a collaborative international research effort, which seeks to establish a reference library of DNA barcodes for all fish species derived from voucher specimens with authorit- ative taxonomic identifications (Hanner et al . 2005). Once completed, FISH-BOL will enable a fast, accurate, and cost-effective system for molecular identification of the world’s ichthyofauna. The benefits of this work include facilitating species identification, flagging potentially previously unrecognized species, and enabling identifica- tions where traditional methods are not applicable, such as for immature stages or body fragments. FISH-BOL will also provide a powerful tool for enhanced understanding of the natural history and ecological interactions of fish species. Obtaining high-quality sequence records from the bar- code region of COI is a key requirement for the FISH-BOL enterprise. Moreover, with the goal of analysing at least 10 specimens per species, this effort will likely involve sequencing more than 0.5 million specimens. Ward et al . (2005) carried out a proof-of-principle study that compiled bar- codes for 200 species of commercially important Australian marine fishes. Since then, an additional 5000 barcodes have been generated from over 2000 species. However, there has not been a serious effort to hone analytical protocols, a gap that this study addresses. Specifically, we seek to identify protocols that enable both efficient polymerase chain rea- ction (PCR) amplification of the barcode region and that deliver high quality sequence data. Ward et al . (2005) used two forward and two reverse primers in all four pairwise combinations to amplify DNA barcodes from the 200 species in their study. This approach delivered amplicons for all but one of the target species, demonstrating the reliable amplification of COI from a diversity of fishes with only a few primers. However, the need for four PCR amplifications of each specimen is unde- sirably complex. Two possible solutions include the gener- ation of a single primer set with degenerate sites or the techniques in this study assembly of a cocktail whose component primers are tailed with M13 to facilitate high throughput sequencing. We employ both, and test their effectiveness on a single species from each of 94 different fish families from diverse habitats (i.e. freshwater, diadram- ous, marine) and divergent evolutionary lineages. Materials and methods Primer design COI sequences from all 159 mitochondrial fish genomes (GenBank, January 2006) were aligned in bioedit (Hall 1999). Potential primer regions were analysed in codehop (Rose et al. 1988) available at (http://blocks.fhcrc.org/ Correspondence: N.V. Ivanova, Fax: (519) 824 5703; E-mail: nivanova@uoguelph.ca