Multi-centre testing and validation of current protocols for the identification of Gyrodactylus salaris (Monogenea).

Despite routine screening requirements for the notifiable fish pathogen Gyrodactylus salaris, no standard operating procedure exists for its rapid identification and discrimination from other species of Gyrodactylus. This study assessed screening and identification efficiencies under real-world conditions for the most commonly employed identification methodologies: visual, morphometric and molecular analyses. Obtained data were used to design a best-practice processing and decision-making protocol allowing rapid specimen throughput and maximal classification accuracy. True specimen identities were established using a consensus from all three identification methods, coupled with the use of host and location information. The most experienced salmonid gyrodactylid expert correctly identified 95.1% of G. salaris specimens. Statistical methods of classification identified 66.7% of the G. salaris, demonstrating the need for much wider training. Molecular techniques (internal transcribed spacer region-restriction fragment length polymorphism (ITS-RFLP)/cytochrome c oxidase I (COI) sequencing) conducted in the diagnostic laboratory most experienced in the analysis of gyrodactylid material, identified 100% of the true G. salaris specimens. Taking into account causes of potential specimen loss, the probabilities of a specimen being accurately identified were 95%, 87% and 92% for visual, morphometric and molecular techniques, respectively, and the probabilities of correctly identifying a specimen of G. salaris by each method were 81%, 58% and 92%. Inter-analyst agreement for 189 gyrodactylids assessed by all three methods using Fleiss' Kappa suggested substantial agreement in identification between the methods. During routine surveillance periods when low numbers of specimens are analysed, we recommend that specimens be analysed using the ITS-RFLP approach followed by sequencing of specimens with a "G. salaris-like" (i.e. G. salaris, Gyrodactylus thymalli) banding pattern. During periods of suspected outbreaks, where a high volume of specimens is expected, we recommended that specimens be identified using visual identification, as the fastest processing method, to select "G. salaris-like" specimens, which are subsequently identified by molecular-based techniques.