Journal article
Threshold models using Gibbs sampling and machine learning genomic predictions for skin fluke disease recorded under field environment in yellowtail kingfish Seriola lalandi
Aquaculture, Vol.547, pp.1-7
2022
Abstract
Ectoparasitic infestations, such as skin fluke due to Benedenia seriolae, represent substantial fish health and welfare challenges that have caused significant industry implications. Unfortunately, existing methods used to control this disease such as bath-administered compounds or therapeutic (dietary) trials are labour intensive, costly and only temporarily effective. Genetic (genomic) methods may provide sustainable long-term solutions to improve host resistance to parasitic diseases in marine finfish species. This study thus explored possibilities for applying genomic selection to improve resistance to B. seriolae and deformity in yellowtail kingfish Seriola lalandi. Specifically, we used a total of 752 animals with individual records and 14,448 single nucleotide polymorphisms (SNPs) developed de novo from Diversity Arrays technology (DArT, a combination of genome complexity reduction methods and next generation sequencing platforms) to assess accuracy of genomic predictions for the incidence of skin fluke and deformity recorded under field (farm) condition. Our multi-locus linear and threshold mixed model analyses showed that heritabilities for the incidence of skin fluke and deformity were small (mean range h2 = 0.02–0.03). Genetic correlations of skin fluke with body weight and deformity were not significant. Our threshold Gibbs sampling and machine learning models revealed that the accuracies of genomic prediction were low for both traits (mean range from 0.151–0.432). Imputation of missing genotypes improved the prediction accuracies for skin fluke and deformity by 0.5–13.2%. Multi-trait analyses outperformed single trait models, only for deformity. Our findings suggest that genomic selection for reduced skin fluke and deformity, albeit possible in yellowtail kingfish, genetic progress made for these traits may be slow because of the low prediction accuracies across models used in this study. To enable the application of genomic selection for skin fluke or disease resistant traits recorded under field condition, it is necessary to sequence a larger number of individuals and families as well as to develop large-scale routine data recording of new phenotypes (e.g., parasite counts) to increase the prediction accuracies for these traits in the breeding program of yellowtail kingfish S. lalandi.
Details
- Title
- Threshold models using Gibbs sampling and machine learning genomic predictions for skin fluke disease recorded under field environment in yellowtail kingfish Seriola lalandi
- Authors
- Nguyen Hong Nguyen (Author) - University of the Sunshine Coast, Queensland, School of Science, Technology and EngineeringNguyen Thanh Vu (Author) - University of the Sunshine Coast, Queensland, School of Science, Technology and Engineering
- Publication details
- Aquaculture, Vol.547, pp.1-7
- Publisher
- Elsevier BV
- DOI
- 10.1016/j.aquaculture.2021.737513
- ISSN
- 1873-5622
- Organisation Unit
- Faculty of Arts, Business and Law; External; Centre for Bioinnovation; University of the Sunshine Coast, Queensland; School of Science, Technology and Engineering; Faculty of Science, Health, Education and Engineering
- Language
- English
- Record Identifier
- 99578906402621
- Output Type
- Journal article
Metrics
17 Record Views
InCites Highlights
These are selected metrics from InCites Benchmarking & Analytics tool, related to this output
- Collaboration types
- International collaboration
- Web Of Science research areas
- Fisheries
- Marine & Freshwater Biology
UN Sustainable Development Goals (SDGs)
This output has contributed to the advancement of the following goals:
Source: InCites