First genomic prediction and genome-wide association for complex growth-related traits in Rock Bream (Oplegnathus fasciatus)

Abstract

Rock Bream (Oplegnathus fasciatus) is an important aquaculture species for offshore cage aquaculture and fish stocking of marine ranching in East Asia. Genomic selection has the potential to expedite genetic gain for the key target traits of a breeding program, but has not yet been evaluated in Oplegnathus. The purposes of the present study were to explore the performance of genomic selection to improve breeding value accuracy through real data analyses using six statistical models and to carry out genome-wide association studies (GWAS) to dissect the genetic architecture of economically vital growth-related traits (body weight, total length, and body depth) in the O. fasciatus population. After quality control, genotypes for 16,162 SNPs were acquired for 455 fish. Heritability was estimated to be moderate for the three traits (0.38 for BW, 0.33 for TL, and 0.24 for BD), and results of GWAS indicated that the underlying genetic architecture was polygenic. Six statistic models (GBLUP, BayesA, BayesB, BayesC, Bayesian Ridge-Regression, and Bayesian LASSO) showed similar performance for the predictability of genomic estimated breeding value (GEBV). The low SNP density (around 1 K selected SNP based on GWAS) is sufficient for accurate prediction on the breeding value for the three growth-related traits in the current studied population, which will provide a good compromise between genotyping costs and predictability in such standard breeding populations advanced. These consequences illustrate that the employment of genomic selection in O. fasciatus breeding could provide advantages for the selection of breeding candidates to facilitate complex economic growth traits.

Keywords: Oplegnathus fasciatus; genome‐wide association; genomic selection; growth trait.

FIGURE 1

Manhattan plot for the significance of each genetic variant in GWAS for (a) body weight (BW), (b) total length (TL), and (c) body depth (BD). The black line denotes the genome‐wide significance threshold, and the red line denotes the suggestive threshold

FIGURE 2

Estimates of predictability with increasing numbers of SNP for body weight (BW) using three cumulative approach to SNP sampling. The orange line indicated gradually increased markers that sorted by GWAS (p‐value in ascending order). The blue line indicated gradually increased markers that were randomly selected from all available SNPs. The red line indicated gradually increased markers that spaced evenly across the genome. The solid line indicated the mean value at each marker number, and the shaded area was formed by connected the dot of positive and negative standard deviations

FIGURE 3

Genomic estimates of predictability for O. fasciatus using different SNP sampling methods. 14 SNP subsets were selected using GWAS informative, randomly and evenly sampled SNPs (0.05 K, 0.1 K, 0.2 K, 0.3 K, 0.5 K, 0.8 K, 1 K, 2 K, 3 K, 5 K, 8 K, 10 K, 13 K, ~16 K SNPs)